[llvm] r333843 - [X86][AVX512] Cleanup intrinsics tests
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Sun Jun 3 07:56:04 PDT 2018
Author: rksimon
Date: Sun Jun 3 07:56:04 2018
New Revision: 333843
URL: http://llvm.org/viewvc/llvm-project?rev=333843&view=rev
Log:
[X86][AVX512] Cleanup intrinsics tests
Ensure we test on 32-bit and 64-bit targets, and strip -mcpu usage.
Part of ongoing work to ensure we test all intrinsic style tests on 32 and 64 bit targets where possible.
Modified:
llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512-gfni-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512-intrinsics-canonical.ll
llvm/trunk/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512-vpclmulqdq.ll
llvm/trunk/test/CodeGen/X86/avx512bw-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512bw-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512bw-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512cd-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512cd-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512cd-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512cdvl-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512cdvl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512dq-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512dq-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512er-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512ifma-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512ifma-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512ifma-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512ifmavl-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512ifmavl-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512ifmavl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vbmi-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512vbmi-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512vbmi-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vbmi2-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512vbmi2-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vbmi2vl-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512vbmi2vl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vbmivl-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512vbmivl-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512vbmivl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vl-intrinsics-canonical.ll
llvm/trunk/test/CodeGen/X86/avx512vl-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll
llvm/trunk/test/CodeGen/X86/avx512vl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vl-vpclmulqdq.ll
llvm/trunk/test/CodeGen/X86/avx512vl_vnni-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vlcd-intrinsics-fast-isel.ll
llvm/trunk/test/CodeGen/X86/avx512vnni-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vpopcntdq-intrinsics.ll
Modified: llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll?rev=333843&r1=333842&r2=333843&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512-fma-intrinsics.ll Sun Jun 3 07:56:04 2018
@@ -1,104 +1,152 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512f | FileCheck %s
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512f --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64
declare <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float>, <16 x float>, <16 x float>, i32)
declare <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double>, <8 x double>, <8 x double>, i32)
define <16 x float> @test_x86_vfnmadd_ps_z(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_x86_vfnmadd_ps_z:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0xac,0xc2]
+; CHECK-NEXT: # zmm0 = -(zmm1 * zmm0) + zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 -1, i32 4) nounwind
ret <16 x float> %res
}
declare <16 x float> @llvm.x86.avx512.mask.vfnmadd.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32) nounwind readnone
define <16 x float> @test_mask_vfnmadd_ps(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
-; CHECK-LABEL: test_mask_vfnmadd_ps:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_vfnmadd_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
+; X86-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_vfnmadd_ps:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
+; X64-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask, i32 4) nounwind
ret <16 x float> %res
}
define <8 x double> @test_x86_vfnmadd_pd_z(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_x86_vfnmadd_pd_z:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xac,0xc2]
+; CHECK-NEXT: # zmm0 = -(zmm1 * zmm0) + zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 -1, i32 4) nounwind
ret <8 x double> %res
}
declare <8 x double> @llvm.x86.avx512.mask.vfnmadd.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32) nounwind readnone
define <8 x double> @test_mask_vfnmadd_pd(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_vfnmadd_pd:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_vfnmadd_pd:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
+; X86-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_vfnmadd_pd:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
+; X64-NEXT: # zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 4) nounwind
ret <8 x double> %res
}
define <16 x float> @test_x86_vfnmsubps_z(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_x86_vfnmsubps_z:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0xae,0xc2]
+; CHECK-NEXT: # zmm0 = -(zmm1 * zmm0) - zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmsub.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 -1, i32 4) nounwind
ret <16 x float> %res
}
declare <16 x float> @llvm.x86.avx512.mask.vfnmsub.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32) nounwind readnone
define <16 x float> @test_mask_vfnmsub_ps(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
-; CHECK-LABEL: test_mask_vfnmsub_ps:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_vfnmsub_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
+; X86-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_vfnmsub_ps:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
+; X64-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmsub.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask, i32 4) nounwind
ret <16 x float> %res
}
define <8 x double> @test_x86_vfnmsubpd_z(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_x86_vfnmsubpd_z:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xae,0xc2]
+; CHECK-NEXT: # zmm0 = -(zmm1 * zmm0) - zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 -1, i32 4) nounwind
ret <8 x double> %res
}
declare <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32) nounwind readnone
define <8 x double> @test_mask_vfnmsub_pd(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_vfnmsub_pd:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_vfnmsub_pd:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
+; X86-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_vfnmsub_pd:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
+; X64-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 4) nounwind
ret <8 x double> %res
}
define <16 x float> @test_x86_vfmaddsubps_z(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_x86_vfmaddsubps_z:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0xa6,0xc2]
+; CHECK-NEXT: # zmm0 = (zmm1 * zmm0) +/- zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 4) nounwind
ret <16 x float> %res
}
define <16 x float> @test_mask_fmaddsub_ps(<16 x float> %a, <16 x float> %b, <16 x float> %c, i16 %mask) {
-; CHECK-LABEL: test_mask_fmaddsub_ps:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_fmaddsub_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
+; X86-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_fmaddsub_ps:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
+; X64-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %c, i32 4)
%bc = bitcast i16 %mask to <16 x i1>
%sel = select <16 x i1> %bc, <16 x float> %res, <16 x float> %a
@@ -109,20 +157,30 @@ declare <16 x float> @llvm.x86.avx512.vf
define <8 x double> @test_x86_vfmaddsubpd_z(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_x86_vfmaddsubpd_z:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xa6,0xc2]
+; CHECK-NEXT: # zmm0 = (zmm1 * zmm0) +/- zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 4) nounwind
ret <8 x double> %res
}
declare <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double>, <8 x double>, <8 x double>, i32) nounwind readnone
define <8 x double> @test_mask_vfmaddsub_pd(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_vfmaddsub_pd:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_vfmaddsub_pd:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
+; X86-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_vfmaddsub_pd:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
+; X64-NEXT: # zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 4) nounwind
%bc = bitcast i8 %mask to <8 x i1>
%sel = select <8 x i1> %bc, <8 x double> %res, <8 x double> %a0
@@ -130,14 +188,26 @@ define <8 x double> @test_mask_vfmaddsub
}
define <8 x double>@test_int_x86_avx512_mask_vfmaddsub_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask_vfmaddsub_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask_vfmaddsub_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd8]
+; X86-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x96,0xd9]
+; X86-NEXT: # zmm3 = (zmm3 * zmm1) +/- zmm2
+; X86-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa6,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask_vfmaddsub_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd8]
+; X64-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x96,0xd9]
+; X64-NEXT: # zmm3 = (zmm3 * zmm1) +/- zmm2
+; X64-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa6,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i32 4)
%bc = bitcast i8 %x3 to <8 x i1>
%sel = select <8 x i1> %bc, <8 x double> %res, <8 x double> %x0
@@ -149,14 +219,26 @@ define <8 x double>@test_int_x86_avx512_
declare <8 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32)
define <8 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfmaddsub_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm2, %zmm3
-; CHECK-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfmaddsub_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X86-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd9]
+; X86-NEXT: # zmm3 = (zmm0 * zmm1) +/- zmm3
+; X86-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa6,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfmaddsub_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X64-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd9]
+; X64-NEXT: # zmm3 = (zmm0 * zmm1) +/- zmm3
+; X64-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa6,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -166,14 +248,26 @@ define <8 x double>@test_int_x86_avx512_
declare <8 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32)
define <8 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm1, %zmm3
-; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm0, %zmm3 {%k1} {z}
-; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm1, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd9]
+; X86-NEXT: vfmaddsub213pd %zmm2, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0xa6,0xda]
+; X86-NEXT: # zmm3 = (zmm0 * zmm3) +/- zmm2
+; X86-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa6,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm1, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd9]
+; X64-NEXT: vfmaddsub213pd %zmm2, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0xa6,0xda]
+; X64-NEXT: # zmm3 = (zmm0 * zmm3) +/- zmm2
+; X64-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa6,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -181,14 +275,25 @@ define <8 x double>@test_int_x86_avx512_
}
define <16 x float>@test_int_x86_avx512_mask_vfmaddsub_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask_vfmaddsub_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask_vfmaddsub_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; X86-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x96,0xd9]
+; X86-NEXT: # zmm3 = (zmm3 * zmm1) +/- zmm2
+; X86-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa6,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask_vfmaddsub_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; X64-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x96,0xd9]
+; X64-NEXT: # zmm3 = (zmm3 * zmm1) +/- zmm2
+; X64-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa6,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i32 4)
%bc = bitcast i16 %x3 to <16 x i1>
%sel = select <16 x i1> %bc, <16 x float> %res, <16 x float> %x0
@@ -200,14 +305,25 @@ define <16 x float>@test_int_x86_avx512_
declare <16 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
define <16 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfmaddsub_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfmaddsub_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X86-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd9]
+; X86-NEXT: # zmm3 = (zmm0 * zmm1) +/- zmm3
+; X86-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa6,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfmaddsub_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X64-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd9]
+; X64-NEXT: # zmm3 = (zmm0 * zmm1) +/- zmm3
+; X64-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa6,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
@@ -217,14 +333,25 @@ define <16 x float>@test_int_x86_avx512_
declare <16 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
define <16 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm1, %zmm3
-; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm0, %zmm3 {%k1} {z}
-; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm1, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; X86-NEXT: vfmaddsub213ps %zmm2, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xa6,0xda]
+; X86-NEXT: # zmm3 = (zmm0 * zmm3) +/- zmm2
+; X86-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa6,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm1, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; X64-NEXT: vfmaddsub213ps %zmm2, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xa6,0xda]
+; X64-NEXT: # zmm3 = (zmm0 * zmm3) +/- zmm2
+; X64-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa6,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
@@ -234,14 +361,26 @@ define <16 x float>@test_int_x86_avx512_
declare <8 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32)
define <8 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsubadd_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm2, %zmm3
-; CHECK-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfmsubadd_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X86-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd9]
+; X86-NEXT: # zmm3 = (zmm0 * zmm1) -/+ zmm3
+; X86-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa7,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfmsubadd_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X64-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd9]
+; X64-NEXT: # zmm3 = (zmm0 * zmm1) -/+ zmm3
+; X64-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa7,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -251,14 +390,25 @@ define <8 x double>@test_int_x86_avx512_
declare <16 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
define <16 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsubadd_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfmsubadd_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X86-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd9]
+; X86-NEXT: # zmm3 = (zmm0 * zmm1) -/+ zmm3
+; X86-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa7,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfmsubadd_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X64-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd9]
+; X64-NEXT: # zmm3 = (zmm0 * zmm1) -/+ zmm3
+; X64-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa7,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
@@ -266,11 +416,17 @@ define <16 x float>@test_int_x86_avx512_
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrb_rne(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_rne:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_ps_rrb_rne:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x19,0x98,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_ps_rrb_rne:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x19,0x98,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 0) nounwind
%bc = bitcast i16 %mask to <16 x i1>
%sel = select <16 x i1> %bc, <16 x float> %res, <16 x float> %a0
@@ -278,11 +434,17 @@ define <16 x float> @test_mask_round_vfm
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrb_rtn(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_rtn:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132ps {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_ps_rrb_rtn:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd132ps {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x39,0x98,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_ps_rrb_rtn:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132ps {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x39,0x98,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 1) nounwind
%bc = bitcast i16 %mask to <16 x i1>
%sel = select <16 x i1> %bc, <16 x float> %res, <16 x float> %a0
@@ -290,11 +452,17 @@ define <16 x float> @test_mask_round_vfm
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrb_rtp(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_rtp:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132ps {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_ps_rrb_rtp:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd132ps {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x59,0x98,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_ps_rrb_rtp:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132ps {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x59,0x98,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 2) nounwind
%bc = bitcast i16 %mask to <16 x i1>
%sel = select <16 x i1> %bc, <16 x float> %res, <16 x float> %a0
@@ -302,11 +470,17 @@ define <16 x float> @test_mask_round_vfm
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrb_rtz(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_rtz:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132ps {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_ps_rrb_rtz:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd132ps {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x79,0x98,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_ps_rrb_rtz:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132ps {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x79,0x98,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 3) nounwind
%bc = bitcast i16 %mask to <16 x i1>
%sel = select <16 x i1> %bc, <16 x float> %res, <16 x float> %a0
@@ -314,11 +488,19 @@ define <16 x float> @test_mask_round_vfm
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrb_current(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i16 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrb_current:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_ps_rrb_current:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
+; X86-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_ps_rrb_current:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
+; X64-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 4) nounwind
%bc = bitcast i16 %mask to <16 x i1>
%sel = select <16 x i1> %bc, <16 x float> %res, <16 x float> %a0
@@ -327,45 +509,46 @@ define <16 x float> @test_mask_round_vfm
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_rne(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_rne:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x18,0xa8,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 0) nounwind
ret <16 x float> %res
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_rtn(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_rtn:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213ps {rd-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213ps {rd-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x38,0xa8,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 1) nounwind
ret <16 x float> %res
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_rtp(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_rtp:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213ps {ru-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213ps {ru-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x58,0xa8,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 2) nounwind
ret <16 x float> %res
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_rtz(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_rtz:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213ps {rz-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213ps {rz-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x78,0xa8,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 3) nounwind
ret <16 x float> %res
}
