[llvm] 39641b1 - [AArch64][SME]: Add precursory tests for D138519

Hassnaa Hamdi via llvm-commits llvm-commits at lists.llvm.org
Thu Nov 24 10:19:22 PST 2022


Author: Hassnaa Hamdi
Date: 2022-11-24T18:19:13Z
New Revision: 39641b11733325e0d8091087f37edca9de6772d4

URL: https://github.com/llvm/llvm-project/commit/39641b11733325e0d8091087f37edca9de6772d4
DIFF: https://github.com/llvm/llvm-project/commit/39641b11733325e0d8091087f37edca9de6772d4.diff

LOG: [AArch64][SME]: Add precursory tests for D138519

Add testing files:
 - fp-select.ll
 - int-select.ll
 - fp-vselect.ll
 - int-vselect.ll

Added: 
    llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-select.ll
    llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-vselect.ll
    llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-select.ll
    llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-vselect.ll

Modified: 
    

Removed: 
    


################################################################################
diff  --git a/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-select.ll b/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-select.ll
new file mode 100644
index 000000000000..bfc25873aa06
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-select.ll
@@ -0,0 +1,286 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -force-streaming-compatible-sve < %s | FileCheck %s
+
+target triple = "aarch64-unknown-linux-gnu"
+
+define <2 x half> @select_v2f16(<2 x half> %op1, <2 x half> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v2f16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI0_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x9, :lo12:.LCPI0_0]
+; CHECK-NEXT:    strh w8, [sp, #14]
+; CHECK-NEXT:    strh w8, [sp, #12]
+; CHECK-NEXT:    strh w8, [sp, #10]
+; CHECK-NEXT:    strh w8, [sp, #8]
+; CHECK-NEXT:    ldr d2, [sp, #8]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <2 x half> %op1, <2 x half> %op2
+  ret <2 x half> %sel
+}
+
+define <4 x half> @select_v4f16(<4 x half> %op1, <4 x half> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v4f16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI1_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x9, :lo12:.LCPI1_0]
+; CHECK-NEXT:    strh w8, [sp, #14]
+; CHECK-NEXT:    strh w8, [sp, #12]
+; CHECK-NEXT:    strh w8, [sp, #10]
+; CHECK-NEXT:    strh w8, [sp, #8]
+; CHECK-NEXT:    ldr d2, [sp, #8]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <4 x half> %op1, <4 x half> %op2
+  ret <4 x half> %sel
+}
+
+define <8 x half> @select_v8f16(<8 x half> %op1, <8 x half> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v8f16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI2_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    ldr q3, [x9, :lo12:.LCPI2_0]
+; CHECK-NEXT:    strh w8, [sp, #14]
+; CHECK-NEXT:    strh w8, [sp, #12]
+; CHECK-NEXT:    strh w8, [sp, #10]
+; CHECK-NEXT:    strh w8, [sp, #8]
+; CHECK-NEXT:    strh w8, [sp, #6]
+; CHECK-NEXT:    strh w8, [sp, #4]
+; CHECK-NEXT:    strh w8, [sp, #2]
+; CHECK-NEXT:    strh w8, [sp]
+; CHECK-NEXT:    ldr q2, [sp]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <8 x half> %op1, <8 x half> %op2
+  ret <8 x half> %sel
+}
+
+define void @select_v16f16(ptr %a, ptr %b, i1 %mask) #0 {
+; CHECK-LABEL: select_v16f16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w2, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI3_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    ldr q0, [x0]
+; CHECK-NEXT:    ldr q1, [x0, #16]
+; CHECK-NEXT:    ldr q2, [x1]
+; CHECK-NEXT:    ldr q3, [x1, #16]
+; CHECK-NEXT:    strh w8, [sp, #14]
+; CHECK-NEXT:    strh w8, [sp, #12]
+; CHECK-NEXT:    ldr q5, [x9, :lo12:.LCPI3_0]
+; CHECK-NEXT:    strh w8, [sp, #10]
+; CHECK-NEXT:    strh w8, [sp, #8]
+; CHECK-NEXT:    strh w8, [sp, #6]
+; CHECK-NEXT:    strh w8, [sp, #4]
+; CHECK-NEXT:    strh w8, [sp, #2]
+; CHECK-NEXT:    strh w8, [sp]
+; CHECK-NEXT:    ldr q4, [sp]
+; CHECK-NEXT:    eor z5.d, z4.d, z5.d
+; CHECK-NEXT:    and z1.d, z1.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    and z3.d, z3.d, z5.d
+; CHECK-NEXT:    orr z0.d, z0.d, z2.d
+; CHECK-NEXT:    orr z1.d, z1.d, z3.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %op1 = load volatile <16 x half>, ptr %a
+  %op2 = load volatile <16 x half>, ptr %b
+  %sel = select i1 %mask, <16 x half> %op1, <16 x half> %op2
+  store <16 x half> %sel, ptr %a
+  ret void
+}
+
+define <2 x float> @select_v2f32(<2 x float> %op1, <2 x float> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v2f32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI4_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x9, :lo12:.LCPI4_0]
+; CHECK-NEXT:    stp w8, w8, [sp, #8]
+; CHECK-NEXT:    ldr d2, [sp, #8]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <2 x float> %op1, <2 x float> %op2
+  ret <2 x float> %sel
+}
+
+define <4 x float> @select_v4f32(<4 x float> %op1, <4 x float> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v4f32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI5_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    ldr q3, [x9, :lo12:.LCPI5_0]
+; CHECK-NEXT:    stp w8, w8, [sp, #8]