define <16 x float> @test_mask_round_vfmadd512_ps_rrbz_current(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_ps_rrbz_current:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0x75,0x48,0xa8,0xc2]
+; CHECK-NEXT: # zmm0 = (zmm1 * zmm0) + zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %a0, <16 x float> %a1, <16 x float> %a2, i32 4) nounwind
ret <16 x float> %res
}
@@ -373,14 +556,26 @@ define <16 x float> @test_mask_round_vfm
declare <8 x double> @llvm.x86.avx512.mask3.vfmsub.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32)
define <8 x double>@test_int_x86_avx512_mask3_vfmsub_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsub_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm2, %zmm3
-; CHECK-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfmsub_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X86-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd9]
+; X86-NEXT: # zmm3 = (zmm0 * zmm1) - zmm3
+; X86-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xaa,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfmsub_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X64-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd9]
+; X64-NEXT: # zmm3 = (zmm0 * zmm1) - zmm3
+; X64-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xaa,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask3.vfmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -390,14 +585,25 @@ define <8 x double>@test_int_x86_avx512_
declare <16 x float> @llvm.x86.avx512.mask3.vfmsub.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
define <16 x float>@test_int_x86_avx512_mask3_vfmsub_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsub_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfmsub_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X86-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd9]
+; X86-NEXT: # zmm3 = (zmm0 * zmm1) - zmm3
+; X86-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xaa,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfmsub_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X64-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd9]
+; X64-NEXT: # zmm3 = (zmm0 * zmm1) - zmm3
+; X64-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xaa,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask3.vfmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
@@ -405,11 +611,18 @@ define <16 x float>@test_int_x86_avx512_
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrb_rne(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_rne:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_pd_rrb_rne:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x19,0x98,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_pd_rrb_rne:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x19,0x98,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 0) nounwind
%bc = bitcast i8 %mask to <8 x i1>
%sel = select <8 x i1> %bc, <8 x double> %res, <8 x double> %a0
@@ -417,11 +630,18 @@ define <8 x double> @test_mask_round_vfm
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrb_rtn(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_rtn:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132pd {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_pd_rrb_rtn:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd132pd {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x39,0x98,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_pd_rrb_rtn:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132pd {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x39,0x98,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 1) nounwind
%bc = bitcast i8 %mask to <8 x i1>
%sel = select <8 x i1> %bc, <8 x double> %res, <8 x double> %a0
@@ -429,11 +649,18 @@ define <8 x double> @test_mask_round_vfm
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrb_rtp(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_rtp:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132pd {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_pd_rrb_rtp:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd132pd {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x59,0x98,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_pd_rrb_rtp:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132pd {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x59,0x98,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 2) nounwind
%bc = bitcast i8 %mask to <8 x i1>
%sel = select <8 x i1> %bc, <8 x double> %res, <8 x double> %a0
@@ -441,11 +668,18 @@ define <8 x double> @test_mask_round_vfm
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrb_rtz(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_rtz:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132pd {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_pd_rrb_rtz:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd132pd {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x79,0x98,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_pd_rrb_rtz:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132pd {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x79,0x98,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 3) nounwind
%bc = bitcast i8 %mask to <8 x i1>
%sel = select <8 x i1> %bc, <8 x double> %res, <8 x double> %a0
@@ -453,11 +687,20 @@ define <8 x double> @test_mask_round_vfm
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrb_current(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrb_current:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfmadd512_pd_rrb_current:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
+; X86-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfmadd512_pd_rrb_current:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
+; X64-NEXT: # zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 4) nounwind
%bc = bitcast i8 %mask to <8 x i1>
%sel = select <8 x i1> %bc, <8 x double> %res, <8 x double> %a0
@@ -466,58 +709,71 @@ define <8 x double> @test_mask_round_vfm
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_rne(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_rne:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x18,0xa8,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 0) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_rtn(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_rtn:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213pd {rd-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213pd {rd-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x38,0xa8,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 1) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_rtp(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_rtp:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213pd {ru-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213pd {ru-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x58,0xa8,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 2) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_rtz(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_rtz:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213pd {rz-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213pd {rz-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x78,0xa8,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 3) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfmadd512_pd_rrbz_current(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfmadd512_pd_rrbz_current:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xa8,0xc2]
+; CHECK-NEXT: # zmm0 = (zmm1 * zmm0) + zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i32 4) nounwind
ret <8 x double> %res
}
define <8 x double>@test_int_x86_avx512_mask_vfmadd_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm3 {%k1}
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask_vfmadd_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd8]
+; X86-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x98,0xd9]
+; X86-NEXT: # zmm3 = (zmm3 * zmm1) + zmm2
+; X86-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa8,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask_vfmadd_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd8]
+; X64-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x98,0xd9]
+; X64-NEXT: # zmm3 = (zmm3 * zmm1) + zmm2
+; X64-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa8,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i32 4)
%bc = bitcast i8 %x3 to <8 x i1>
%sel = select <8 x i1> %bc, <8 x double> %res, <8 x double> %x0
@@ -529,14 +785,26 @@ define <8 x double>@test_int_x86_avx512_
declare <8 x double> @llvm.x86.avx512.mask3.vfmadd.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32)
define <8 x double>@test_int_x86_avx512_mask3_vfmadd_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfmadd_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm2, %zmm3
-; CHECK-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfmadd_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X86-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd9]
+; X86-NEXT: # zmm3 = (zmm0 * zmm1) + zmm3
+; X86-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa8,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfmadd_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X64-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd9]
+; X64-NEXT: # zmm3 = (zmm0 * zmm1) + zmm3
+; X64-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa8,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask3.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -546,14 +814,26 @@ define <8 x double>@test_int_x86_avx512_
declare <8 x double> @llvm.x86.avx512.maskz.vfmadd.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32)
define <8 x double>@test_int_x86_avx512_maskz_vfmadd_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm1, %zmm3
-; CHECK-NEXT: vfmadd213pd %zmm2, %zmm0, %zmm3 {%k1} {z}
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm1, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd9]
+; X86-NEXT: vfmadd213pd %zmm2, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0xa8,0xda]
+; X86-NEXT: # zmm3 = (zmm0 * zmm3) + zmm2
+; X86-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa8,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm1, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd9]
+; X64-NEXT: vfmadd213pd %zmm2, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf2,0xfd,0xc9,0xa8,0xda]
+; X64-NEXT: # zmm3 = (zmm0 * zmm3) + zmm2
+; X64-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xa8,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.maskz.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.maskz.vfmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -561,14 +841,25 @@ define <8 x double>@test_int_x86_avx512_
}
define <16 x float>@test_int_x86_avx512_mask_vfmadd_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm3 {%k1}
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask_vfmadd_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; X86-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x98,0xd9]
+; X86-NEXT: # zmm3 = (zmm3 * zmm1) + zmm2
+; X86-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa8,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask_vfmadd_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; X64-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x98,0xd9]
+; X64-NEXT: # zmm3 = (zmm3 * zmm1) + zmm2
+; X64-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa8,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i32 4)
%bc = bitcast i16 %x3 to <16 x i1>
%sel = select <16 x i1> %bc, <16 x float> %res, <16 x float> %x0
@@ -580,14 +871,25 @@ define <16 x float>@test_int_x86_avx512_
declare <16 x float> @llvm.x86.avx512.mask3.vfmadd.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
define <16 x float>@test_int_x86_avx512_mask3_vfmadd_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfmadd_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfmadd_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X86-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd9]
+; X86-NEXT: # zmm3 = (zmm0 * zmm1) + zmm3
+; X86-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa8,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfmadd_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X64-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd9]
+; X64-NEXT: # zmm3 = (zmm0 * zmm1) + zmm3
+; X64-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa8,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask3.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
@@ -597,14 +899,25 @@ define <16 x float>@test_int_x86_avx512_
declare <16 x float> @llvm.x86.avx512.maskz.vfmadd.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
define <16 x float>@test_int_x86_avx512_maskz_vfmadd_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm1, %zmm3
-; CHECK-NEXT: vfmadd213ps %zmm2, %zmm0, %zmm3 {%k1} {z}
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm1, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; X86-NEXT: vfmadd213ps %zmm2, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xa8,0xda]
+; X86-NEXT: # zmm3 = (zmm0 * zmm3) + zmm2
+; X86-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa8,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm1, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; X64-NEXT: vfmadd213ps %zmm2, %zmm0, %zmm3 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xa8,0xda]
+; X64-NEXT: # zmm3 = (zmm0 * zmm3) + zmm2
+; X64-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xa8,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.maskz.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.maskz.vfmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
@@ -613,109 +926,159 @@ define <16 x float>@test_int_x86_avx512_
define <8 x double> @test_mask_round_vfnmsub512_pd_rrb_rne(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_rne:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfnmsub512_pd_rrb_rne:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x19,0x9e,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfnmsub512_pd_rrb_rne:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x19,0x9e,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 0) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrb_rtn(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtn:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtn:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub132pd {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x39,0x9e,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtn:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132pd {rd-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x39,0x9e,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 1) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrb_rtp(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtp:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtp:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub132pd {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x59,0x9e,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtp:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132pd {ru-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x59,0x9e,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 2) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrb_rtz(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtz:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtz:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub132pd {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x79,0x9e,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfnmsub512_pd_rrb_rtz:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132pd {rz-sae}, %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x79,0x9e,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 3) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrb_current(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask) {
-; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrb_current:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1}
-; CHECK-NEXT: retq
+; X86-LABEL: test_mask_round_vfnmsub512_pd_rrb_current:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
+; X86-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_mask_round_vfnmsub512_pd_rrb_current:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
+; X64-NEXT: # zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 %mask, i32 4) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_rne(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_rne:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x18,0xae,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 -1, i32 0) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_rtn(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_rtn:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmsub213pd {rd-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmsub213pd {rd-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x38,0xae,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 -1, i32 1) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_rtp(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_rtp:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmsub213pd {ru-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmsub213pd {ru-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x58,0xae,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 -1, i32 2) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_rtz(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_rtz:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmsub213pd {rz-sae}, %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmsub213pd {rz-sae}, %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x78,0xae,0xc2]
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 -1, i32 3) nounwind
ret <8 x double> %res
}
define <8 x double> @test_mask_round_vfnmsub512_pd_rrbz_current(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2) {
; CHECK-LABEL: test_mask_round_vfnmsub512_pd_rrbz_current:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0
-; CHECK-NEXT: retq
+; CHECK: # %bb.0:
+; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 # encoding: [0x62,0xf2,0xf5,0x48,0xae,0xc2]
+; CHECK-NEXT: # zmm0 = -(zmm1 * zmm0) - zmm2
+; CHECK-NEXT: ret{{[l|q]}} # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %a0, <8 x double> %a1, <8 x double> %a2, i8 -1, i32 4) nounwind
ret <8 x double> %res
}
define <8 x double>@test_int_x86_avx512_mask_vfnmsub_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask_vfnmsub_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm3 {%k1}
-; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask_vfnmsub_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd8]
+; X86-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xd9]
+; X86-NEXT: # zmm3 = -(zmm3 * zmm1) - zmm2
+; X86-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xae,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask_vfnmsub_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd8]
+; X64-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9e,0xd9]
+; X64-NEXT: # zmm3 = -(zmm3 * zmm1) - zmm2
+; X64-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xae,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.mask.vfnmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -725,14 +1088,26 @@ define <8 x double>@test_int_x86_avx512_
declare <8 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.512(<8 x double>, <8 x double>, <8 x double>, i8, i32)
define <8 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfnmsub_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm2, %zmm3
-; CHECK-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfnmsub_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X86-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd9]
+; X86-NEXT: # zmm3 = -(zmm0 * zmm1) - zmm3
+; X86-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xae,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfnmsub_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm2, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xda]
+; X64-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd9]
+; X64-NEXT: # zmm3 = -(zmm0 * zmm1) - zmm3
+; X64-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xae,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -740,14 +1115,25 @@ define <8 x double>@test_int_x86_avx512_
}
define <16 x float>@test_int_x86_avx512_mask_vfnmsub_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask_vfnmsub_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm3 {%k1}
-; CHECK-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask_vfnmsub_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; X86-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xd9]
+; X86-NEXT: # zmm3 = -(zmm3 * zmm1) - zmm2
+; X86-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xae,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask_vfnmsub_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; X64-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xd9]
+; X64-NEXT: # zmm3 = -(zmm3 * zmm1) - zmm2
+; X64-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xae,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.mask.vfnmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
@@ -757,14 +1143,25 @@ define <16 x float>@test_int_x86_avx512_
declare <16 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.512(<16 x float>, <16 x float>, <16 x float>, i16, i32)
define <16 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask3_vfnmsub_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm3 {%k1}
-; CHECK-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask3_vfnmsub_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X86-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd9]
+; X86-NEXT: # zmm3 = -(zmm0 * zmm1) - zmm3
+; X86-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xae,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask3_vfnmsub_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm2, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; X64-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm3 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd9]
+; X64-NEXT: # zmm3 = -(zmm0 * zmm1) - zmm3
+; X64-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xae,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
@@ -772,14 +1169,26 @@ define <16 x float>@test_int_x86_avx512_
}
define <8 x double>@test_int_x86_avx512_mask_vfnmadd_pd_512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask_vfnmadd_pd_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovapd %zmm0, %zmm3
-; CHECK-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm3 {%k1}
-; CHECK-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddpd %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask_vfnmadd_pd_512:
+; X86: # %bb.0:
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax # encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 # encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vmovapd %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd8]
+; X86-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9c,0xd9]
+; X86-NEXT: # zmm3 = -(zmm3 * zmm1) + zmm2
+; X86-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xac,0xca]
+; X86-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask_vfnmadd_pd_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovapd %zmm0, %zmm3 # encoding: [0x62,0xf1,0xfd,0x48,0x28,0xd8]
+; X64-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0xed,0x49,0x9c,0xd9]
+; X64-NEXT: # zmm3 = -(zmm3 * zmm1) + zmm2
+; X64-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0xfd,0x18,0xac,0xca]
+; X64-NEXT: vaddpd %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0xe5,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <8 x double> @llvm.x86.avx512.mask.vfnmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 %x3, i32 4)
%res1 = call <8 x double> @llvm.x86.avx512.mask.vfnmadd.pd.512(<8 x double> %x0, <8 x double> %x1, <8 x double> %x2, i8 -1, i32 0)
%res2 = fadd <8 x double> %res, %res1
@@ -787,14 +1196,25 @@ define <8 x double>@test_int_x86_avx512_
}
define <16 x float>@test_int_x86_avx512_mask_vfnmadd_ps_512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3){
-; CHECK-LABEL: test_int_x86_avx512_mask_vfnmadd_ps_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm3 {%k1}
-; CHECK-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1
-; CHECK-NEXT: vaddps %zmm1, %zmm3, %zmm0
-; CHECK-NEXT: retq
+; X86-LABEL: test_int_x86_avx512_mask_vfnmadd_ps_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vmovaps %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; X86-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xd9]
+; X86-NEXT: # zmm3 = -(zmm3 * zmm1) + zmm2
+; X86-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xac,0xca]
+; X86-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_int_x86_avx512_mask_vfnmadd_ps_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovw %edi, %k1 # encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vmovaps %zmm0, %zmm3 # encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; X64-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm3 {%k1} # encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xd9]
+; X64-NEXT: # zmm3 = -(zmm3 * zmm1) + zmm2
+; X64-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm0, %zmm1 # encoding: [0x62,0xf2,0x7d,0x18,0xac,0xca]
+; X64-NEXT: vaddps %zmm1, %zmm3, %zmm0 # encoding: [0x62,0xf1,0x64,0x48,0x58,0xc1]
+; X64-NEXT: retq # encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.vfnmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 %x3, i32 4)
%res1 = call <16 x float> @llvm.x86.avx512.mask.vfnmadd.ps.512(<16 x float> %x0, <16 x float> %x1, <16 x float> %x2, i16 -1, i32 0)
%res2 = fadd <16 x float> %res, %res1
Modified: llvm/trunk/test/CodeGen/X86/avx512-gfni-intrinsics.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512-gfni-intrinsics.ll?rev=333843&r1=333842&r2=333843&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-gfni-intrinsics.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512-gfni-intrinsics.ll Sun Jun 3 07:56:04 2018
@@ -1,17 +1,28 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512vl,+gfni,+avx512bw --show-mc-encoding | FileCheck %s
+; RUN: llc < %s -mtriple=i686-unknown-unknown -mattr=+avx512vl,+gfni,+avx512bw --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86
+; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512vl,+gfni,+avx512bw --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64
declare <16 x i8> @llvm.x86.vgf2p8affineinvqb.128(<16 x i8>, <16 x i8>, i8)
define <16 x i8> @test_vgf2p8affineinvqb_128(<16 x i8> %src1, <16 x i8> %src2, <16 x i8> %passthru, i16 %mask) {
-; CHECK-LABEL: test_vgf2p8affineinvqb_128:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm3 ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xcf,0xd9,0x03]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0x89,0xcf,0xe1,0x03]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0xcf,0xd1,0x03]
-; CHECK-NEXT: vpxor %xmm3, %xmm2, %xmm0 ## EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
-; CHECK-NEXT: vpxor %xmm0, %xmm4, %xmm0 ## EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8affineinvqb_128:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xcf,0xd9,0x03]
+; X86-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xcf,0xe1,0x03]
+; X86-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xcf,0xd1,0x03]
+; X86-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
+; X86-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8affineinvqb_128:
+; X64: # %bb.0:
+; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xcf,0xd9,0x03]
+; X64-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xcf,0xe1,0x03]
+; X64-NEXT: vgf2p8affineinvqb $3, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xcf,0xd1,0x03]
+; X64-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
+; X64-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i16 %mask to <16 x i1>
%2 = call <16 x i8> @llvm.x86.vgf2p8affineinvqb.128(<16 x i8> %src1, <16 x i8> %src2, i8 3)
%3 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> zeroinitializer
@@ -23,15 +34,25 @@ define <16 x i8> @test_vgf2p8affineinvqb
declare <32 x i8> @llvm.x86.vgf2p8affineinvqb.256(<32 x i8>, <32 x i8>, i8)
define <32 x i8> @test_vgf2p8affineinvqb_256(<32 x i8> %src1, <32 x i8> %src2, <32 x i8> %passthru, i32 %mask) {
-; CHECK-LABEL: test_vgf2p8affineinvqb_256:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm3 ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xcf,0xd9,0x03]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0xcf,0xe1,0x03]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0xcf,0xd1,0x03]
-; CHECK-NEXT: vpxor %ymm3, %ymm2, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
-; CHECK-NEXT: vpxor %ymm0, %ymm4, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8affineinvqb_256:
+; X86: # %bb.0:
+; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xcf,0xd9,0x03]
+; X86-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xcf,0xe1,0x03]
+; X86-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xcf,0xd1,0x03]
+; X86-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
+; X86-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8affineinvqb_256:
+; X64: # %bb.0:
+; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xcf,0xd9,0x03]
+; X64-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xcf,0xe1,0x03]
+; X64-NEXT: vgf2p8affineinvqb $3, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xcf,0xd1,0x03]
+; X64-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
+; X64-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i32 %mask to <32 x i1>
%2 = call <32 x i8> @llvm.x86.vgf2p8affineinvqb.256(<32 x i8> %src1, <32 x i8> %src2, i8 3)
%3 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> zeroinitializer
@@ -43,15 +64,25 @@ define <32 x i8> @test_vgf2p8affineinvqb
declare <64 x i8> @llvm.x86.vgf2p8affineinvqb.512(<64 x i8>, <64 x i8>, i8)
define <64 x i8> @test_vgf2p8affineinvqb_512(<64 x i8> %src1, <64 x i8> %src2, <64 x i8> %passthru, i64 %mask) {
-; CHECK-LABEL: test_vgf2p8affineinvqb_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovq %rdi, %k1 ## encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm3 ## encoding: [0x62,0xf3,0xfd,0x48,0xcf,0xd9,0x03]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xc9,0xcf,0xe1,0x03]
-; CHECK-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x49,0xcf,0xd1,0x03]
-; CHECK-NEXT: vpxorq %zmm3, %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
-; CHECK-NEXT: vpxorq %zmm0, %zmm4, %zmm0 ## encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8affineinvqb_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0xcf,0xd9,0x03]
+; X86-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xcf,0xe1,0x03]
+; X86-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xcf,0xd1,0x03]
+; X86-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
+; X86-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8affineinvqb_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0xcf,0xd9,0x03]
+; X64-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xcf,0xe1,0x03]
+; X64-NEXT: vgf2p8affineinvqb $3, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xcf,0xd1,0x03]
+; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
+; X64-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i64 %mask to <64 x i1>
%2 = call <64 x i8> @llvm.x86.vgf2p8affineinvqb.512(<64 x i8> %src1, <64 x i8> %src2, i8 3)
%3 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> zeroinitializer
@@ -63,15 +94,25 @@ define <64 x i8> @test_vgf2p8affineinvqb
declare <16 x i8> @llvm.x86.vgf2p8affineqb.128(<16 x i8>, <16 x i8>, i8)
define <16 x i8> @test_vgf2p8affineqb_128(<16 x i8> %src1, <16 x i8> %src2, <16 x i8> %passthru, i16 %mask) {
-; CHECK-LABEL: test_vgf2p8affineqb_128:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm3 ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xce,0xd9,0x03]
-; CHECK-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0x89,0xce,0xe1,0x03]
-; CHECK-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0xce,0xd1,0x03]
-; CHECK-NEXT: vpxor %xmm3, %xmm2, %xmm0 ## EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
-; CHECK-NEXT: vpxor %xmm0, %xmm4, %xmm0 ## EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8affineqb_128:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xce,0xd9,0x03]
+; X86-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xce,0xe1,0x03]
+; X86-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xce,0xd1,0x03]
+; X86-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
+; X86-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8affineqb_128:
+; X64: # %bb.0:
+; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xf9,0xce,0xd9,0x03]
+; X64-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0x89,0xce,0xe1,0x03]
+; X64-NEXT: vgf2p8affineqb $3, %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x09,0xce,0xd1,0x03]
+; X64-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
+; X64-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i16 %mask to <16 x i1>
%2 = call <16 x i8> @llvm.x86.vgf2p8affineqb.128(<16 x i8> %src1, <16 x i8> %src2, i8 3)
%3 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> zeroinitializer
@@ -83,15 +124,25 @@ define <16 x i8> @test_vgf2p8affineqb_12
declare <32 x i8> @llvm.x86.vgf2p8affineqb.256(<32 x i8>, <32 x i8>, i8)
define <32 x i8> @test_vgf2p8affineqb_256(<32 x i8> %src1, <32 x i8> %src2, <32 x i8> %passthru, i32 %mask) {
-; CHECK-LABEL: test_vgf2p8affineqb_256:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm3 ## EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xce,0xd9,0x03]
-; CHECK-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0xce,0xe1,0x03]
-; CHECK-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0xce,0xd1,0x03]
-; CHECK-NEXT: vpxor %ymm3, %ymm2, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
-; CHECK-NEXT: vpxor %ymm0, %ymm4, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8affineqb_256:
+; X86: # %bb.0:
+; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xce,0xd9,0x03]
+; X86-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xce,0xe1,0x03]
+; X86-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xce,0xd1,0x03]
+; X86-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
+; X86-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8affineqb_256:
+; X64: # %bb.0:
+; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe3,0xfd,0xce,0xd9,0x03]
+; X64-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xa9,0xce,0xe1,0x03]
+; X64-NEXT: vgf2p8affineqb $3, %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x29,0xce,0xd1,0x03]
+; X64-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
+; X64-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i32 %mask to <32 x i1>
%2 = call <32 x i8> @llvm.x86.vgf2p8affineqb.256(<32 x i8> %src1, <32 x i8> %src2, i8 3)
%3 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> zeroinitializer
@@ -103,15 +154,25 @@ define <32 x i8> @test_vgf2p8affineqb_25
declare <64 x i8> @llvm.x86.vgf2p8affineqb.512(<64 x i8>, <64 x i8>, i8)
define <64 x i8> @test_vgf2p8affineqb_512(<64 x i8> %src1, <64 x i8> %src2, <64 x i8> %passthru, i64 %mask) {
-; CHECK-LABEL: test_vgf2p8affineqb_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovq %rdi, %k1 ## encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm3 ## encoding: [0x62,0xf3,0xfd,0x48,0xce,0xd9,0x03]
-; CHECK-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xc9,0xce,0xe1,0x03]
-; CHECK-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x49,0xce,0xd1,0x03]
-; CHECK-NEXT: vpxorq %zmm3, %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
-; CHECK-NEXT: vpxorq %zmm0, %zmm4, %zmm0 ## encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8affineqb_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0xce,0xd9,0x03]
+; X86-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xce,0xe1,0x03]
+; X86-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xce,0xd1,0x03]
+; X86-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
+; X86-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8affineqb_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf3,0xfd,0x48,0xce,0xd9,0x03]
+; X64-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf3,0xfd,0xc9,0xce,0xe1,0x03]
+; X64-NEXT: vgf2p8affineqb $3, %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf3,0xfd,0x49,0xce,0xd1,0x03]
+; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
+; X64-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i64 %mask to <64 x i1>
%2 = call <64 x i8> @llvm.x86.vgf2p8affineqb.512(<64 x i8> %src1, <64 x i8> %src2, i8 3)
%3 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> zeroinitializer
@@ -123,15 +184,25 @@ define <64 x i8> @test_vgf2p8affineqb_51
declare <16 x i8> @llvm.x86.vgf2p8mulb.128(<16 x i8>, <16 x i8>)
define <16 x i8> @test_vgf2p8mulb_128(<16 x i8> %src1, <16 x i8> %src2, <16 x i8> %passthru, i16 %mask) {
-; CHECK-LABEL: test_vgf2p8mulb_128:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm3 ## EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0xcf,0xd9]
-; CHECK-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm4 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0xcf,0xe1]
-; CHECK-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xcf,0xd1]
-; CHECK-NEXT: vpxor %xmm3, %xmm2, %xmm0 ## EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
-; CHECK-NEXT: vpxor %xmm0, %xmm4, %xmm0 ## EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8mulb_128:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 # encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0xcf,0xd9]
+; X86-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x89,0xcf,0xe1]
+; X86-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xcf,0xd1]
+; X86-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
+; X86-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8mulb_128:
+; X64: # %bb.0:
+; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x79,0xcf,0xd9]
+; X64-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0x89,0xcf,0xe1]
+; X64-NEXT: vgf2p8mulb %xmm1, %xmm0, %xmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x09,0xcf,0xd1]
+; X64-NEXT: vpxor %xmm3, %xmm2, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xe9,0xef,0xc3]
+; X64-NEXT: vpxor %xmm0, %xmm4, %xmm0 # EVEX TO VEX Compression encoding: [0xc5,0xd9,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i16 %mask to <16 x i1>
%2 = call <16 x i8> @llvm.x86.vgf2p8mulb.128(<16 x i8> %src1, <16 x i8> %src2)
%3 = select <16 x i1> %1, <16 x i8> %2, <16 x i8> zeroinitializer
@@ -143,15 +214,25 @@ define <16 x i8> @test_vgf2p8mulb_128(<1
declare <32 x i8> @llvm.x86.vgf2p8mulb.256(<32 x i8>, <32 x i8>)
define <32 x i8> @test_vgf2p8mulb_256(<32 x i8> %src1, <32 x i8> %src2, <32 x i8> %passthru, i32 %mask) {
-; CHECK-LABEL: test_vgf2p8mulb_256:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm3 ## EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0xcf,0xd9]
-; CHECK-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm4 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0xcf,0xe1]
-; CHECK-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xcf,0xd1]
-; CHECK-NEXT: vpxor %ymm3, %ymm2, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
-; CHECK-NEXT: vpxor %ymm0, %ymm4, %ymm0 ## EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8mulb_256:
+; X86: # %bb.0:
+; X86-NEXT: kmovd {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf9,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0xcf,0xd9]
+; X86-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xa9,0xcf,0xe1]
+; X86-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xcf,0xd1]
+; X86-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
+; X86-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8mulb_256:
+; X64: # %bb.0:
+; X64-NEXT: kmovd %edi, %k1 # encoding: [0xc5,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm3 # EVEX TO VEX Compression encoding: [0xc4,0xe2,0x7d,0xcf,0xd9]
+; X64-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xa9,0xcf,0xe1]
+; X64-NEXT: vgf2p8mulb %ymm1, %ymm0, %ymm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x29,0xcf,0xd1]
+; X64-NEXT: vpxor %ymm3, %ymm2, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xed,0xef,0xc3]
+; X64-NEXT: vpxor %ymm0, %ymm4, %ymm0 # EVEX TO VEX Compression encoding: [0xc5,0xdd,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i32 %mask to <32 x i1>
%2 = call <32 x i8> @llvm.x86.vgf2p8mulb.256(<32 x i8> %src1, <32 x i8> %src2)
%3 = select <32 x i1> %1, <32 x i8> %2, <32 x i8> zeroinitializer
@@ -163,15 +244,25 @@ define <32 x i8> @test_vgf2p8mulb_256(<3
declare <64 x i8> @llvm.x86.vgf2p8mulb.512(<64 x i8>, <64 x i8>)
define <64 x i8> @test_vgf2p8mulb_512(<64 x i8> %src1, <64 x i8> %src2, <64 x i8> %passthru, i64 %mask) {
-; CHECK-LABEL: test_vgf2p8mulb_512:
-; CHECK: ## %bb.0:
-; CHECK-NEXT: kmovq %rdi, %k1 ## encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
-; CHECK-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm3 ## encoding: [0x62,0xf2,0x7d,0x48,0xcf,0xd9]
-; CHECK-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm4 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xc9,0xcf,0xe1]
-; CHECK-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xcf,0xd1]
-; CHECK-NEXT: vpxorq %zmm3, %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
-; CHECK-NEXT: vpxorq %zmm0, %zmm4, %zmm0 ## encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_vgf2p8mulb_512:
+; X86: # %bb.0:
+; X86-NEXT: kmovq {{[0-9]+}}(%esp), %k1 # encoding: [0xc4,0xe1,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0x7d,0x48,0xcf,0xd9]
+; X86-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xcf,0xe1]
+; X86-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xcf,0xd1]
+; X86-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
+; X86-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
+; X86-NEXT: retl # encoding: [0xc3]
+;
+; X64-LABEL: test_vgf2p8mulb_512:
+; X64: # %bb.0:
+; X64-NEXT: kmovq %rdi, %k1 # encoding: [0xc4,0xe1,0xfb,0x92,0xcf]
+; X64-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm3 # encoding: [0x62,0xf2,0x7d,0x48,0xcf,0xd9]
+; X64-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm4 {%k1} {z} # encoding: [0x62,0xf2,0x7d,0xc9,0xcf,0xe1]
+; X64-NEXT: vgf2p8mulb %zmm1, %zmm0, %zmm2 {%k1} # encoding: [0x62,0xf2,0x7d,0x49,0xcf,0xd1]
+; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0 # encoding: [0x62,0xf1,0xed,0x48,0xef,0xc3]
+; X64-NEXT: vpxorq %zmm0, %zmm4, %zmm0 # encoding: [0x62,0xf1,0xdd,0x48,0xef,0xc0]
+; X64-NEXT: retq # encoding: [0xc3]
%1 = bitcast i64 %mask to <64 x i1>
%2 = call <64 x i8> @llvm.x86.vgf2p8mulb.512(<64 x i8> %src1, <64 x i8> %src2)
%3 = select <64 x i1> %1, <64 x i8> %2, <64 x i8> zeroinitializer