+; CHECK-NEXT:    stp w8, w8, [sp]
+; CHECK-NEXT:    ldr q2, [sp]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <4 x float> %op1, <4 x float> %op2
+  ret <4 x float> %sel
+}
+
+define void @select_v8f32(ptr %a, ptr %b, i1 %mask) #0 {
+; CHECK-LABEL: select_v8f32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w2, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI6_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    ldr q0, [x0]
+; CHECK-NEXT:    ldr q1, [x0, #16]
+; CHECK-NEXT:    ldr q2, [x1]
+; CHECK-NEXT:    ldr q3, [x1, #16]
+; CHECK-NEXT:    stp w8, w8, [sp, #8]
+; CHECK-NEXT:    stp w8, w8, [sp]
+; CHECK-NEXT:    ldr q5, [x9, :lo12:.LCPI6_0]
+; CHECK-NEXT:    ldr q4, [sp]
+; CHECK-NEXT:    eor z5.d, z4.d, z5.d
+; CHECK-NEXT:    and z1.d, z1.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    and z3.d, z3.d, z5.d
+; CHECK-NEXT:    orr z0.d, z0.d, z2.d
+; CHECK-NEXT:    orr z1.d, z1.d, z3.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %op1 = load volatile <8 x float>, ptr %a
+  %op2 = load volatile <8 x float>, ptr %b
+  %sel = select i1 %mask, <8 x float> %op1, <8 x float> %op2
+  store <8 x float> %sel, ptr %a
+  ret void
+}
+
+define <1 x double> @select_v1f64(<1 x double> %op1, <1 x double> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v1f64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    mov x9, #-1
+; CHECK-NEXT:    csetm x8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    fmov d3, x9
+; CHECK-NEXT:    fmov d2, x8
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <1 x double> %op1, <1 x double> %op2
+  ret <1 x double> %sel
+}
+
+define <2 x double> @select_v2f64(<2 x double> %op1, <2 x double> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v2f64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    adrp x9, .LCPI8_0
+; CHECK-NEXT:    csetm x8, ne
+; CHECK-NEXT:    stp x8, x8, [sp, #-16]!
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    ldr q2, [sp]
+; CHECK-NEXT:    ldr q3, [x9, :lo12:.LCPI8_0]
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <2 x double> %op1, <2 x double> %op2
+  ret <2 x double> %sel
+}
+
+define void @select_v4f64(ptr %a, ptr %b, i1 %mask) #0 {
+; CHECK-LABEL: select_v4f64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w2, #0x1
+; CHECK-NEXT:    ldr q0, [x0]
+; CHECK-NEXT:    csetm x8, ne
+; CHECK-NEXT:    ldr q1, [x0, #16]
+; CHECK-NEXT:    ldr q2, [x1]
+; CHECK-NEXT:    adrp x9, .LCPI9_0
+; CHECK-NEXT:    ldr q3, [x1, #16]
+; CHECK-NEXT:    stp x8, x8, [sp, #-16]!
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI9_0]
+; CHECK-NEXT:    ldr q5, [sp]
+; CHECK-NEXT:    eor z4.d, z5.d, z4.d
+; CHECK-NEXT:    and z1.d, z1.d, z5.d
+; CHECK-NEXT:    and z0.d, z0.d, z5.d
+; CHECK-NEXT:    and z2.d, z2.d, z4.d
+; CHECK-NEXT:    and z3.d, z3.d, z4.d
+; CHECK-NEXT:    orr z0.d, z0.d, z2.d
+; CHECK-NEXT:    orr z1.d, z1.d, z3.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %op1 = load volatile <4 x double>, ptr %a
+  %op2 = load volatile <4 x double>, ptr %b
+  %sel = select i1 %mask, <4 x double> %op1, <4 x double> %op2
+  store <4 x double> %sel, ptr %a
+  ret void
+}
+
+attributes #0 = { "target-features"="+sve" }

diff  --git a/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-vselect.ll b/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-vselect.ll
new file mode 100644
index 000000000000..c29f02a14fde
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-fp-vselect.ll
@@ -0,0 +1,257 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -force-streaming-compatible-sve < %s | FileCheck %s
+
+target triple = "aarch64-unknown-linux-gnu"
+
+define <2 x half> @select_v2f16(<2 x half> %op1, <2 x half> %op2, <2 x i1> %mask) #0 {
+; CHECK-LABEL: select_v2f16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    mov z3.s, z2.s[1]
+; CHECK-NEXT:    adrp x8, .LCPI0_0
+; CHECK-NEXT:    fmov w9, s2
+; CHECK-NEXT:    fmov w10, s3
+; CHECK-NEXT:    ptrue p0.h, vl4
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d2, [x8, :lo12:.LCPI0_0]
+; CHECK-NEXT:    adrp x8, .LCPI0_1
+; CHECK-NEXT:    strh w9, [sp, #8]
+; CHECK-NEXT:    strh w10, [sp, #10]
+; CHECK-NEXT:    ldr d3, [sp, #8]
+; CHECK-NEXT:    ldr d4, [x8, :lo12:.LCPI0_1]
+; CHECK-NEXT:    lsl z3.h, p0/m, z3.h, z2.h
+; CHECK-NEXT:    asrr z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select <2 x i1> %mask, <2 x half> %op1, <2 x half> %op2
+  ret <2 x half> %sel
+}
+
+define <4 x half> @select_v4f16(<4 x half> %op1, <4 x half> %op2, <4 x i1> %mask) #0 {
+; CHECK-LABEL: select_v4f16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI1_0
+; CHECK-NEXT:    adrp x9, .LCPI1_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.h, vl4
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x8, :lo12:.LCPI1_0]
+; CHECK-NEXT:    ldr d4, [x9, :lo12:.LCPI1_1]
+; CHECK-NEXT:    lsl z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    asr z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <4 x i1> %mask, <4 x half> %op1, <4 x half> %op2
+  ret <4 x half> %sel
+}
+
+define <8 x half> @select_v8f16(<8 x half> %op1, <8 x half> %op2, <8 x i1> %mask) #0 {