Modified: llvm/trunk/test/CodeGen/X86/avx512-intrinsics-canonical.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512-intrinsics-canonical.ll?rev=333843&r1=333842&r2=333843&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-intrinsics-canonical.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512-intrinsics-canonical.ll Sun Jun 3 07:56:04 2018
@@ -1,5 +1,6 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl --show-mc-encoding | FileCheck %s
+; RUN: llc < %s -mtriple=i686-apple-darwin -mattr=+avx512f --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X86
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512f --show-mc-encoding | FileCheck %s --check-prefixes=CHECK,X64
; NOTE: This should use IR equivalent to what is generated by clang/test/CodeGen/avx512f-builtins.c
@@ -7,7 +8,7 @@ define <8 x double> @test_mm512_fmadd_ro
; CHECK-LABEL: test_mm512_fmadd_round_pd:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x18,0xa8,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
ret <8 x double> %0
@@ -16,11 +17,18 @@ entry:
declare <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double>, <8 x double>, <8 x double>, i32) #1
define <8 x double> @test_mm512_mask_fmadd_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x98,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x98,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x98,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
%1 = bitcast i8 %__U to <8 x i1>
@@ -29,12 +37,20 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb8,0xd1]
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb8,0xd1]
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb8,0xd1]
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
%1 = bitcast i8 %__U to <8 x i1>
@@ -43,11 +59,18 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmadd_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa8,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa8,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa8,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
%1 = bitcast i8 %__U to <8 x i1>
@@ -59,7 +82,7 @@ define <8 x double> @test_mm512_fmsub_ro
; CHECK-LABEL: test_mm512_fmsub_round_pd:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x18,0xaa,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub, i32 8)
@@ -67,11 +90,18 @@ entry:
}
define <8 x double> @test_mm512_mask_fmsub_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9a,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9a,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9a,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub, i32 8)
@@ -81,11 +111,18 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmsub_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xaa,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xaa,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xaa,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub, i32 8)
@@ -98,7 +135,7 @@ define <8 x double> @test_mm512_fnmadd_r
; CHECK-LABEL: test_mm512_fnmadd_round_pd:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x18,0xac,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %sub, <8 x double> %__B, <8 x double> %__C, i32 8)
@@ -106,12 +143,20 @@ entry:
}
define <8 x double> @test_mm512_mask3_fnmadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fnmadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xbc,0xd1]
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fnmadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xbc,0xd1]
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fnmadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xbc,0xd1]
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %sub, <8 x double> %__B, <8 x double> %__C, i32 8)
@@ -121,11 +166,18 @@ entry:
}
define <8 x double> @test_mm512_maskz_fnmadd_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fnmadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xac,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fnmadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xac,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fnmadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xac,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %sub, <8 x double> %__B, <8 x double> %__C, i32 8)
@@ -138,7 +190,7 @@ define <8 x double> @test_mm512_fnmsub_r
; CHECK-LABEL: test_mm512_fnmsub_round_pd:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x18,0xae,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%sub1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -147,11 +199,18 @@ entry:
}
define <8 x double> @test_mm512_maskz_fnmsub_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fnmsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xae,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fnmsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xae,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fnmsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xae,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%sub1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -166,19 +225,27 @@ define <8 x double> @test_mm512_fmadd_pd
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x48,0xa8,0xc2]
; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
ret <8 x double> %0
}
define <8 x double> @test_mm512_mask_fmadd_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
-; CHECK-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
+; X86-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x98,0xc1]
+; X64-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
%1 = bitcast i8 %__U to <8 x i1>
@@ -187,13 +254,22 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd1]
-; CHECK-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd1]
+; X86-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb8,0xd1]
+; X64-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
%1 = bitcast i8 %__U to <8 x i1>
@@ -202,12 +278,20 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmadd_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa8,0xc2]
-; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa8,0xc2]
+; X86-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa8,0xc2]
+; X64-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
%1 = bitcast i8 %__U to <8 x i1>
@@ -220,7 +304,7 @@ define <8 x double> @test_mm512_fmsub_pd
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmsub213pd %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x48,0xaa,0xc2]
; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -228,12 +312,20 @@ entry:
}
define <8 x double> @test_mm512_mask_fmsub_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9a,0xc1]
-; CHECK-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9a,0xc1]
+; X86-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9a,0xc1]
+; X64-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -243,12 +335,20 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmsub_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xaa,0xc2]
-; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xaa,0xc2]
+; X86-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xaa,0xc2]
+; X64-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -262,7 +362,7 @@ define <8 x double> @test_mm512_fnmadd_p
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x48,0xac,0xc2]
; CHECK-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %sub.i, <8 x double> %__B, <8 x double> %__C) #10
@@ -270,13 +370,22 @@ entry:
}
define <8 x double> @test_mm512_mask3_fnmadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fnmadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbc,0xd1]
-; CHECK-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fnmadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbc,0xd1]
+; X86-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fnmadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbc,0xd1]
+; X64-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %sub.i, <8 x double> %__B, <8 x double> %__C) #10
@@ -286,12 +395,20 @@ entry:
}
define <8 x double> @test_mm512_maskz_fnmadd_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fnmadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xac,0xc2]
-; CHECK-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fnmadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xac,0xc2]
+; X86-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fnmadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xac,0xc2]
+; X64-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %sub.i, <8 x double> %__B, <8 x double> %__C) #10
@@ -305,7 +422,7 @@ define <8 x double> @test_mm512_fnmsub_p
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x48,0xae,0xc2]
; CHECK-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%sub1.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -314,12 +431,20 @@ entry:
}
define <8 x double> @test_mm512_maskz_fnmsub_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fnmsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xae,0xc2]
-; CHECK-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fnmsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xae,0xc2]
+; X86-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fnmsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xae,0xc2]
+; X64-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%sub1.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -333,7 +458,7 @@ define <16 x float> @test_mm512_fmadd_ro
; CHECK-LABEL: test_mm512_fmadd_round_ps:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa8,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
ret <16 x float> %0
@@ -342,11 +467,17 @@ entry:
declare <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float>, <16 x float>, <16 x float>, i32) #1
define <16 x float> @test_mm512_mask_fmadd_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x98,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x98,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x98,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
%1 = bitcast i16 %__U to <16 x i1>
@@ -355,12 +486,19 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb8,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb8,0xd1]
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb8,0xd1]
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
%1 = bitcast i16 %__U to <16 x i1>
@@ -369,11 +507,17 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmadd_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa8,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa8,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa8,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
%1 = bitcast i16 %__U to <16 x i1>
@@ -385,7 +529,7 @@ define <16 x float> @test_mm512_fmsub_ro
; CHECK-LABEL: test_mm512_fmsub_round_ps:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xaa,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub, i32 8)
@@ -393,11 +537,17 @@ entry:
}
define <16 x float> @test_mm512_mask_fmsub_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9a,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9a,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9a,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub, i32 8)
@@ -407,11 +557,17 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmsub_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xaa,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xaa,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xaa,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub, i32 8)
@@ -424,7 +580,7 @@ define <16 x float> @test_mm512_fnmadd_r
; CHECK-LABEL: test_mm512_fnmadd_round_ps:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xac,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %sub, <16 x float> %__B, <16 x float> %__C, i32 8)
@@ -432,12 +588,19 @@ entry:
}
define <16 x float> @test_mm512_mask3_fnmadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fnmadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xbc,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fnmadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xbc,0xd1]
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fnmadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xbc,0xd1]
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %sub, <16 x float> %__B, <16 x float> %__C, i32 8)
@@ -447,11 +610,17 @@ entry:
}
define <16 x float> @test_mm512_maskz_fnmadd_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fnmadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xac,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fnmadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xac,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fnmadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xac,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %sub, <16 x float> %__B, <16 x float> %__C, i32 8)
@@ -464,7 +633,7 @@ define <16 x float> @test_mm512_fnmsub_r
; CHECK-LABEL: test_mm512_fnmsub_round_ps:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xae,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%sub1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -473,11 +642,17 @@ entry:
}
define <16 x float> @test_mm512_maskz_fnmsub_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fnmsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xae,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fnmsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xae,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fnmsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xae,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%sub1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -492,19 +667,26 @@ define <16 x float> @test_mm512_fmadd_ps
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x48,0xa8,0xc2]
; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
ret <16 x float> %0
}
define <16 x float> @test_mm512_mask_fmadd_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
-; CHECK-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
+; X86-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x98,0xc1]
+; X64-NEXT: ## zmm0 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
%1 = bitcast i16 %__U to <16 x i1>
@@ -513,13 +695,21 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd1]
-; CHECK-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd1]
+; X86-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb8,0xd1]
+; X64-NEXT: ## zmm2 = (zmm0 * zmm1) + zmm2
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
%1 = bitcast i16 %__U to <16 x i1>
@@ -528,12 +718,19 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmadd_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa8,0xc2]
-; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa8,0xc2]
+; X86-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa8,0xc2]
+; X64-NEXT: ## zmm0 = (zmm1 * zmm0) + zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
%1 = bitcast i16 %__U to <16 x i1>
@@ -546,7 +743,7 @@ define <16 x float> @test_mm512_fmsub_ps
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmsub213ps %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x48,0xaa,0xc2]
; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -554,12 +751,19 @@ entry:
}
define <16 x float> @test_mm512_mask_fmsub_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9a,0xc1]
-; CHECK-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9a,0xc1]
+; X86-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9a,0xc1]
+; X64-NEXT: ## zmm0 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -569,12 +773,19 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmsub_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xaa,0xc2]
-; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xaa,0xc2]
+; X86-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xaa,0xc2]
+; X64-NEXT: ## zmm0 = (zmm1 * zmm0) - zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -588,7 +799,7 @@ define <16 x float> @test_mm512_fnmadd_p
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x48,0xac,0xc2]
; CHECK-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %sub.i, <16 x float> %__B, <16 x float> %__C) #10
@@ -596,13 +807,21 @@ entry:
}
define <16 x float> @test_mm512_mask3_fnmadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fnmadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbc,0xd1]
-; CHECK-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fnmadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbc,0xd1]
+; X86-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fnmadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbc,0xd1]
+; X64-NEXT: ## zmm2 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %sub.i, <16 x float> %__B, <16 x float> %__C) #10
@@ -612,12 +831,19 @@ entry:
}
define <16 x float> @test_mm512_maskz_fnmadd_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fnmadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xac,0xc2]
-; CHECK-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fnmadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xac,0xc2]
+; X86-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fnmadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xac,0xc2]
+; X64-NEXT: ## zmm0 = -(zmm1 * zmm0) + zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %sub.i, <16 x float> %__B, <16 x float> %__C) #10
@@ -631,7 +857,7 @@ define <16 x float> @test_mm512_fnmsub_p
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x48,0xae,0xc2]
; CHECK-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%sub1.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -640,12 +866,19 @@ entry:
}
define <16 x float> @test_mm512_maskz_fnmsub_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fnmsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xae,0xc2]
-; CHECK-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fnmsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xae,0xc2]
+; X86-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fnmsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xae,0xc2]
+; X64-NEXT: ## zmm0 = -(zmm1 * zmm0) - zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%sub1.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -659,7 +892,7 @@ define <8 x double> @test_mm512_fmaddsub
; CHECK-LABEL: test_mm512_fmaddsub_round_pd:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x18,0xa6,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
ret <8 x double> %0
@@ -668,11 +901,18 @@ entry:
declare <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double>, <8 x double>, <8 x double>, i32) #1
define <8 x double> @test_mm512_mask_fmaddsub_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmaddsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x96,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmaddsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmaddsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x96,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmaddsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x96,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
%1 = bitcast i8 %__U to <8 x i1>
@@ -681,12 +921,20 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmaddsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmaddsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb6,0xd1]
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmaddsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmaddsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb6,0xd1]
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmaddsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb6,0xd1]
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
%1 = bitcast i8 %__U to <8 x i1>
@@ -695,11 +943,18 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmaddsub_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmaddsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa6,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmaddsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa6,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmaddsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa6,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
%1 = bitcast i8 %__U to <8 x i1>
@@ -711,7 +966,7 @@ define <8 x double> @test_mm512_fmsubadd
; CHECK-LABEL: test_mm512_fmsubadd_round_pd:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x18,0xa7,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub, i32 8)
@@ -719,11 +974,18 @@ entry:
}
define <8 x double> @test_mm512_mask_fmsubadd_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmsubadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x97,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmsubadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsubadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x97,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmsubadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x97,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub, i32 8)
@@ -733,11 +995,18 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmsubadd_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmsubadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa7,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmsubadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa7,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmsubadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x99,0xa7,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub, i32 8)
@@ -751,7 +1020,7 @@ define <8 x double> @test_mm512_fmaddsub
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x48,0xa6,0xc2]
; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -761,12 +1030,20 @@ entry:
}
define <8 x double> @test_mm512_mask_fmaddsub_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmaddsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