+; CHECK-LABEL: select_v8f16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI2_0
+; CHECK-NEXT:    adrp x9, .LCPI2_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.h, vl8
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    uunpklo z2.h, z2.b
+; CHECK-NEXT:    ldr q3, [x8, :lo12:.LCPI2_0]
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI2_1]
+; CHECK-NEXT:    lsl z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    asr z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <8 x i1> %mask, <8 x half> %op1, <8 x half> %op2
+  ret <8 x half> %sel
+}
+
+define void @select_v16f16(ptr %a, ptr %b) #0 {
+; CHECK-LABEL: select_v16f16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ldp q0, q1, [x1]
+; CHECK-NEXT:    adrp x8, .LCPI3_0
+; CHECK-NEXT:    ldp q3, q2, [x0]
+; CHECK-NEXT:    fcmeq v5.8h, v3.8h, v0.8h
+; CHECK-NEXT:    fcmeq v4.8h, v2.8h, v1.8h
+; CHECK-NEXT:    ldr q6, [x8, :lo12:.LCPI3_0]
+; CHECK-NEXT:    and z3.d, z3.d, z5.d
+; CHECK-NEXT:    and z2.d, z2.d, z4.d
+; CHECK-NEXT:    eor z4.d, z4.d, z6.d
+; CHECK-NEXT:    eor z6.d, z5.d, z6.d
+; CHECK-NEXT:    and z1.d, z1.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z6.d
+; CHECK-NEXT:    orr z1.d, z2.d, z1.d
+; CHECK-NEXT:    orr z0.d, z3.d, z0.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    ret
+  %op1 = load <16 x half>, ptr %a
+  %op2 = load <16 x half>, ptr %b
+  %mask = fcmp oeq <16 x half> %op1, %op2
+  %sel = select <16 x i1> %mask, <16 x half> %op1, <16 x half> %op2
+  store <16 x half> %sel, ptr %a
+  ret void
+}
+
+define <2 x float> @select_v2f32(<2 x float> %op1, <2 x float> %op2, <2 x i1> %mask) #0 {
+; CHECK-LABEL: select_v2f32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI4_0
+; CHECK-NEXT:    adrp x9, .LCPI4_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.s, vl2
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x8, :lo12:.LCPI4_0]
+; CHECK-NEXT:    ldr d4, [x9, :lo12:.LCPI4_1]
+; CHECK-NEXT:    lsl z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    asr z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <2 x i1> %mask, <2 x float> %op1, <2 x float> %op2
+  ret <2 x float> %sel
+}
+
+define <4 x float> @select_v4f32(<4 x float> %op1, <4 x float> %op2, <4 x i1> %mask) #0 {
+; CHECK-LABEL: select_v4f32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI5_0
+; CHECK-NEXT:    adrp x9, .LCPI5_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.s, vl4
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    uunpklo z2.s, z2.h
+; CHECK-NEXT:    ldr q3, [x8, :lo12:.LCPI5_0]
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI5_1]
+; CHECK-NEXT:    lsl z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    asr z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <4 x i1> %mask, <4 x float> %op1, <4 x float> %op2
+  ret <4 x float> %sel
+}
+
+define void @select_v8f32(ptr %a, ptr %b) #0 {
+; CHECK-LABEL: select_v8f32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ldp q0, q1, [x1]
+; CHECK-NEXT:    adrp x8, .LCPI6_0
+; CHECK-NEXT:    ldp q3, q2, [x0]
+; CHECK-NEXT:    fcmeq v5.4s, v3.4s, v0.4s
+; CHECK-NEXT:    fcmeq v4.4s, v2.4s, v1.4s
+; CHECK-NEXT:    ldr q6, [x8, :lo12:.LCPI6_0]
+; CHECK-NEXT:    and z3.d, z3.d, z5.d
+; CHECK-NEXT:    and z2.d, z2.d, z4.d
+; CHECK-NEXT:    eor z4.d, z4.d, z6.d
+; CHECK-NEXT:    eor z6.d, z5.d, z6.d
+; CHECK-NEXT:    and z1.d, z1.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z6.d
+; CHECK-NEXT:    orr z1.d, z2.d, z1.d
+; CHECK-NEXT:    orr z0.d, z3.d, z0.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    ret
+  %op1 = load <8 x float>, ptr %a
+  %op2 = load <8 x float>, ptr %b
+  %mask = fcmp oeq <8 x float> %op1, %op2
+  %sel = select <8 x i1> %mask, <8 x float> %op1, <8 x float> %op2
+  store <8 x float> %sel, ptr %a
+  ret void
+}
+
+define <1 x double> @select_v1f64(<1 x double> %op1, <1 x double> %op2, <1 x i1> %mask) #0 {
+; CHECK-LABEL: select_v1f64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    mov x9, #-1
+; CHECK-NEXT:    csetm x8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    fmov d3, x9
+; CHECK-NEXT:    fmov d2, x8
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <1 x i1> %mask, <1 x double> %op1, <1 x double> %op2
+  ret <1 x double> %sel
+}
+
+define <2 x double> @select_v2f64(<2 x double> %op1, <2 x double> %op2, <2 x i1> %mask) #0 {
+; CHECK-LABEL: select_v2f64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI8_0
+; CHECK-NEXT:    adrp x9, .LCPI8_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.d, vl2
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    uunpklo z2.d, z2.s
+; CHECK-NEXT:    ldr q3, [x8, :lo12:.LCPI8_0]
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI8_1]
+; CHECK-NEXT:    lsl z2.d, p0/m, z2.d, z3.d
+; CHECK-NEXT:    asr z2.d, p0/m, z2.d, z3.d
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <2 x i1> %mask, <2 x double> %op1, <2 x double> %op2
+  ret <2 x double> %sel
+}
+
+define void @select_v4f64(ptr %a, ptr %b) #0 {
+; CHECK-LABEL: select_v4f64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ldp q0, q1, [x1]
+; CHECK-NEXT:    adrp x8, .LCPI9_0
+; CHECK-NEXT:    ldp q3, q2, [x0]
+; CHECK-NEXT:    fcmeq v5.2d, v3.2d, v0.2d
+; CHECK-NEXT:    fcmeq v4.2d, v2.2d, v1.2d
+; CHECK-NEXT:    ldr q6, [x8, :lo12:.LCPI9_0]
+; CHECK-NEXT:    and z3.d, z3.d, z5.d
+; CHECK-NEXT:    and z2.d, z2.d, z4.d
+; CHECK-NEXT:    eor z4.d, z4.d, z6.d
+; CHECK-NEXT:    eor z6.d, z5.d, z6.d
+; CHECK-NEXT:    and z1.d, z1.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z6.d
+; CHECK-NEXT:    orr z1.d, z2.d, z1.d
+; CHECK-NEXT:    orr z0.d, z3.d, z0.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    ret
+  %op1 = load <4 x double>, ptr %a
+  %op2 = load <4 x double>, ptr %b
+  %mask = fcmp oeq <4 x double> %op1, %op2
+  %sel = select <4 x i1> %mask, <4 x double> %op1, <4 x double> %op2
+  store <4 x double> %sel, ptr %a
+  ret void
+}
+
+attributes #0 = { "target-features"="+sve" }

diff  --git a/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-select.ll b/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-select.ll
new file mode 100644
index 000000000000..989fb79060ce
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-select.ll
@@ -0,0 +1,427 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -force-streaming-compatible-sve < %s | FileCheck %s
+
+target triple = "aarch64-unknown-linux-gnu"
+
+define <4 x i8> @select_v4i8(<4 x i8> %op1, <4 x i8> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v4i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI0_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x9, :lo12:.LCPI0_0]
+; CHECK-NEXT:    strh w8, [sp, #14]
+; CHECK-NEXT:    strh w8, [sp, #12]
+; CHECK-NEXT:    strh w8, [sp, #10]
+; CHECK-NEXT:    strh w8, [sp, #8]
+; CHECK-NEXT:    ldr d2, [sp, #8]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <4 x i8> %op1, <4 x i8> %op2
+  ret <4 x i8> %sel
+}
+
+define <8 x i8> @select_v8i8(<8 x i8> %op1, <8 x i8> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v8i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI1_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x9, :lo12:.LCPI1_0]
+; CHECK-NEXT:    strb w8, [sp, #15]
+; CHECK-NEXT:    strb w8, [sp, #14]
+; CHECK-NEXT:    strb w8, [sp, #13]
+; CHECK-NEXT:    strb w8, [sp, #12]
+; CHECK-NEXT:    strb w8, [sp, #11]
+; CHECK-NEXT:    strb w8, [sp, #10]
+; CHECK-NEXT:    strb w8, [sp, #9]
+; CHECK-NEXT:    strb w8, [sp, #8]
+; CHECK-NEXT:    ldr d2, [sp, #8]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <8 x i8> %op1, <8 x i8> %op2
+  ret <8 x i8> %sel
+}
+
+define <16 x i8> @select_v16i8(<16 x i8> %op1, <16 x i8> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v16i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI2_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    ldr q3, [x9, :lo12:.LCPI2_0]
+; CHECK-NEXT:    strb w8, [sp, #15]
+; CHECK-NEXT:    strb w8, [sp, #14]
+; CHECK-NEXT:    strb w8, [sp, #13]
+; CHECK-NEXT:    strb w8, [sp, #12]
+; CHECK-NEXT:    strb w8, [sp, #11]
+; CHECK-NEXT:    strb w8, [sp, #10]
+; CHECK-NEXT:    strb w8, [sp, #9]
+; CHECK-NEXT:    strb w8, [sp, #8]
+; CHECK-NEXT:    strb w8, [sp, #7]
+; CHECK-NEXT:    strb w8, [sp, #6]
+; CHECK-NEXT:    strb w8, [sp, #5]
+; CHECK-NEXT:    strb w8, [sp, #4]
+; CHECK-NEXT:    strb w8, [sp, #3]
+; CHECK-NEXT:    strb w8, [sp, #2]
+; CHECK-NEXT:    strb w8, [sp, #1]
+; CHECK-NEXT:    strb w8, [sp]
+; CHECK-NEXT:    ldr q2, [sp]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <16 x i8> %op1, <16 x i8> %op2
+  ret <16 x i8> %sel
+}
+
+define void @select_v32i8(ptr %a, ptr %b, i1 %mask) #0 {
+; CHECK-LABEL: select_v32i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w2, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI3_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    ldr q0, [x0]
+; CHECK-NEXT:    ldr q1, [x0, #16]
+; CHECK-NEXT:    ldr q2, [x1]
+; CHECK-NEXT:    ldr q3, [x1, #16]
+; CHECK-NEXT:    strb w8, [sp, #15]
+; CHECK-NEXT:    strb w8, [sp, #14]
+; CHECK-NEXT:    ldr q5, [x9, :lo12:.LCPI3_0]
+; CHECK-NEXT:    strb w8, [sp, #13]
+; CHECK-NEXT:    strb w8, [sp, #12]
+; CHECK-NEXT:    strb w8, [sp, #11]
+; CHECK-NEXT:    strb w8, [sp, #10]
+; CHECK-NEXT:    strb w8, [sp, #9]
+; CHECK-NEXT:    strb w8, [sp, #8]
+; CHECK-NEXT:    strb w8, [sp, #7]
+; CHECK-NEXT:    strb w8, [sp, #6]
+; CHECK-NEXT:    strb w8, [sp, #5]
+; CHECK-NEXT:    strb w8, [sp, #4]
+; CHECK-NEXT:    strb w8, [sp, #3]
+; CHECK-NEXT:    strb w8, [sp, #2]
+; CHECK-NEXT:    strb w8, [sp, #1]
+; CHECK-NEXT:    strb w8, [sp]
+; CHECK-NEXT:    ldr q4, [sp]
+; CHECK-NEXT:    eor z5.d, z4.d, z5.d
+; CHECK-NEXT:    and z1.d, z1.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    and z3.d, z3.d, z5.d
+; CHECK-NEXT:    orr z0.d, z0.d, z2.d
+; CHECK-NEXT:    orr z1.d, z1.d, z3.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %op1 = load volatile <32 x i8>, ptr %a
+  %op2 = load volatile <32 x i8>, ptr %b
+  %sel = select i1 %mask, <32 x i8> %op1, <32 x i8> %op2
+  store <32 x i8> %sel, ptr %a
+  ret void
+}
+
+define <2 x i16> @select_v2i16(<2 x i16> %op1, <2 x i16> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v2i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI4_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x9, :lo12:.LCPI4_0]
+; CHECK-NEXT:    stp w8, w8, [sp, #8]