-; CHECK-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmaddsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
+; X86-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmaddsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x96,0xc1]
+; X64-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -778,13 +1055,22 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmaddsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmaddsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd1]
-; CHECK-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmaddsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd1]
+; X86-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmaddsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb6,0xd1]
+; X64-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -796,12 +1082,20 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmaddsub_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmaddsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa6,0xc2]
-; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmaddsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa6,0xc2]
+; X86-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmaddsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa6,0xc2]
+; X64-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -817,7 +1111,7 @@ define <8 x double> @test_mm512_fmsubadd
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmsubadd213pd %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0xf5,0x48,0xa7,0xc2]
; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -827,12 +1121,20 @@ entry:
}
define <8 x double> @test_mm512_mask_fmsubadd_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmsubadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x97,0xc1]
-; CHECK-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmsubadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsubadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x97,0xc1]
+; X86-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmsubadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x97,0xc1]
+; X64-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -844,12 +1146,20 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmsubadd_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmsubadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa7,0xc2]
-; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmsubadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsubadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa7,0xc2]
+; X86-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmsubadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd213pd %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xc9,0xa7,0xc2]
+; X64-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -864,7 +1174,7 @@ define <16 x float> @test_mm512_fmaddsub
; CHECK-LABEL: test_mm512_fmaddsub_round_ps:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa6,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
ret <16 x float> %0
@@ -873,11 +1183,17 @@ entry:
declare <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float>, <16 x float>, <16 x float>, i32) #1
define <16 x float> @test_mm512_mask_fmaddsub_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmaddsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x96,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmaddsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmaddsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x96,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmaddsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x96,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
%1 = bitcast i16 %__U to <16 x i1>
@@ -886,12 +1202,19 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmaddsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmaddsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb6,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmaddsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmaddsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb6,0xd1]
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmaddsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb6,0xd1]
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
%1 = bitcast i16 %__U to <16 x i1>
@@ -900,11 +1223,17 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmaddsub_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmaddsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa6,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmaddsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa6,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmaddsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa6,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
%1 = bitcast i16 %__U to <16 x i1>
@@ -916,7 +1245,7 @@ define <16 x float> @test_mm512_fmsubadd
; CHECK-LABEL: test_mm512_fmsubadd_round_ps:
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa7,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub, i32 8)
@@ -924,11 +1253,17 @@ entry:
}
define <16 x float> @test_mm512_mask_fmsubadd_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmsubadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x97,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmsubadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsubadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x97,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmsubadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x97,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub, i32 8)
@@ -938,11 +1273,17 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmsubadd_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmsubadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa7,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmsubadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa7,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmsubadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x99,0xa7,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub, i32 8)
@@ -956,7 +1297,7 @@ define <16 x float> @test_mm512_fmaddsub
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x48,0xa6,0xc2]
; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -966,12 +1307,19 @@ entry:
}
define <16 x float> @test_mm512_mask_fmaddsub_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmaddsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
-; CHECK-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmaddsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
+; X86-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmaddsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x96,0xc1]
+; X64-NEXT: ## zmm0 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -983,13 +1331,21 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmaddsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmaddsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd1]
-; CHECK-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmaddsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd1]
+; X86-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmaddsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb6,0xd1]
+; X64-NEXT: ## zmm2 = (zmm0 * zmm1) +/- zmm2
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1001,12 +1357,19 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmaddsub_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmaddsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa6,0xc2]
-; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmaddsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa6,0xc2]
+; X86-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmaddsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmaddsub213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa6,0xc2]
+; X64-NEXT: ## zmm0 = (zmm1 * zmm0) +/- zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1022,7 +1385,7 @@ define <16 x float> @test_mm512_fmsubadd
; CHECK: ## %bb.0: ## %entry
; CHECK-NEXT: vfmsubadd213ps %zmm2, %zmm1, %zmm0 ## encoding: [0x62,0xf2,0x75,0x48,0xa7,0xc2]
; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; CHECK-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -1032,12 +1395,19 @@ entry:
}
define <16 x float> @test_mm512_mask_fmsubadd_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fmsubadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x97,0xc1]
-; CHECK-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fmsubadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsubadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x97,0xc1]
+; X86-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fmsubadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x97,0xc1]
+; X64-NEXT: ## zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -1049,12 +1419,19 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmsubadd_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_maskz_fmsubadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa7,0xc2]
-; CHECK-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_maskz_fmsubadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsubadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa7,0xc2]
+; X86-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_maskz_fmsubadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd213ps %zmm2, %zmm1, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xc9,0xa7,0xc2]
+; X64-NEXT: ## zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -1066,12 +1443,20 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xba,0xd1]
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xba,0xd1]
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xba,0xd1]
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub, i32 8)
@@ -1081,13 +1466,22 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd1]
-; CHECK-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd1]
+; X86-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xba,0xd1]
+; X64-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -1097,12 +1491,19 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xba,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xba,0xd1]
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xba,0xd1]
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub, i32 8)
@@ -1112,13 +1513,21 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd1]
-; CHECK-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd1]
+; X86-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xba,0xd1]
+; X64-NEXT: ## zmm2 = (zmm0 * zmm1) - zmm2
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -1128,12 +1537,20 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmsubadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmsubadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb7,0xd1]
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmsubadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsubadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb7,0xd1]
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmsubadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xb7,0xd1]
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub, i32 8)
@@ -1143,13 +1560,22 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmsubadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmsubadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd1]
-; CHECK-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmsubadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd1]
+; X86-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmsubadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xb7,0xd1]
+; X64-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -1161,12 +1587,19 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmsubadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmsubadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb7,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmsubadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsubadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb7,0xd1]
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmsubadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xb7,0xd1]
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub, i32 8)
@@ -1176,13 +1609,21 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmsubadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fmsubadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd1]
-; CHECK-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fmsubadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd1]
+; X86-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fmsubadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsubadd231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xb7,0xd1]
+; X64-NEXT: ## zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -1194,11 +1635,18 @@ entry:
}
define <8 x double> @test_mm512_mask_fnmadd_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fnmadd_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9c,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fnmadd_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9c,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fnmadd_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9c,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %sub, <8 x double> %__B, <8 x double> %__C, i32 8)
@@ -1208,12 +1656,20 @@ entry:
}
define <8 x double> @test_mm512_mask_fnmadd_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fnmadd_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
-; CHECK-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fnmadd_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
+; X86-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fnmadd_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9c,0xc1]
+; X64-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %sub.i, <8 x double> %__B, <8 x double> %__C) #10
@@ -1223,11 +1679,17 @@ entry:
}
define <16 x float> @test_mm512_mask_fnmadd_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fnmadd_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9c,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fnmadd_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9c,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fnmadd_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9c,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %sub, <16 x float> %__B, <16 x float> %__C, i32 8)
@@ -1237,12 +1699,19 @@ entry:
}
define <16 x float> @test_mm512_mask_fnmadd_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fnmadd_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
-; CHECK-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fnmadd_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
+; X86-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fnmadd_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9c,0xc1]
+; X64-NEXT: ## zmm0 = -(zmm0 * zmm1) + zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %sub.i, <16 x float> %__B, <16 x float> %__C) #10
@@ -1252,11 +1721,18 @@ entry:
}
define <8 x double> @test_mm512_mask_fnmsub_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fnmsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9e,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fnmsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9e,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fnmsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x19,0x9e,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
%sub1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -1267,12 +1743,20 @@ entry:
}
define <8 x double> @test_mm512_mask3_fnmsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fnmsub_round_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xbe,0xd1]
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fnmsub_round_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xbe,0xd1]
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fnmsub_round_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0xbe,0xd1]
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
%sub1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -1283,12 +1767,20 @@ entry:
}
define <8 x double> @test_mm512_mask_fnmsub_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; CHECK-LABEL: test_mm512_mask_fnmsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
-; CHECK-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fnmsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
+; X86-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fnmsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132pd %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0xed,0x49,0x9e,0xc1]
+; X64-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
%sub2.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -1299,13 +1791,22 @@ entry:
}
define <8 x double> @test_mm512_mask3_fnmsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fnmsub_pd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd1]
-; CHECK-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
-; CHECK-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fnmsub_pd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movzbl {{[0-9]+}}(%esp), %eax ## encoding: [0x0f,0xb6,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd1]
+; X86-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fnmsub_pd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub231pd %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0xbe,0xd1]
+; X64-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: vmovapd %zmm2, %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__B
%sub2.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -1316,11 +1817,17 @@ entry:
}
define <16 x float> @test_mm512_mask_fnmsub_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fnmsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9e,0xc1]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fnmsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9e,0xc1]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fnmsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x19,0x9e,0xc1]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
%sub1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1331,12 +1838,19 @@ entry:
}
define <16 x float> @test_mm512_mask3_fnmsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fnmsub_round_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xbe,0xd1]
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fnmsub_round_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xbe,0xd1]
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fnmsub_round_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0xbe,0xd1]
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
%sub1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1347,12 +1861,19 @@ entry:
}
define <16 x float> @test_mm512_mask_fnmsub_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; CHECK-LABEL: test_mm512_mask_fnmsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
-; CHECK-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask_fnmsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
+; X86-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask_fnmsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub132ps %zmm1, %zmm2, %zmm0 {%k1} ## encoding: [0x62,0xf2,0x6d,0x49,0x9e,0xc1]
+; X64-NEXT: ## zmm0 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
%sub1.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1363,13 +1884,21 @@ entry:
}
define <16 x float> @test_mm512_mask3_fnmsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; CHECK-LABEL: test_mm512_mask3_fnmsub_ps:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd1]
-; CHECK-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
-; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm512_mask3_fnmsub_ps:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1 ## encoding: [0xc5,0xf8,0x90,0x4c,0x24,0x04]
+; X86-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd1]
+; X86-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm512_mask3_fnmsub_ps:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub231ps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x49,0xbe,0xd1]
+; X64-NEXT: ## zmm2 = -(zmm0 * zmm1) - zmm2
+; X64-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
%sub1.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1380,12 +1909,20 @@ entry:
}
define <4 x float> @test_mm_mask_fmadd_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; CHECK-LABEL: test_mm_mask_fmadd_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fmadd_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fmadd_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
%1 = extractelement <4 x float> %__A, i64 0
@@ -1400,12 +1937,20 @@ entry:
}
define <4 x float> @test_mm_mask_fmadd_round_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; CHECK-LABEL: test_mm_mask_fmadd_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fmadd_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fmadd_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa9,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.x86.avx512.mask.vfmadd.ss(<4 x float> %__W, <4 x float> %__A, <4 x float> %__B, i8 %__U, i32 4)
ret <4 x float> %0
@@ -1414,12 +1959,20 @@ entry:
declare <4 x float> @llvm.x86.avx512.mask.vfmadd.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) #1