+; CHECK-NEXT:    ldr d2, [sp, #8]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <2 x i16> %op1, <2 x i16> %op2
+  ret <2 x i16> %sel
+}
+
+define <4 x i16> @select_v4i16(<4 x i16> %op1, <4 x i16> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v4i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI5_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x9, :lo12:.LCPI5_0]
+; CHECK-NEXT:    strh w8, [sp, #14]
+; CHECK-NEXT:    strh w8, [sp, #12]
+; CHECK-NEXT:    strh w8, [sp, #10]
+; CHECK-NEXT:    strh w8, [sp, #8]
+; CHECK-NEXT:    ldr d2, [sp, #8]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <4 x i16> %op1, <4 x i16> %op2
+  ret <4 x i16> %sel
+}
+
+define <8 x i16> @select_v8i16(<8 x i16> %op1, <8 x i16> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v8i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI6_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    ldr q3, [x9, :lo12:.LCPI6_0]
+; CHECK-NEXT:    strh w8, [sp, #14]
+; CHECK-NEXT:    strh w8, [sp, #12]
+; CHECK-NEXT:    strh w8, [sp, #10]
+; CHECK-NEXT:    strh w8, [sp, #8]
+; CHECK-NEXT:    strh w8, [sp, #6]
+; CHECK-NEXT:    strh w8, [sp, #4]
+; CHECK-NEXT:    strh w8, [sp, #2]
+; CHECK-NEXT:    strh w8, [sp]
+; CHECK-NEXT:    ldr q2, [sp]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <8 x i16> %op1, <8 x i16> %op2
+  ret <8 x i16> %sel
+}
+
+define void @select_v16i16(ptr %a, ptr %b, i1 %mask) #0 {
+; CHECK-LABEL: select_v16i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w2, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI7_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    ldr q0, [x0]
+; CHECK-NEXT:    ldr q1, [x0, #16]
+; CHECK-NEXT:    ldr q2, [x1]
+; CHECK-NEXT:    ldr q3, [x1, #16]
+; CHECK-NEXT:    strh w8, [sp, #14]
+; CHECK-NEXT:    strh w8, [sp, #12]
+; CHECK-NEXT:    ldr q5, [x9, :lo12:.LCPI7_0]
+; CHECK-NEXT:    strh w8, [sp, #10]
+; CHECK-NEXT:    strh w8, [sp, #8]
+; CHECK-NEXT:    strh w8, [sp, #6]
+; CHECK-NEXT:    strh w8, [sp, #4]
+; CHECK-NEXT:    strh w8, [sp, #2]
+; CHECK-NEXT:    strh w8, [sp]
+; CHECK-NEXT:    ldr q4, [sp]
+; CHECK-NEXT:    eor z5.d, z4.d, z5.d
+; CHECK-NEXT:    and z1.d, z1.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    and z3.d, z3.d, z5.d
+; CHECK-NEXT:    orr z0.d, z0.d, z2.d
+; CHECK-NEXT:    orr z1.d, z1.d, z3.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %op1 = load volatile <16 x i16>, ptr %a
+  %op2 = load volatile <16 x i16>, ptr %b
+  %sel = select i1 %mask, <16 x i16> %op1, <16 x i16> %op2
+  store <16 x i16> %sel, ptr %a
+  ret void
+}
+
+define <2 x i32> @select_v2i32(<2 x i32> %op1, <2 x i32> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v2i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI8_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x9, :lo12:.LCPI8_0]
+; CHECK-NEXT:    stp w8, w8, [sp, #8]
+; CHECK-NEXT:    ldr d2, [sp, #8]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <2 x i32> %op1, <2 x i32> %op2
+  ret <2 x i32> %sel
+}
+
+define <4 x i32> @select_v4i32(<4 x i32> %op1, <4 x i32> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v4i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI9_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    ldr q3, [x9, :lo12:.LCPI9_0]
+; CHECK-NEXT:    stp w8, w8, [sp, #8]
+; CHECK-NEXT:    stp w8, w8, [sp]
+; CHECK-NEXT:    ldr q2, [sp]
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <4 x i32> %op1, <4 x i32> %op2
+  ret <4 x i32> %sel
+}
+
+define void @select_v8i32(ptr %a, ptr %b, i1 %mask) #0 {
+; CHECK-LABEL: select_v8i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    sub sp, sp, #16
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    tst w2, #0x1
+; CHECK-NEXT:    adrp x9, .LCPI10_0
+; CHECK-NEXT:    csetm w8, ne
+; CHECK-NEXT:    ldr q0, [x0]
+; CHECK-NEXT:    ldr q1, [x0, #16]
+; CHECK-NEXT:    ldr q2, [x1]
+; CHECK-NEXT:    ldr q3, [x1, #16]
+; CHECK-NEXT:    stp w8, w8, [sp, #8]
+; CHECK-NEXT:    stp w8, w8, [sp]
+; CHECK-NEXT:    ldr q5, [x9, :lo12:.LCPI10_0]
+; CHECK-NEXT:    ldr q4, [sp]
+; CHECK-NEXT:    eor z5.d, z4.d, z5.d
+; CHECK-NEXT:    and z1.d, z1.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    and z3.d, z3.d, z5.d
+; CHECK-NEXT:    orr z0.d, z0.d, z2.d
+; CHECK-NEXT:    orr z1.d, z1.d, z3.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %op1 = load volatile <8 x i32>, ptr %a
+  %op2 = load volatile <8 x i32>, ptr %b
+  %sel = select i1 %mask, <8 x i32> %op1, <8 x i32> %op2
+  store <8 x i32> %sel, ptr %a
+  ret void
+}
+
+define <1 x i64> @select_v1i64(<1 x i64> %op1, <1 x i64> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v1i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    mov x9, #-1
+; CHECK-NEXT:    csetm x8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    fmov d3, x9
+; CHECK-NEXT:    fmov d2, x8
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <1 x i64> %op1, <1 x i64> %op2
+  ret <1 x i64> %sel
+}
+
+define <2 x i64> @select_v2i64(<2 x i64> %op1, <2 x i64> %op2, i1 %mask) #0 {
+; CHECK-LABEL: select_v2i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    adrp x9, .LCPI12_0
+; CHECK-NEXT:    csetm x8, ne
+; CHECK-NEXT:    stp x8, x8, [sp, #-16]!
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    ldr q2, [sp]
+; CHECK-NEXT:    ldr q3, [x9, :lo12:.LCPI12_0]
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %sel = select i1 %mask, <2 x i64> %op1, <2 x i64> %op2
+  ret <2 x i64> %sel
+}
+
+define void @select_v4i64(ptr %a, ptr %b, i1 %mask) #0 {
+; CHECK-LABEL: select_v4i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w2, #0x1
+; CHECK-NEXT:    ldr q0, [x0]
+; CHECK-NEXT:    csetm x8, ne
+; CHECK-NEXT:    ldr q1, [x0, #16]
+; CHECK-NEXT:    ldr q2, [x1]
+; CHECK-NEXT:    adrp x9, .LCPI13_0
+; CHECK-NEXT:    ldr q3, [x1, #16]
+; CHECK-NEXT:    stp x8, x8, [sp, #-16]!
+; CHECK-NEXT:    .cfi_def_cfa_offset 16
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI13_0]
+; CHECK-NEXT:    ldr q5, [sp]
+; CHECK-NEXT:    eor z4.d, z5.d, z4.d
+; CHECK-NEXT:    and z1.d, z1.d, z5.d
+; CHECK-NEXT:    and z0.d, z0.d, z5.d
+; CHECK-NEXT:    and z2.d, z2.d, z4.d
+; CHECK-NEXT:    and z3.d, z3.d, z4.d
+; CHECK-NEXT:    orr z0.d, z0.d, z2.d
+; CHECK-NEXT:    orr z1.d, z1.d, z3.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    add sp, sp, #16
+; CHECK-NEXT:    ret
+  %op1 = load volatile <4 x i64>, ptr %a
+  %op2 = load volatile <4 x i64>, ptr %b
+  %sel = select i1 %mask, <4 x i64> %op1, <4 x i64> %op2
+  store <4 x i64> %sel, ptr %a
+  ret void
+}
+
+attributes #0 = { "target-features"="+sve" }

diff  --git a/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-vselect.ll b/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-vselect.ll
new file mode 100644
index 000000000000..33fa29f3b7f6
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/sve-streaming-mode-fixed-length-int-vselect.ll
@@ -0,0 +1,344 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -force-streaming-compatible-sve < %s | FileCheck %s
+
+target triple = "aarch64-unknown-linux-gnu"
+
+define <4 x i8> @select_v4i8(<4 x i8> %op1, <4 x i8> %op2, <4 x i1> %mask) #0 {
+; CHECK-LABEL: select_v4i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI0_0
+; CHECK-NEXT:    adrp x9, .LCPI0_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.h, vl4
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x8, :lo12:.LCPI0_0]
+; CHECK-NEXT:    ldr d4, [x9, :lo12:.LCPI0_1]
+; CHECK-NEXT:    lsl z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    asr z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <4 x i1> %mask, <4 x i8> %op1, <4 x i8> %op2
+  ret <4 x i8> %sel
+}
+
+define <8 x i8> @select_v8i8(<8 x i8> %op1, <8 x i8> %op2, <8 x i1> %mask) #0 {
+; CHECK-LABEL: select_v8i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI1_0
+; CHECK-NEXT:    adrp x9, .LCPI1_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.b, vl8
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x8, :lo12:.LCPI1_0]
+; CHECK-NEXT:    ldr d4, [x9, :lo12:.LCPI1_1]
+; CHECK-NEXT:    lsl z2.b, p0/m, z2.b, z3.b
+; CHECK-NEXT:    asr z2.b, p0/m, z2.b, z3.b
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <8 x i1> %mask, <8 x i8> %op1, <8 x i8> %op2
+  ret <8 x i8> %sel
+}
+
+define <16 x i8> @select_v16i8(<16 x i8> %op1, <16 x i8> %op2, <16 x i1> %mask) #0 {
+; CHECK-LABEL: select_v16i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI2_0
+; CHECK-NEXT:    adrp x9, .LCPI2_1
+; CHECK-NEXT:    // kill: def $q2 killed $q2 def $z2
+; CHECK-NEXT:    ptrue p0.b, vl16
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    ldr q3, [x8, :lo12:.LCPI2_0]
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI2_1]
+; CHECK-NEXT:    lsl z2.b, p0/m, z2.b, z3.b
+; CHECK-NEXT:    asr z2.b, p0/m, z2.b, z3.b
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <16 x i1> %mask, <16 x i8> %op1, <16 x i8> %op2
+  ret <16 x i8> %sel
+}
+
+define void @select_v32i8(ptr %a, ptr %b) #0 {
+; CHECK-LABEL: select_v32i8:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ldp q0, q1, [x1]
+; CHECK-NEXT:    adrp x8, .LCPI3_0
+; CHECK-NEXT:    ldp q3, q2, [x0]
+; CHECK-NEXT:    cmeq v6.16b, v3.16b, v0.16b
+; CHECK-NEXT:    ldr q4, [x8, :lo12:.LCPI3_0]
+; CHECK-NEXT:    and z3.d, z3.d, z6.d
+; CHECK-NEXT:    cmeq v5.16b, v2.16b, v1.16b