define <4 x float> @test_mm_maskz_fmadd_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; CHECK-LABEL: test_mm_maskz_fmadd_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fmadd_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fmadd_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__A, i64 0
%1 = extractelement <4 x float> %__B, i64 0
@@ -1433,12 +1986,20 @@ entry:
}
define <4 x float> @test_mm_maskz_fmadd_round_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; CHECK-LABEL: test_mm_maskz_fmadd_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fmadd_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fmadd_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa9,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float> %__A, <4 x float> %__B, <4 x float> %__C, i8 %__U, i32 4)
ret <4 x float> %0
@@ -1447,13 +2008,22 @@ entry:
declare <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) #1
define <4 x float> @test_mm_mask3_fmadd_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fmadd_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fmadd_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fmadd_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
%1 = extractelement <4 x float> %__X, i64 0
@@ -1468,13 +2038,22 @@ entry:
}
define <4 x float> @test_mm_mask3_fmadd_round_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fmadd_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fmadd_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fmadd_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb9,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.x86.avx512.mask3.vfmadd.ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 %__U, i32 4)
ret <4 x float> %0
@@ -1483,12 +2062,20 @@ entry:
declare <4 x float> @llvm.x86.avx512.mask3.vfmadd.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) #1
define <4 x float> @test_mm_mask_fmsub_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; CHECK-LABEL: test_mm_mask_fmsub_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fmsub_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fmsub_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
%1 = extractelement <4 x float> %__A, i64 0
@@ -1504,12 +2091,20 @@ entry:
}
define <4 x float> @test_mm_mask_fmsub_round_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; CHECK-LABEL: test_mm_mask_fmsub_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fmsub_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fmsub_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xab,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
%0 = tail call <4 x float> @llvm.x86.avx512.mask.vfmadd.ss(<4 x float> %__W, <4 x float> %__A, <4 x float> %sub, i8 %__U, i32 4)
@@ -1517,12 +2112,20 @@ entry:
}
define <4 x float> @test_mm_maskz_fmsub_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; CHECK-LABEL: test_mm_maskz_fmsub_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fmsub_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fmsub_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__A, i64 0
%1 = extractelement <4 x float> %__B, i64 0
@@ -1537,12 +2140,20 @@ entry:
}
define <4 x float> @test_mm_maskz_fmsub_round_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; CHECK-LABEL: test_mm_maskz_fmsub_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fmsub_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fmsub_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xab,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float> %__A, <4 x float> %__B, <4 x float> %sub, i8 %__U, i32 4)
@@ -1550,13 +2161,22 @@ entry:
}
define <4 x float> @test_mm_mask3_fmsub_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fmsub_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fmsub_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fmsub_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
%1 = extractelement <4 x float> %__X, i64 0
@@ -1572,13 +2192,22 @@ entry:
}
define <4 x float> @test_mm_mask3_fmsub_round_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fmsub_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fmsub_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fmsub_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <4 x float> @llvm.x86.avx512.mask3.vfmsub.ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 %__U, i32 4)
ret <4 x float> %0
@@ -1587,12 +2216,20 @@ entry:
declare <4 x float> @llvm.x86.avx512.mask3.vfmsub.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) #1
define <4 x float> @test_mm_mask_fnmadd_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; CHECK-LABEL: test_mm_mask_fnmadd_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fnmadd_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fnmadd_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
%.rhs.i = extractelement <4 x float> %__A, i64 0
@@ -1608,12 +2245,20 @@ entry:
}
define <4 x float> @test_mm_mask_fnmadd_round_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; CHECK-LABEL: test_mm_mask_fnmadd_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fnmadd_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fnmadd_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xad,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%0 = tail call <4 x float> @llvm.x86.avx512.mask.vfmadd.ss(<4 x float> %__W, <4 x float> %sub, <4 x float> %__B, i8 %__U, i32 4)
@@ -1621,12 +2266,20 @@ entry:
}
define <4 x float> @test_mm_maskz_fnmadd_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; CHECK-LABEL: test_mm_maskz_fnmadd_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fnmadd_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fnmadd_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__A, i64 0
%.rhs.i = extractelement <4 x float> %__B, i64 0
@@ -1641,12 +2294,20 @@ entry:
}
define <4 x float> @test_mm_maskz_fnmadd_round_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; CHECK-LABEL: test_mm_maskz_fnmadd_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fnmadd_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fnmadd_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xad,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
%0 = tail call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float> %__A, <4 x float> %sub, <4 x float> %__C, i8 %__U, i32 4)
@@ -1654,13 +2315,22 @@ entry:
}
define <4 x float> @test_mm_mask3_fnmadd_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fnmadd_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
-; CHECK-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fnmadd_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
+; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fnmadd_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
+; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
%.rhs.i = extractelement <4 x float> %__X, i64 0
@@ -1676,13 +2346,22 @@ entry:
}
define <4 x float> @test_mm_mask3_fnmadd_round_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fnmadd_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
-; CHECK-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fnmadd_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
+; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fnmadd_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbd,0xd1]
+; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__X
%0 = tail call <4 x float> @llvm.x86.avx512.mask3.vfmadd.ss(<4 x float> %__W, <4 x float> %sub, <4 x float> %__Y, i8 %__U, i32 4)
@@ -1690,12 +2369,20 @@ entry:
}
define <4 x float> @test_mm_mask_fnmsub_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; CHECK-LABEL: test_mm_mask_fnmsub_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fnmsub_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fnmsub_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
%.rhs.i = extractelement <4 x float> %__A, i64 0
@@ -1712,12 +2399,20 @@ entry:
}
define <4 x float> @test_mm_mask_fnmsub_round_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; CHECK-LABEL: test_mm_mask_fnmsub_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fnmsub_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fnmsub_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xaf,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%sub1 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
@@ -1726,12 +2421,20 @@ entry:
}
define <4 x float> @test_mm_maskz_fnmsub_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; CHECK-LABEL: test_mm_maskz_fnmsub_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fnmsub_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fnmsub_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__A, i64 0
%.rhs.i = extractelement <4 x float> %__B, i64 0
@@ -1747,12 +2450,20 @@ entry:
}
define <4 x float> @test_mm_maskz_fnmsub_round_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; CHECK-LABEL: test_mm_maskz_fnmsub_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fnmsub_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fnmsub_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213ss %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xaf,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__B
%sub1 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -1761,13 +2472,22 @@ entry:
}
define <4 x float> @test_mm_mask3_fnmsub_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fnmsub_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd1]
-; CHECK-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fnmsub_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd1]
+; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fnmsub_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbf,0xd1]
+; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <4 x float> %__W, i64 0
%.rhs.i = extractelement <4 x float> %__X, i64 0
@@ -1784,17 +2504,30 @@ entry:
}
define <4 x float> @test_mm_mask3_fnmsub_round_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fnmsub_round_ss:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; CHECK-NEXT: ## encoding: [0xc4,0xe2,0x79,0x18,0x1d,A,A,A,A]
-; CHECK-NEXT: ## fixup A - offset: 5, value: LCPI119_0-4, kind: reloc_riprel_4byte
-; CHECK-NEXT: vxorps %xmm3, %xmm1, %xmm1 ## encoding: [0xc5,0xf0,0x57,0xcb]
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
-; CHECK-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fnmsub_round_ss:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: ## encoding: [0xc4,0xe2,0x79,0x18,0x1d,A,A,A,A]
+; X86-NEXT: ## fixup A - offset: 5, value: LCPI119_0, kind: FK_Data_4
+; X86-NEXT: vxorps %xmm3, %xmm1, %xmm1 ## encoding: [0xc5,0xf0,0x57,0xcb]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fnmsub_round_ss:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X64-NEXT: ## encoding: [0xc4,0xe2,0x79,0x18,0x1d,A,A,A,A]
+; X64-NEXT: ## fixup A - offset: 5, value: LCPI119_0-4, kind: reloc_riprel_4byte
+; X64-NEXT: vxorps %xmm3, %xmm1, %xmm1 ## encoding: [0xc5,0xf0,0x57,0xcb]
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231ss %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbb,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vmovaps %xmm2, %xmm0 ## encoding: [0xc5,0xf8,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__X
%0 = tail call <4 x float> @llvm.x86.avx512.mask3.vfmsub.ss(<4 x float> %__W, <4 x float> %sub, <4 x float> %__Y, i8 %__U, i32 4)
@@ -1802,12 +2535,20 @@ entry:
}
define <2 x double> @test_mm_mask_fmadd_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; CHECK-LABEL: test_mm_mask_fmadd_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fmadd_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fmadd_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
%1 = extractelement <2 x double> %__A, i64 0
@@ -1822,12 +2563,20 @@ entry:
}
define <2 x double> @test_mm_mask_fmadd_round_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; CHECK-LABEL: test_mm_mask_fmadd_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fmadd_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fmadd_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa9,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.x86.avx512.mask.vfmadd.sd(<2 x double> %__W, <2 x double> %__A, <2 x double> %__B, i8 %__U, i32 4)
ret <2 x double> %0
@@ -1836,12 +2585,20 @@ entry:
declare <2 x double> @llvm.x86.avx512.mask.vfmadd.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) #1
define <2 x double> @test_mm_maskz_fmadd_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; CHECK-LABEL: test_mm_maskz_fmadd_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fmadd_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fmadd_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__A, i64 0
%1 = extractelement <2 x double> %__B, i64 0
@@ -1855,12 +2612,20 @@ entry:
}
define <2 x double> @test_mm_maskz_fmadd_round_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; CHECK-LABEL: test_mm_maskz_fmadd_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fmadd_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fmadd_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa9,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.x86.avx512.maskz.vfmadd.sd(<2 x double> %__A, <2 x double> %__B, <2 x double> %__C, i8 %__U, i32 4)
ret <2 x double> %0
@@ -1869,13 +2634,22 @@ entry:
declare <2 x double> @llvm.x86.avx512.maskz.vfmadd.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) #1
define <2 x double> @test_mm_mask3_fmadd_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fmadd_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
-; CHECK-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fmadd_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fmadd_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
%1 = extractelement <2 x double> %__X, i64 0
@@ -1890,13 +2664,22 @@ entry:
}
define <2 x double> @test_mm_mask3_fmadd_round_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fmadd_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
-; CHECK-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fmadd_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fmadd_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb9,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) + xmm2
+; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.x86.avx512.mask3.vfmadd.sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 %__U, i32 4)
ret <2 x double> %0
@@ -1905,12 +2688,20 @@ entry:
declare <2 x double> @llvm.x86.avx512.mask3.vfmadd.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) #1
define <2 x double> @test_mm_mask_fmsub_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; CHECK-LABEL: test_mm_mask_fmsub_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fmsub_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fmsub_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
%1 = extractelement <2 x double> %__A, i64 0
@@ -1926,12 +2717,20 @@ entry:
}
define <2 x double> @test_mm_mask_fmsub_round_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; CHECK-LABEL: test_mm_mask_fmsub_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fmsub_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fmsub_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xab,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
%0 = tail call <2 x double> @llvm.x86.avx512.mask.vfmadd.sd(<2 x double> %__W, <2 x double> %__A, <2 x double> %sub, i8 %__U, i32 4)
@@ -1939,12 +2738,20 @@ entry:
}
define <2 x double> @test_mm_maskz_fmsub_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; CHECK-LABEL: test_mm_maskz_fmsub_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fmsub_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fmsub_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__A, i64 0
%1 = extractelement <2 x double> %__B, i64 0
@@ -1959,12 +2766,20 @@ entry:
}
define <2 x double> @test_mm_maskz_fmsub_round_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; CHECK-LABEL: test_mm_maskz_fmsub_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
-; CHECK-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fmsub_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
+; X86-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fmsub_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xab,0xc2]
+; X64-NEXT: ## xmm0 = (xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <2 x double> @llvm.x86.avx512.maskz.vfmadd.sd(<2 x double> %__A, <2 x double> %__B, <2 x double> %sub, i8 %__U, i32 4)
@@ -1972,13 +2787,22 @@ entry:
}
define <2 x double> @test_mm_mask3_fmsub_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fmsub_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
-; CHECK-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fmsub_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fmsub_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
%1 = extractelement <2 x double> %__X, i64 0
@@ -1994,13 +2818,22 @@ entry:
}
define <2 x double> @test_mm_mask3_fmsub_round_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fmsub_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
-; CHECK-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fmsub_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fmsub_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = tail call <2 x double> @llvm.x86.avx512.mask3.vfmsub.sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 %__U, i32 4)
ret <2 x double> %0
@@ -2009,12 +2842,20 @@ entry:
declare <2 x double> @llvm.x86.avx512.mask3.vfmsub.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) #1
define <2 x double> @test_mm_mask_fnmadd_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; CHECK-LABEL: test_mm_mask_fnmadd_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fnmadd_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fnmadd_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
%.rhs.i = extractelement <2 x double> %__A, i64 0
@@ -2030,12 +2871,20 @@ entry:
}
define <2 x double> @test_mm_mask_fnmadd_round_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; CHECK-LABEL: test_mm_mask_fnmadd_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fnmadd_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fnmadd_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xad,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__A
%0 = tail call <2 x double> @llvm.x86.avx512.mask.vfmadd.sd(<2 x double> %__W, <2 x double> %sub, <2 x double> %__B, i8 %__U, i32 4)
@@ -2043,12 +2892,20 @@ entry:
}
define <2 x double> @test_mm_maskz_fnmadd_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; CHECK-LABEL: test_mm_maskz_fnmadd_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fnmadd_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fnmadd_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__A, i64 0
%.rhs.i = extractelement <2 x double> %__B, i64 0
@@ -2063,12 +2920,20 @@ entry:
}
define <2 x double> @test_mm_maskz_fnmadd_round_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; CHECK-LABEL: test_mm_maskz_fnmadd_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fnmadd_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fnmadd_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xad,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) + xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
%0 = tail call <2 x double> @llvm.x86.avx512.maskz.vfmadd.sd(<2 x double> %__A, <2 x double> %sub, <2 x double> %__C, i8 %__U, i32 4)
@@ -2076,13 +2941,22 @@ entry:
}
define <2 x double> @test_mm_mask3_fnmadd_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fnmadd_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
-; CHECK-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
-; CHECK-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fnmadd_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
+; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fnmadd_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
+; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
%.rhs.i = extractelement <2 x double> %__X, i64 0
@@ -2098,13 +2972,22 @@ entry:
}
define <2 x double> @test_mm_mask3_fnmadd_round_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fnmadd_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
-; CHECK-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
-; CHECK-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fnmadd_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
+; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fnmadd_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmadd231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbd,0xd1]
+; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) + xmm2
+; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__X
%0 = tail call <2 x double> @llvm.x86.avx512.mask3.vfmadd.sd(<2 x double> %__W, <2 x double> %sub, <2 x double> %__Y, i8 %__U, i32 4)
@@ -2112,12 +2995,20 @@ entry:
}
define <2 x double> @test_mm_mask_fnmsub_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; CHECK-LABEL: test_mm_mask_fnmsub_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fnmsub_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fnmsub_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
%.rhs.i = extractelement <2 x double> %__A, i64 0
@@ -2134,12 +3025,20 @@ entry:
}
define <2 x double> @test_mm_mask_fnmsub_round_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; CHECK-LABEL: test_mm_mask_fnmsub_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask_fnmsub_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask_fnmsub_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xaf,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__A
%sub1 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
@@ -2148,12 +3047,20 @@ entry:
}