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    eor z5.d, z5.d, z4.d
+; CHECK-NEXT:    eor z4.d, z6.d, z4.d
+; CHECK-NEXT:    and z1.d, z1.d, z5.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    orr z1.d, z2.d, z1.d
+; CHECK-NEXT:    orr z0.d, z3.d, z0.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    ret
+  %op1 = load <32 x i8>, ptr %a
+  %op2 = load <32 x i8>, ptr %b
+  %mask = icmp eq <32 x i8> %op1, %op2
+  %sel = select <32 x i1> %mask, <32 x i8> %op1, <32 x i8> %op2
+  store <32 x i8> %sel, ptr %a
+  ret void
+}
+
+define <2 x i16> @select_v2i16(<2 x i16> %op1, <2 x i16> %op2, <2 x i1> %mask) #0 {
+; CHECK-LABEL: select_v2i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI4_0
+; CHECK-NEXT:    adrp x9, .LCPI4_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.s, vl2
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x8, :lo12:.LCPI4_0]
+; CHECK-NEXT:    ldr d4, [x9, :lo12:.LCPI4_1]
+; CHECK-NEXT:    lsl z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    asr z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <2 x i1> %mask, <2 x i16> %op1, <2 x i16> %op2
+  ret <2 x i16> %sel
+}
+
+define <4 x i16> @select_v4i16(<4 x i16> %op1, <4 x i16> %op2, <4 x i1> %mask) #0 {
+; CHECK-LABEL: select_v4i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI5_0
+; CHECK-NEXT:    adrp x9, .LCPI5_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.h, vl4
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x8, :lo12:.LCPI5_0]
+; CHECK-NEXT:    ldr d4, [x9, :lo12:.LCPI5_1]
+; CHECK-NEXT:    lsl z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    asr z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <4 x i1> %mask, <4 x i16> %op1, <4 x i16> %op2
+  ret <4 x i16> %sel
+}
+
+define <8 x i16> @select_v8i16(<8 x i16> %op1, <8 x i16> %op2, <8 x i1> %mask) #0 {
+; CHECK-LABEL: select_v8i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI6_0
+; CHECK-NEXT:    adrp x9, .LCPI6_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.h, vl8
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    uunpklo z2.h, z2.b
+; CHECK-NEXT:    ldr q3, [x8, :lo12:.LCPI6_0]
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI6_1]
+; CHECK-NEXT:    lsl z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    asr z2.h, p0/m, z2.h, z3.h
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <8 x i1> %mask, <8 x i16> %op1, <8 x i16> %op2
+  ret <8 x i16> %sel
+}
+
+define void @select_v16i16(ptr %a, ptr %b) #0 {
+; CHECK-LABEL: select_v16i16:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ldp q0, q1, [x1]
+; CHECK-NEXT:    adrp x8, .LCPI7_0
+; CHECK-NEXT:    ldp q3, q2, [x0]
+; CHECK-NEXT:    cmeq v6.8h, v3.8h, v0.8h
+; CHECK-NEXT:    ldr q4, [x8, :lo12:.LCPI7_0]
+; CHECK-NEXT:    and z3.d, z3.d, z6.d
+; CHECK-NEXT:    cmeq v5.8h, v2.8h, v1.8h
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    eor z5.d, z5.d, z4.d
+; CHECK-NEXT:    eor z4.d, z6.d, z4.d
+; CHECK-NEXT:    and z1.d, z1.d, z5.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    orr z1.d, z2.d, z1.d
+; CHECK-NEXT:    orr z0.d, z3.d, z0.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    ret
+  %op1 = load <16 x i16>, ptr %a
+  %op2 = load <16 x i16>, ptr %b
+  %mask = icmp eq <16 x i16> %op1, %op2
+  %sel = select <16 x i1> %mask, <16 x i16> %op1, <16 x i16> %op2
+  store <16 x i16> %sel, ptr %a
+  ret void
+}
+
+define <2 x i32> @select_v2i32(<2 x i32> %op1, <2 x i32> %op2, <2 x i1> %mask) #0 {
+; CHECK-LABEL: select_v2i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI8_0
+; CHECK-NEXT:    adrp x9, .LCPI8_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.s, vl2
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    ldr d3, [x8, :lo12:.LCPI8_0]
+; CHECK-NEXT:    ldr d4, [x9, :lo12:.LCPI8_1]
+; CHECK-NEXT:    lsl z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    asr z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <2 x i1> %mask, <2 x i32> %op1, <2 x i32> %op2
+  ret <2 x i32> %sel
+}
+
+define <4 x i32> @select_v4i32(<4 x i32> %op1, <4 x i32> %op2, <4 x i1> %mask) #0 {
+; CHECK-LABEL: select_v4i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI9_0
+; CHECK-NEXT:    adrp x9, .LCPI9_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.s, vl4
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    uunpklo z2.s, z2.h
+; CHECK-NEXT:    ldr q3, [x8, :lo12:.LCPI9_0]
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI9_1]
+; CHECK-NEXT:    lsl z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    asr z2.s, p0/m, z2.s, z3.s
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <4 x i1> %mask, <4 x i32> %op1, <4 x i32> %op2
+  ret <4 x i32> %sel
+}
+
+define void @select_v8i32(ptr %a, ptr %b) #0 {
+; CHECK-LABEL: select_v8i32:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ldp q0, q1, [x1]
+; CHECK-NEXT:    adrp x8, .LCPI10_0
+; CHECK-NEXT:    ldp q3, q2, [x0]
+; CHECK-NEXT:    cmeq v6.4s, v3.4s, v0.4s
+; CHECK-NEXT:    ldr q4, [x8, :lo12:.LCPI10_0]
+; CHECK-NEXT:    and z3.d, z3.d, z6.d
+; CHECK-NEXT:    cmeq v5.4s, v2.4s, v1.4s
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    eor z5.d, z5.d, z4.d
+; CHECK-NEXT:    eor z4.d, z6.d, z4.d
+; CHECK-NEXT:    and z1.d, z1.d, z5.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    orr z1.d, z2.d, z1.d
+; CHECK-NEXT:    orr z0.d, z3.d, z0.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    ret
+  %op1 = load <8 x i32>, ptr %a
+  %op2 = load <8 x i32>, ptr %b
+  %mask = icmp eq <8 x i32> %op1, %op2
+  %sel = select <8 x i1> %mask, <8 x i32> %op1, <8 x i32> %op2
+  store <8 x i32> %sel, ptr %a
+  ret void
+}
+
+define <1 x i64> @select_v1i64(<1 x i64> %op1, <1 x i64> %op2, <1 x i1> %mask) #0 {
+; CHECK-LABEL: select_v1i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    tst w0, #0x1
+; CHECK-NEXT:    mov x9, #-1
+; CHECK-NEXT:    csetm x8, ne
+; CHECK-NEXT:    // kill: def $d1 killed $d1 def $z1
+; CHECK-NEXT:    // kill: def $d0 killed $d0 def $z0
+; CHECK-NEXT:    fmov d3, x9
+; CHECK-NEXT:    fmov d2, x8
+; CHECK-NEXT:    eor z3.d, z2.d, z3.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $d0 killed $d0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <1 x i1> %mask, <1 x i64> %op1, <1 x i64> %op2
+  ret <1 x i64> %sel
+}
+
+define <2 x i64> @select_v2i64(<2 x i64> %op1, <2 x i64> %op2, <2 x i1> %mask) #0 {
+; CHECK-LABEL: select_v2i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    adrp x8, .LCPI12_0
+; CHECK-NEXT:    adrp x9, .LCPI12_1
+; CHECK-NEXT:    // kill: def $d2 killed $d2 def $z2
+; CHECK-NEXT:    ptrue p0.d, vl2
+; CHECK-NEXT:    // kill: def $q1 killed $q1 def $z1
+; CHECK-NEXT:    // kill: def $q0 killed $q0 def $z0
+; CHECK-NEXT:    uunpklo z2.d, z2.s
+; CHECK-NEXT:    ldr q3, [x8, :lo12:.LCPI12_0]
+; CHECK-NEXT:    ldr q4, [x9, :lo12:.LCPI12_1]
+; CHECK-NEXT:    lsl z2.d, p0/m, z2.d, z3.d
+; CHECK-NEXT:    asr z2.d, p0/m, z2.d, z3.d
+; CHECK-NEXT:    eor z3.d, z2.d, z4.d
+; CHECK-NEXT:    and z0.d, z0.d, z2.d
+; CHECK-NEXT:    and z1.d, z1.d, z3.d
+; CHECK-NEXT:    orr z0.d, z0.d, z1.d
+; CHECK-NEXT:    // kill: def $q0 killed $q0 killed $z0
+; CHECK-NEXT:    ret
+  %sel = select <2 x i1> %mask, <2 x i64> %op1, <2 x i64> %op2
+  ret <2 x i64> %sel
+}
+
+define void @select_v4i64(ptr %a, ptr %b) #0 {
+; CHECK-LABEL: select_v4i64:
+; CHECK:       // %bb.0:
+; CHECK-NEXT:    ldp q0, q1, [x1]
+; CHECK-NEXT:    adrp x8, .LCPI13_0
+; CHECK-NEXT:    ldp q3, q2, [x0]
+; CHECK-NEXT:    cmeq v6.2d, v3.2d, v0.2d
+; CHECK-NEXT:    ldr q4, [x8, :lo12:.LCPI13_0]
+; CHECK-NEXT:    and z3.d, z3.d, z6.d
+; CHECK-NEXT:    cmeq v5.2d, v2.2d, v1.2d
+; CHECK-NEXT:    and z2.d, z2.d, z5.d
+; CHECK-NEXT:    eor z5.d, z5.d, z4.d
+; CHECK-NEXT:    eor z4.d, z6.d, z4.d
+; CHECK-NEXT:    and z1.d, z1.d, z5.d
+; CHECK-NEXT:    and z0.d, z0.d, z4.d
+; CHECK-NEXT:    orr z1.d, z2.d, z1.d
+; CHECK-NEXT:    orr z0.d, z3.d, z0.d
+; CHECK-NEXT:    stp q0, q1, [x0]
+; CHECK-NEXT:    ret
+  %op1 = load <4 x i64>, ptr %a
+  %op2 = load <4 x i64>, ptr %b
+  %mask = icmp eq <4 x i64> %op1, %op2
+  %sel = select <4 x i1> %mask, <4 x i64> %op1, <4 x i64> %op2
+  store <4 x i64> %sel, ptr %a
+  ret void
+}
+
+attributes #0 = { "target-features"="+sve" uwtable }


        


More information about the llvm-commits mailing list