define <2 x double> @test_mm_maskz_fnmsub_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; CHECK-LABEL: test_mm_maskz_fnmsub_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fnmsub_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fnmsub_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__A, i64 0
%.rhs.i = extractelement <2 x double> %__B, i64 0
@@ -2169,12 +3076,20 @@ entry:
}
define <2 x double> @test_mm_maskz_fnmsub_round_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; CHECK-LABEL: test_mm_maskz_fnmsub_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
-; CHECK-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_maskz_fnmsub_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
+; X86-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_maskz_fnmsub_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub213sd %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xaf,0xc2]
+; X64-NEXT: ## xmm0 = -(xmm1 * xmm0) - xmm2
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__B
%sub1 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__C
@@ -2183,13 +3098,22 @@ entry:
}
define <2 x double> @test_mm_mask3_fnmsub_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fnmsub_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd1]
-; CHECK-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
-; CHECK-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fnmsub_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd1]
+; X86-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fnmsub_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfnmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbf,0xd1]
+; X64-NEXT: ## xmm2 = -(xmm0 * xmm1) - xmm2
+; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%0 = extractelement <2 x double> %__W, i64 0
%.rhs.i = extractelement <2 x double> %__X, i64 0
@@ -2206,15 +3130,26 @@ entry:
}
define <2 x double> @test_mm_mask3_fnmsub_round_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; CHECK-LABEL: test_mm_mask3_fnmsub_round_sd:
-; CHECK: ## %bb.0: ## %entry
-; CHECK-NEXT: vxorpd {{.*}}(%rip), %xmm1, %xmm1 ## encoding: [0xc5,0xf1,0x57,0x0d,A,A,A,A]
-; CHECK-NEXT: ## fixup A - offset: 4, value: LCPI143_0-4, kind: reloc_riprel_4byte
-; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
-; CHECK-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
-; CHECK-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
-; CHECK-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
-; CHECK-NEXT: retq ## encoding: [0xc3]
+; X86-LABEL: test_mm_mask3_fnmsub_round_sd:
+; X86: ## %bb.0: ## %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al ## encoding: [0x8a,0x44,0x24,0x04]
+; X86-NEXT: vxorpd LCPI143_0, %xmm1, %xmm1 ## encoding: [0xc5,0xf1,0x57,0x0d,A,A,A,A]
+; X86-NEXT: ## fixup A - offset: 4, value: LCPI143_0, kind: FK_Data_4
+; X86-NEXT: kmovw %eax, %k1 ## encoding: [0xc5,0xf8,0x92,0xc8]
+; X86-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
+; X86-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X86-NEXT: retl ## encoding: [0xc3]
+;
+; X64-LABEL: test_mm_mask3_fnmsub_round_sd:
+; X64: ## %bb.0: ## %entry
+; X64-NEXT: vxorpd {{.*}}(%rip), %xmm1, %xmm1 ## encoding: [0xc5,0xf1,0x57,0x0d,A,A,A,A]
+; X64-NEXT: ## fixup A - offset: 4, value: LCPI143_0-4, kind: reloc_riprel_4byte
+; X64-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; X64-NEXT: vfmsub231sd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbb,0xd1]
+; X64-NEXT: ## xmm2 = (xmm0 * xmm1) - xmm2
+; X64-NEXT: vmovapd %xmm2, %xmm0 ## encoding: [0xc5,0xf9,0x28,0xc2]
+; X64-NEXT: retq ## encoding: [0xc3]
entry:
%sub = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %__X
%0 = tail call <2 x double> @llvm.x86.avx512.mask3.vfmsub.sd(<2 x double> %__W, <2 x double> %sub, <2 x double> %__Y, i8 %__U, i32 4)
Modified: llvm/trunk/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll?rev=333843&r1=333842&r2=333843&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512-intrinsics-fast-isel.ll Sun Jun 3 07:56:04 2018
@@ -1,32 +1,32 @@
; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc < %s -fast-isel -mtriple=i386-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=ALL --check-prefix=X32
-; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=ALL --check-prefix=X64
+; RUN: llc < %s -fast-isel -mtriple=i386-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,X86
+; RUN: llc < %s -fast-isel -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,X64
; NOTE: This should use IR equivalent to what is generated by clang/test/CodeGen/avx512f-builtins.c
define zeroext i16 @test_mm512_kunpackb(<8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__C, <8 x i64> %__D, <8 x i64> %__E, <8 x i64> %__F) local_unnamed_addr #0 {
-; X32-LABEL: test_mm512_kunpackb:
-; X32: # %bb.0: # %entry
-; X32-NEXT: pushl %ebp
-; X32-NEXT: .cfi_def_cfa_offset 8
-; X32-NEXT: .cfi_offset %ebp, -8
-; X32-NEXT: movl %esp, %ebp
-; X32-NEXT: .cfi_def_cfa_register %ebp
-; X32-NEXT: andl $-64, %esp
-; X32-NEXT: subl $64, %esp
-; X32-NEXT: vmovdqa64 136(%ebp), %zmm3
-; X32-NEXT: vpcmpneqd %zmm1, %zmm0, %k0
-; X32-NEXT: vpcmpneqd 8(%ebp), %zmm2, %k1
-; X32-NEXT: kunpckbw %k0, %k1, %k1
-; X32-NEXT: vpcmpneqd 72(%ebp), %zmm3, %k0 {%k1}
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: movzwl %ax, %eax
-; X32-NEXT: movl %ebp, %esp
-; X32-NEXT: popl %ebp
-; X32-NEXT: .cfi_def_cfa %esp, 4
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_kunpackb:
+; X86: # %bb.0: # %entry
+; X86-NEXT: pushl %ebp
+; X86-NEXT: .cfi_def_cfa_offset 8
+; X86-NEXT: .cfi_offset %ebp, -8
+; X86-NEXT: movl %esp, %ebp
+; X86-NEXT: .cfi_def_cfa_register %ebp
+; X86-NEXT: andl $-64, %esp
+; X86-NEXT: subl $64, %esp
+; X86-NEXT: vmovdqa64 136(%ebp), %zmm3
+; X86-NEXT: vpcmpneqd %zmm1, %zmm0, %k0
+; X86-NEXT: vpcmpneqd 8(%ebp), %zmm2, %k1
+; X86-NEXT: kunpckbw %k0, %k1, %k1
+; X86-NEXT: vpcmpneqd 72(%ebp), %zmm3, %k0 {%k1}
+; X86-NEXT: kmovw %k0, %eax
+; X86-NEXT: movzwl %ax, %eax
+; X86-NEXT: movl %ebp, %esp
+; X86-NEXT: popl %ebp
+; X86-NEXT: .cfi_def_cfa %esp, 4
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_kunpackb:
; X64: # %bb.0: # %entry
@@ -57,28 +57,28 @@ entry:
}
define i32 @test_mm512_kortestc(<8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__C, <8 x i64> %__D) {
-; X32-LABEL: test_mm512_kortestc:
-; X32: # %bb.0: # %entry
-; X32-NEXT: pushl %ebp
-; X32-NEXT: .cfi_def_cfa_offset 8
-; X32-NEXT: .cfi_offset %ebp, -8
-; X32-NEXT: movl %esp, %ebp
-; X32-NEXT: .cfi_def_cfa_register %ebp
-; X32-NEXT: andl $-64, %esp
-; X32-NEXT: subl $64, %esp
-; X32-NEXT: vpcmpneqd %zmm1, %zmm0, %k0
-; X32-NEXT: vpcmpneqd 8(%ebp), %zmm2, %k1
-; X32-NEXT: korw %k0, %k1, %k0
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: cmpw $-1, %ax
-; X32-NEXT: sete %al
-; X32-NEXT: andb $1, %al
-; X32-NEXT: movzbl %al, %eax
-; X32-NEXT: movl %ebp, %esp
-; X32-NEXT: popl %ebp
-; X32-NEXT: .cfi_def_cfa %esp, 4
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_kortestc:
+; X86: # %bb.0: # %entry
+; X86-NEXT: pushl %ebp
+; X86-NEXT: .cfi_def_cfa_offset 8
+; X86-NEXT: .cfi_offset %ebp, -8
+; X86-NEXT: movl %esp, %ebp
+; X86-NEXT: .cfi_def_cfa_register %ebp
+; X86-NEXT: andl $-64, %esp
+; X86-NEXT: subl $64, %esp
+; X86-NEXT: vpcmpneqd %zmm1, %zmm0, %k0
+; X86-NEXT: vpcmpneqd 8(%ebp), %zmm2, %k1
+; X86-NEXT: korw %k0, %k1, %k0
+; X86-NEXT: kmovw %k0, %eax
+; X86-NEXT: cmpw $-1, %ax
+; X86-NEXT: sete %al
+; X86-NEXT: andb $1, %al
+; X86-NEXT: movzbl %al, %eax
+; X86-NEXT: movl %ebp, %esp
+; X86-NEXT: popl %ebp
+; X86-NEXT: .cfi_def_cfa %esp, 4
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_kortestc:
; X64: # %bb.0: # %entry
@@ -106,28 +106,28 @@ entry:
}
define i32 @test_mm512_kortestz(<8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__C, <8 x i64> %__D) {
-; X32-LABEL: test_mm512_kortestz:
-; X32: # %bb.0: # %entry
-; X32-NEXT: pushl %ebp
-; X32-NEXT: .cfi_def_cfa_offset 8
-; X32-NEXT: .cfi_offset %ebp, -8
-; X32-NEXT: movl %esp, %ebp
-; X32-NEXT: .cfi_def_cfa_register %ebp
-; X32-NEXT: andl $-64, %esp
-; X32-NEXT: subl $64, %esp
-; X32-NEXT: vpcmpneqd %zmm1, %zmm0, %k0
-; X32-NEXT: vpcmpneqd 8(%ebp), %zmm2, %k1
-; X32-NEXT: korw %k0, %k1, %k0
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: cmpw $0, %ax
-; X32-NEXT: sete %al
-; X32-NEXT: andb $1, %al
-; X32-NEXT: movzbl %al, %eax
-; X32-NEXT: movl %ebp, %esp
-; X32-NEXT: popl %ebp
-; X32-NEXT: .cfi_def_cfa %esp, 4
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_kortestz:
+; X86: # %bb.0: # %entry
+; X86-NEXT: pushl %ebp
+; X86-NEXT: .cfi_def_cfa_offset 8
+; X86-NEXT: .cfi_offset %ebp, -8
+; X86-NEXT: movl %esp, %ebp
+; X86-NEXT: .cfi_def_cfa_register %ebp
+; X86-NEXT: andl $-64, %esp
+; X86-NEXT: subl $64, %esp
+; X86-NEXT: vpcmpneqd %zmm1, %zmm0, %k0
+; X86-NEXT: vpcmpneqd 8(%ebp), %zmm2, %k1
+; X86-NEXT: korw %k0, %k1, %k0
+; X86-NEXT: kmovw %k0, %eax
+; X86-NEXT: cmpw $0, %ax
+; X86-NEXT: sete %al
+; X86-NEXT: andb $1, %al
+; X86-NEXT: movzbl %al, %eax
+; X86-NEXT: movl %ebp, %esp
+; X86-NEXT: popl %ebp
+; X86-NEXT: .cfi_def_cfa %esp, 4
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_kortestz:
; X64: # %bb.0: # %entry
@@ -156,15 +156,10 @@ entry:
}
define <16 x float> @test_mm512_shuffle_f32x4(<16 x float> %__A, <16 x float> %__B) {
-; X32-LABEL: test_mm512_shuffle_f32x4:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vshuff32x4 {{.*#+}} zmm0 = zmm0[0,1,2,3,4,5,6,7],zmm1[0,1,2,3,0,1,2,3]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_shuffle_f32x4:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vshuff32x4 {{.*#+}} zmm0 = zmm0[0,1,2,3,4,5,6,7],zmm1[0,1,2,3,0,1,2,3]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_shuffle_f32x4:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vshuff32x4 {{.*#+}} zmm0 = zmm0[0,1,2,3,4,5,6,7],zmm1[0,1,2,3,0,1,2,3]
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%shuffle = shufflevector <16 x float> %__A, <16 x float> %__B, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 16, i32 17, i32 18, i32 19, i32 16, i32 17, i32 18, i32 19>
ret <16 x float> %shuffle
@@ -172,11 +167,11 @@ entry:
define <16 x float> @test_mm512_mask_shuffle_f32x4(<16 x float> %__W, i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B) {
-; X32-LABEL: test_mm512_mask_shuffle_f32x4:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vshuff32x4 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,4,5,6,7],zmm2[0,1,2,3,0,1,2,3]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_shuffle_f32x4:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vshuff32x4 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,4,5,6,7],zmm2[0,1,2,3,0,1,2,3]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_shuffle_f32x4:
; X64: # %bb.0: # %entry
@@ -191,11 +186,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_shuffle_f32x4(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B) {
-; X32-LABEL: test_mm512_maskz_shuffle_f32x4:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vshuff32x4 {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,3,4,5,6,7],zmm1[0,1,2,3,0,1,2,3]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_shuffle_f32x4:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vshuff32x4 {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,3,4,5,6,7],zmm1[0,1,2,3,0,1,2,3]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_shuffle_f32x4:
; X64: # %bb.0: # %entry
@@ -210,27 +205,22 @@ entry:
}
define <8 x double> @test_mm512_shuffle_f64x2(<8 x double> %__A, <8 x double> %__B) {
-; X32-LABEL: test_mm512_shuffle_f64x2:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vshuff64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_shuffle_f64x2:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vshuff64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_shuffle_f64x2:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vshuff64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%shuffle = shufflevector <8 x double> %__A, <8 x double> %__B, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 8, i32 9>
ret <8 x double> %shuffle
}
define <8 x double> @test_mm512_mask_shuffle_f64x2(<8 x double> %__W, i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B) {
-; X32-LABEL: test_mm512_mask_shuffle_f64x2:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vshuff64x2 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3],zmm2[0,1,0,1]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_shuffle_f64x2:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vshuff64x2 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3],zmm2[0,1,0,1]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_shuffle_f64x2:
; X64: # %bb.0: # %entry
@@ -245,12 +235,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_shuffle_f64x2(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B) {
-; X32-LABEL: test_mm512_maskz_shuffle_f64x2:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vshuff64x2 {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,3],zmm1[0,1,0,1]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_shuffle_f64x2:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vshuff64x2 {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,3],zmm1[0,1,0,1]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_shuffle_f64x2:
; X64: # %bb.0: # %entry
@@ -265,26 +255,21 @@ entry:
}
define <8 x i64> @test_mm512_shuffle_i32x4(<8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
-; X32-LABEL: test_mm512_shuffle_i32x4:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_shuffle_i32x4:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_shuffle_i32x4:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%shuffle = shufflevector <8 x i64> %__A, <8 x i64> %__B, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 8, i32 9>
ret <8 x i64> %shuffle
}
define <8 x i64> @test_mm512_mask_shuffle_i32x4(<8 x i64> %__W, i16 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
-; X32-LABEL: test_mm512_mask_shuffle_i32x4:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vshufi32x4 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,4,5,6,7],zmm2[0,1,2,3,0,1,2,3]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_shuffle_i32x4:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vshufi32x4 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3,4,5,6,7],zmm2[0,1,2,3,0,1,2,3]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_shuffle_i32x4:
; X64: # %bb.0: # %entry
@@ -302,11 +287,11 @@ entry:
}
define <8 x i64> @test_mm512_maskz_shuffle_i32x4(i16 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
-; X32-LABEL: test_mm512_maskz_shuffle_i32x4:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vshufi32x4 {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,3,4,5,6,7],zmm1[0,1,2,3,0,1,2,3]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_shuffle_i32x4:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vshufi32x4 {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,3,4,5,6,7],zmm1[0,1,2,3,0,1,2,3]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_shuffle_i32x4:
; X64: # %bb.0: # %entry
@@ -323,27 +308,22 @@ entry:
}
define <8 x i64> @test_mm512_shuffle_i64x2(<8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
-; X32-LABEL: test_mm512_shuffle_i64x2:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_shuffle_i64x2:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_shuffle_i64x2:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vshufi64x2 {{.*#+}} zmm0 = zmm0[0,1,2,3],zmm1[0,1,0,1]
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%shuffle = shufflevector <8 x i64> %__A, <8 x i64> %__B, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 8, i32 9>
ret <8 x i64> %shuffle
}
define <8 x i64> @test_mm512_mask_shuffle_i64x2(<8 x i64> %__W, i8 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
-; X32-LABEL: test_mm512_mask_shuffle_i64x2:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vshufi64x2 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3],zmm2[0,1,0,1]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_shuffle_i64x2:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vshufi64x2 {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,3],zmm2[0,1,0,1]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_shuffle_i64x2:
; X64: # %bb.0: # %entry
@@ -358,12 +338,12 @@ entry:
}
define <8 x i64> @test_mm512_maskz_shuffle_i64x2(i8 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B) local_unnamed_addr #0 {
-; X32-LABEL: test_mm512_maskz_shuffle_i64x2:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vshufi64x2 {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,3],zmm1[0,1,0,1]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_shuffle_i64x2:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vshufi64x2 {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,3],zmm1[0,1,0,1]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_shuffle_i64x2:
; X64: # %bb.0: # %entry
@@ -379,21 +359,13 @@ entry:
define zeroext i16 @test_mm512_testn_epi32_mask(<8 x i64> %__A, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_testn_epi32_mask:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vptestnmd %zmm0, %zmm1, %k0
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: movzwl %ax, %eax
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_testn_epi32_mask:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vptestnmd %zmm0, %zmm1, %k0
-; X64-NEXT: kmovw %k0, %eax
-; X64-NEXT: movzwl %ax, %eax
-; X64-NEXT: vzeroupper
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_testn_epi32_mask:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vptestnmd %zmm0, %zmm1, %k0
+; CHECK-NEXT: kmovw %k0, %eax
+; CHECK-NEXT: movzwl %ax, %eax
+; CHECK-NEXT: vzeroupper
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%and1.i.i = and <8 x i64> %__B, %__A
%0 = bitcast <8 x i64> %and1.i.i to <16 x i32>
@@ -403,14 +375,14 @@ entry:
}
define zeroext i16 @test_mm512_mask_testn_epi32_mask(i16 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_mask_testn_epi32_mask:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vptestnmd %zmm0, %zmm1, %k0 {%k1}
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: movzwl %ax, %eax
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_testn_epi32_mask:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vptestnmd %zmm0, %zmm1, %k0 {%k1}
+; X86-NEXT: kmovw %k0, %eax
+; X86-NEXT: movzwl %ax, %eax
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_testn_epi32_mask:
; X64: # %bb.0: # %entry
@@ -431,21 +403,13 @@ entry:
}
define zeroext i8 @test_mm512_testn_epi64_mask(<8 x i64> %__A, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_testn_epi64_mask:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vptestnmq %zmm0, %zmm1, %k0
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: movzbl %al, %eax
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_testn_epi64_mask:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vptestnmq %zmm0, %zmm1, %k0
-; X64-NEXT: kmovw %k0, %eax
-; X64-NEXT: movzbl %al, %eax
-; X64-NEXT: vzeroupper
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_testn_epi64_mask:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vptestnmq %zmm0, %zmm1, %k0
+; CHECK-NEXT: kmovw %k0, %eax
+; CHECK-NEXT: movzbl %al, %eax
+; CHECK-NEXT: vzeroupper
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%and1.i.i = and <8 x i64> %__B, %__A
%0 = icmp eq <8 x i64> %and1.i.i, zeroinitializer
@@ -454,15 +418,15 @@ entry:
}
define zeroext i8 @test_mm512_mask_testn_epi64_mask(i8 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_mask_testn_epi64_mask:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vptestnmq %zmm0, %zmm1, %k0 {%k1}
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: movzbl %al, %eax
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_testn_epi64_mask:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vptestnmq %zmm0, %zmm1, %k0 {%k1}
+; X86-NEXT: kmovw %k0, %eax
+; X86-NEXT: movzbl %al, %eax
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_testn_epi64_mask:
; X64: # %bb.0: # %entry
@@ -482,14 +446,14 @@ entry:
}
define zeroext i16 @test_mm512_mask_test_epi32_mask(i16 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_mask_test_epi32_mask:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vptestmd %zmm0, %zmm1, %k0 {%k1}
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: movzwl %ax, %eax
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_test_epi32_mask:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vptestmd %zmm0, %zmm1, %k0 {%k1}
+; X86-NEXT: kmovw %k0, %eax
+; X86-NEXT: movzwl %ax, %eax
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_test_epi32_mask:
; X64: # %bb.0: # %entry
@@ -510,15 +474,15 @@ entry:
}
define zeroext i8 @test_mm512_mask_test_epi64_mask(i8 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_mask_test_epi64_mask:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vptestmq %zmm0, %zmm1, %k0 {%k1}
-; X32-NEXT: kmovw %k0, %eax
-; X32-NEXT: movzbl %al, %eax
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_test_epi64_mask:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vptestmq %zmm0, %zmm1, %k0 {%k1}
+; X86-NEXT: kmovw %k0, %eax
+; X86-NEXT: movzbl %al, %eax
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_test_epi64_mask:
; X64: # %bb.0: # %entry
@@ -538,12 +502,12 @@ entry:
}
define <8 x i64> @test_mm512_mask_set1_epi32(<8 x i64> %__O, i16 zeroext %__M, i32 %__A) {
-; X32-LABEL: test_mm512_mask_set1_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpbroadcastd %eax, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_set1_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpbroadcastd %eax, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_set1_epi32:
; X64: # %bb.0: # %entry
@@ -561,12 +525,12 @@ entry:
}
define <8 x i64> @test_mm512_maskz_set1_epi32(i16 zeroext %__M, i32 %__A) {
-; X32-LABEL: test_mm512_maskz_set1_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpbroadcastd %eax, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_set1_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpbroadcastd %eax, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_set1_epi32:
; X64: # %bb.0: # %entry
@@ -583,14 +547,14 @@ entry:
}
define <8 x i64> @test_mm512_mask_set1_epi64(<8 x i64> %__O, i8 zeroext %__M, i64 %__A) {
-; X32-LABEL: test_mm512_mask_set1_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; X32-NEXT: vpinsrd $1, {{[0-9]+}}(%esp), %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpbroadcastq %xmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_set1_epi64:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X86-NEXT: vpinsrd $1, {{[0-9]+}}(%esp), %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpbroadcastq %xmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_set1_epi64:
; X64: # %bb.0: # %entry
@@ -606,14 +570,14 @@ entry:
}
define <8 x i64> @test_mm512_maskz_set1_epi64(i8 zeroext %__M, i64 %__A) {
-; X32-LABEL: test_mm512_maskz_set1_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
-; X32-NEXT: vpinsrd $1, {{[0-9]+}}(%esp), %xmm0, %xmm0
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpbroadcastq %xmm0, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_set1_epi64:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero
+; X86-NEXT: vpinsrd $1, {{[0-9]+}}(%esp), %xmm0, %xmm0
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpbroadcastq %xmm0, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_set1_epi64:
; X64: # %bb.0: # %entry
@@ -630,15 +594,10 @@ entry:
define <8 x i64> @test_mm512_broadcastd_epi32(<2 x i64> %a0) {
-; X32-LABEL: test_mm512_broadcastd_epi32:
-; X32: # %bb.0:
-; X32-NEXT: vbroadcastss %xmm0, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_broadcastd_epi32:
-; X64: # %bb.0:
-; X64-NEXT: vbroadcastss %xmm0, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_broadcastd_epi32:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vbroadcastss %xmm0, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
%arg0 = bitcast <2 x i64> %a0 to <4 x i32>
%res0 = shufflevector <4 x i32> %arg0, <4 x i32> undef, <16 x i32> zeroinitializer
%res1 = bitcast <16 x i32> %res0 to <8 x i64>
@@ -646,11 +605,11 @@ define <8 x i64> @test_mm512_broadcastd_
}
define <8 x i64> @test_mm512_mask_broadcastd_epi32(<8 x i64> %a0, i16 %a1, <2 x i64> %a2) {
-; X32-LABEL: test_mm512_mask_broadcastd_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpbroadcastd %xmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_broadcastd_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpbroadcastd %xmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_broadcastd_epi32:
; X64: # %bb.0:
@@ -667,11 +626,11 @@ define <8 x i64> @test_mm512_mask_broadc
}
define <8 x i64> @test_mm512_maskz_broadcastd_epi32(i16 %a0, <2 x i64> %a1) {
-; X32-LABEL: test_mm512_maskz_broadcastd_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpbroadcastd %xmm0, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_broadcastd_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpbroadcastd %xmm0, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_broadcastd_epi32:
; X64: # %bb.0:
@@ -687,26 +646,21 @@ define <8 x i64> @test_mm512_maskz_broad
}
define <8 x i64> @test_mm512_broadcastq_epi64(<2 x i64> %a0) {
-; X32-LABEL: test_mm512_broadcastq_epi64:
-; X32: # %bb.0:
-; X32-NEXT: vbroadcastsd %xmm0, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_broadcastq_epi64:
-; X64: # %bb.0:
-; X64-NEXT: vbroadcastsd %xmm0, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_broadcastq_epi64:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vbroadcastsd %xmm0, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <2 x i64> %a0, <2 x i64> undef, <8 x i32> zeroinitializer
ret <8 x i64> %res
}
define <8 x i64> @test_mm512_mask_broadcastq_epi64(<8 x i64> %a0, i8 %a1, <2 x i64> %a2) {
-; X32-LABEL: test_mm512_mask_broadcastq_epi64:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpbroadcastq %xmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_broadcastq_epi64:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpbroadcastq %xmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_broadcastq_epi64:
; X64: # %bb.0:
@@ -720,12 +674,12 @@ define <8 x i64> @test_mm512_mask_broadc
}
define <8 x i64> @test_mm512_maskz_broadcastq_epi64(i8 %a0, <2 x i64> %a1) {
-; X32-LABEL: test_mm512_maskz_broadcastq_epi64:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpbroadcastq %xmm0, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_broadcastq_epi64:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpbroadcastq %xmm0, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_broadcastq_epi64:
; X64: # %bb.0:
@@ -739,26 +693,21 @@ define <8 x i64> @test_mm512_maskz_broad
}
define <8 x double> @test_mm512_broadcastsd_pd(<2 x double> %a0) {
-; X32-LABEL: test_mm512_broadcastsd_pd:
-; X32: # %bb.0:
-; X32-NEXT: vbroadcastsd %xmm0, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_broadcastsd_pd:
-; X64: # %bb.0:
-; X64-NEXT: vbroadcastsd %xmm0, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_broadcastsd_pd:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vbroadcastsd %xmm0, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <2 x double> %a0, <2 x double> undef, <8 x i32> zeroinitializer
ret <8 x double> %res
}
define <8 x double> @test_mm512_mask_broadcastsd_pd(<8 x double> %a0, i8 %a1, <2 x double> %a2) {
-; X32-LABEL: test_mm512_mask_broadcastsd_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vbroadcastsd %xmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_broadcastsd_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vbroadcastsd %xmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_broadcastsd_pd:
; X64: # %bb.0:
@@ -772,12 +721,12 @@ define <8 x double> @test_mm512_mask_bro
}
define <8 x double> @test_mm512_maskz_broadcastsd_pd(i8 %a0, <2 x double> %a1) {
-; X32-LABEL: test_mm512_maskz_broadcastsd_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vbroadcastsd %xmm0, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_broadcastsd_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vbroadcastsd %xmm0, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_broadcastsd_pd:
; X64: # %bb.0:
@@ -791,25 +740,20 @@ define <8 x double> @test_mm512_maskz_br
}
define <16 x float> @test_mm512_broadcastss_ps(<4 x float> %a0) {
-; X32-LABEL: test_mm512_broadcastss_ps:
-; X32: # %bb.0:
-; X32-NEXT: vbroadcastss %xmm0, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_broadcastss_ps:
-; X64: # %bb.0:
-; X64-NEXT: vbroadcastss %xmm0, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_broadcastss_ps:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vbroadcastss %xmm0, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <4 x float> %a0, <4 x float> undef, <16 x i32> zeroinitializer
ret <16 x float> %res
}
define <16 x float> @test_mm512_mask_broadcastss_ps(<16 x float> %a0, i16 %a1, <4 x float> %a2) {
-; X32-LABEL: test_mm512_mask_broadcastss_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vbroadcastss %xmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_broadcastss_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vbroadcastss %xmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_broadcastss_ps:
; X64: # %bb.0:
@@ -823,11 +767,11 @@ define <16 x float> @test_mm512_mask_bro
}
define <16 x float> @test_mm512_maskz_broadcastss_ps(i16 %a0, <4 x float> %a1) {
-; X32-LABEL: test_mm512_maskz_broadcastss_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vbroadcastss %xmm0, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_broadcastss_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vbroadcastss %xmm0, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_broadcastss_ps:
; X64: # %bb.0:
@@ -841,26 +785,21 @@ define <16 x float> @test_mm512_maskz_br
}
define <8 x double> @test_mm512_movddup_pd(<8 x double> %a0) {
-; X32-LABEL: test_mm512_movddup_pd:
-; X32: # %bb.0:
-; X32-NEXT: vmovddup {{.*#+}} zmm0 = zmm0[0,0,2,2,4,4,6,6]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_movddup_pd:
-; X64: # %bb.0:
-; X64-NEXT: vmovddup {{.*#+}} zmm0 = zmm0[0,0,2,2,4,4,6,6]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_movddup_pd:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovddup {{.*#+}} zmm0 = zmm0[0,0,2,2,4,4,6,6]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x double> %a0, <8 x double> undef, <8 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
ret <8 x double> %res
}
define <8 x double> @test_mm512_mask_movddup_pd(<8 x double> %a0, i8 %a1, <8 x double> %a2) {
-; X32-LABEL: test_mm512_mask_movddup_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vmovddup {{.*#+}} zmm0 {%k1} = zmm1[0,0,2,2,4,4,6,6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_movddup_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vmovddup {{.*#+}} zmm0 {%k1} = zmm1[0,0,2,2,4,4,6,6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_movddup_pd:
; X64: # %bb.0:
@@ -874,12 +813,12 @@ define <8 x double> @test_mm512_mask_mov
}
define <8 x double> @test_mm512_maskz_movddup_pd(i8 %a0, <8 x double> %a1) {
-; X32-LABEL: test_mm512_maskz_movddup_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vmovddup {{.*#+}} zmm0 {%k1} {z} = zmm0[0,0,2,2,4,4,6,6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_movddup_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vmovddup {{.*#+}} zmm0 {%k1} {z} = zmm0[0,0,2,2,4,4,6,6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_movddup_pd:
; X64: # %bb.0:
@@ -893,25 +832,20 @@ define <8 x double> @test_mm512_maskz_mo
}
define <16 x float> @test_mm512_movehdup_ps(<16 x float> %a0) {
-; X32-LABEL: test_mm512_movehdup_ps:
-; X32: # %bb.0:
-; X32-NEXT: vmovshdup {{.*#+}} zmm0 = zmm0[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_movehdup_ps:
-; X64: # %bb.0:
-; X64-NEXT: vmovshdup {{.*#+}} zmm0 = zmm0[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_movehdup_ps:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovshdup {{.*#+}} zmm0 = zmm0[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <16 x float> %a0, <16 x float> undef, <16 x i32> <i32 1, i32 1, i32 3, i32 3, i32 5, i32 5, i32 7, i32 7, i32 9, i32 9, i32 11, i32 11, i32 13, i32 13, i32 15, i32 15>
ret <16 x float> %res
}
define <16 x float> @test_mm512_mask_movehdup_ps(<16 x float> %a0, i16 %a1, <16 x float> %a2) {
-; X32-LABEL: test_mm512_mask_movehdup_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vmovshdup {{.*#+}} zmm0 {%k1} = zmm1[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_movehdup_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vmovshdup {{.*#+}} zmm0 {%k1} = zmm1[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_movehdup_ps:
; X64: # %bb.0:
@@ -925,11 +859,11 @@ define <16 x float> @test_mm512_mask_mov
}
define <16 x float> @test_mm512_maskz_movehdup_ps(i16 %a0, <16 x float> %a1) {
-; X32-LABEL: test_mm512_maskz_movehdup_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vmovshdup {{.*#+}} zmm0 {%k1} {z} = zmm0[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_movehdup_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vmovshdup {{.*#+}} zmm0 {%k1} {z} = zmm0[1,1,3,3,5,5,7,7,9,9,11,11,13,13,15,15]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_movehdup_ps:
; X64: # %bb.0:
@@ -943,25 +877,20 @@ define <16 x float> @test_mm512_maskz_mo
}
define <16 x float> @test_mm512_moveldup_ps(<16 x float> %a0) {
-; X32-LABEL: test_mm512_moveldup_ps:
-; X32: # %bb.0:
-; X32-NEXT: vmovsldup {{.*#+}} zmm0 = zmm0[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_moveldup_ps:
-; X64: # %bb.0:
-; X64-NEXT: vmovsldup {{.*#+}} zmm0 = zmm0[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_moveldup_ps:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovsldup {{.*#+}} zmm0 = zmm0[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <16 x float> %a0, <16 x float> undef, <16 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14>
ret <16 x float> %res
}
define <16 x float> @test_mm512_mask_moveldup_ps(<16 x float> %a0, i16 %a1, <16 x float> %a2) {
-; X32-LABEL: test_mm512_mask_moveldup_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vmovsldup {{.*#+}} zmm0 {%k1} = zmm1[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_moveldup_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vmovsldup {{.*#+}} zmm0 {%k1} = zmm1[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_moveldup_ps:
; X64: # %bb.0:
@@ -975,11 +904,11 @@ define <16 x float> @test_mm512_mask_mov
}
define <16 x float> @test_mm512_maskz_moveldup_ps(i16 %a0, <16 x float> %a1) {
-; X32-LABEL: test_mm512_maskz_moveldup_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vmovsldup {{.*#+}} zmm0 {%k1} {z} = zmm0[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_moveldup_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vmovsldup {{.*#+}} zmm0 {%k1} {z} = zmm0[0,0,2,2,4,4,6,6,8,8,10,10,12,12,14,14]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_moveldup_ps:
; X64: # %bb.0:
@@ -993,26 +922,21 @@ define <16 x float> @test_mm512_maskz_mo
}
define <8 x double> @test_mm512_permute_pd(<8 x double> %a0) {
-; X32-LABEL: test_mm512_permute_pd:
-; X32: # %bb.0:
-; X32-NEXT: vpermilpd {{.*#+}} zmm0 = zmm0[0,1,2,2,4,4,6,6]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_permute_pd:
-; X64: # %bb.0:
-; X64-NEXT: vpermilpd {{.*#+}} zmm0 = zmm0[0,1,2,2,4,4,6,6]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_permute_pd:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vpermilpd {{.*#+}} zmm0 = zmm0[0,1,2,2,4,4,6,6]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x double> %a0, <8 x double> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6>
ret <8 x double> %res
}
define <8 x double> @test_mm512_mask_permute_pd(<8 x double> %a0, i8 %a1, <8 x double> %a2) {
-; X32-LABEL: test_mm512_mask_permute_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermilpd {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,2,4,4,6,6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_permute_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermilpd {{.*#+}} zmm0 {%k1} = zmm1[0,1,2,2,4,4,6,6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_permute_pd:
; X64: # %bb.0:
@@ -1026,12 +950,12 @@ define <8 x double> @test_mm512_mask_per
}
define <8 x double> @test_mm512_maskz_permute_pd(i8 %a0, <8 x double> %a1) {
-; X32-LABEL: test_mm512_maskz_permute_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermilpd {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,2,4,4,6,6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_permute_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermilpd {{.*#+}} zmm0 {%k1} {z} = zmm0[0,1,2,2,4,4,6,6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_permute_pd:
; X64: # %bb.0:
@@ -1045,25 +969,20 @@ define <8 x double> @test_mm512_maskz_pe
}
define <16 x float> @test_mm512_permute_ps(<16 x float> %a0) {
-; X32-LABEL: test_mm512_permute_ps:
-; X32: # %bb.0:
-; X32-NEXT: vpermilps {{.*#+}} zmm0 = zmm0[2,0,0,0,6,4,4,4,10,8,8,8,14,12,12,12]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_permute_ps:
-; X64: # %bb.0:
-; X64-NEXT: vpermilps {{.*#+}} zmm0 = zmm0[2,0,0,0,6,4,4,4,10,8,8,8,14,12,12,12]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_permute_ps:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vpermilps {{.*#+}} zmm0 = zmm0[2,0,0,0,6,4,4,4,10,8,8,8,14,12,12,12]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <16 x float> %a0, <16 x float> undef, <16 x i32> <i32 2, i32 0, i32 0, i32 0, i32 6, i32 4, i32 4, i32 4, i32 10, i32 8, i32 8, i32 8, i32 14, i32 12, i32 12, i32 12>
ret <16 x float> %res
}
define <16 x float> @test_mm512_mask_permute_ps(<16 x float> %a0, i16 %a1, <16 x float> %a2) {
-; X32-LABEL: test_mm512_mask_permute_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpermilps {{.*#+}} zmm0 {%k1} = zmm1[2,0,0,0,6,4,4,4,10,8,8,8,14,12,12,12]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_permute_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpermilps {{.*#+}} zmm0 {%k1} = zmm1[2,0,0,0,6,4,4,4,10,8,8,8,14,12,12,12]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_permute_ps:
; X64: # %bb.0:
@@ -1077,11 +996,11 @@ define <16 x float> @test_mm512_mask_per
}
define <16 x float> @test_mm512_maskz_permute_ps(i16 %a0, <16 x float> %a1) {
-; X32-LABEL: test_mm512_maskz_permute_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpermilps {{.*#+}} zmm0 {%k1} {z} = zmm0[2,0,0,0,6,4,4,4,10,8,8,8,14,12,12,12]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_permute_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpermilps {{.*#+}} zmm0 {%k1} {z} = zmm0[2,0,0,0,6,4,4,4,10,8,8,8,14,12,12,12]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_permute_ps:
; X64: # %bb.0:
@@ -1095,26 +1014,21 @@ define <16 x float> @test_mm512_maskz_pe
}
define <8 x i64> @test_mm512_permutex_epi64(<8 x i64> %a0) {
-; X32-LABEL: test_mm512_permutex_epi64:
-; X32: # %bb.0:
-; X32-NEXT: vpermpd {{.*#+}} zmm0 = zmm0[0,0,0,0,4,4,4,4]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_permutex_epi64:
-; X64: # %bb.0:
-; X64-NEXT: vpermpd {{.*#+}} zmm0 = zmm0[0,0,0,0,4,4,4,4]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_permutex_epi64:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vpermpd {{.*#+}} zmm0 = zmm0[0,0,0,0,4,4,4,4]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x i64> %a0, <8 x i64> undef, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4>
ret <8 x i64> %res
}
define <8 x i64> @test_mm512_mask_permutex_epi64(<8 x i64> %a0, i8 %a1, <8 x i64> %a2) {
-; X32-LABEL: test_mm512_mask_permutex_epi64:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermq {{.*#+}} zmm0 {%k1} = zmm1[0,0,0,0,4,4,4,4]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_permutex_epi64:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermq {{.*#+}} zmm0 {%k1} = zmm1[0,0,0,0,4,4,4,4]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_permutex_epi64:
; X64: # %bb.0:
@@ -1128,12 +1042,12 @@ define <8 x i64> @test_mm512_mask_permut
}
define <8 x i64> @test_mm512_maskz_permutex_epi64(i8 %a0, <8 x i64> %a1) {
-; X32-LABEL: test_mm512_maskz_permutex_epi64:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermq {{.*#+}} zmm0 {%k1} {z} = zmm0[0,0,0,0,4,4,4,4]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_permutex_epi64:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermq {{.*#+}} zmm0 {%k1} {z} = zmm0[0,0,0,0,4,4,4,4]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_permutex_epi64:
; X64: # %bb.0:
@@ -1147,26 +1061,21 @@ define <8 x i64> @test_mm512_maskz_permu
}
define <8 x double> @test_mm512_permutex_pd(<8 x double> %a0) {
-; X32-LABEL: test_mm512_permutex_pd:
-; X32: # %bb.0:
-; X32-NEXT: vpermpd {{.*#+}} zmm0 = zmm0[0,0,0,0,4,4,4,4]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_permutex_pd:
-; X64: # %bb.0:
-; X64-NEXT: vpermpd {{.*#+}} zmm0 = zmm0[0,0,0,0,4,4,4,4]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_permutex_pd:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vpermpd {{.*#+}} zmm0 = zmm0[0,0,0,0,4,4,4,4]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x double> %a0, <8 x double> undef, <8 x i32> <i32 0, i32 0, i32 0, i32 0, i32 4, i32 4, i32 4, i32 4>
ret <8 x double> %res
}
define <8 x double> @test_mm512_mask_permutex_pd(<8 x double> %a0, i8 %a1, <8 x double> %a2) {
-; X32-LABEL: test_mm512_mask_permutex_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermpd {{.*#+}} zmm0 {%k1} = zmm1[0,0,0,0,4,4,4,4]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_permutex_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermpd {{.*#+}} zmm0 {%k1} = zmm1[0,0,0,0,4,4,4,4]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_permutex_pd:
; X64: # %bb.0:
@@ -1180,12 +1089,12 @@ define <8 x double> @test_mm512_mask_per
}
define <8 x double> @test_mm512_maskz_permutex_pd(i8 %a0, <8 x double> %a1) {
-; X32-LABEL: test_mm512_maskz_permutex_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermpd {{.*#+}} zmm0 {%k1} {z} = zmm0[0,0,0,0,4,4,4,4]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_permutex_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermpd {{.*#+}} zmm0 {%k1} {z} = zmm0[0,0,0,0,4,4,4,4]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_permutex_pd:
; X64: # %bb.0:
@@ -1199,15 +1108,10 @@ define <8 x double> @test_mm512_maskz_pe
}
define <8 x i64> @test_mm512_shuffle_epi32(<8 x i64> %a0) {
-; X32-LABEL: test_mm512_shuffle_epi32:
-; X32: # %bb.0:
-; X32-NEXT: vpermilps {{.*#+}} zmm0 = zmm0[1,0,0,0,5,4,4,4,9,8,8,8,13,12,12,12]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_shuffle_epi32:
-; X64: # %bb.0:
-; X64-NEXT: vpermilps {{.*#+}} zmm0 = zmm0[1,0,0,0,5,4,4,4,9,8,8,8,13,12,12,12]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_shuffle_epi32:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vpermilps {{.*#+}} zmm0 = zmm0[1,0,0,0,5,4,4,4,9,8,8,8,13,12,12,12]
+; CHECK-NEXT: ret{{[l|q]}}
%arg0 = bitcast <8 x i64> %a0 to <16 x i32>
%res0 = shufflevector <16 x i32> %arg0, <16 x i32> undef, <16 x i32> <i32 1, i32 0, i32 0, i32 0, i32 5, i32 4, i32 4, i32 4, i32 9, i32 8, i32 8, i32 8, i32 13, i32 12, i32 12, i32 12>
%res1 = bitcast <16 x i32> %res0 to <8 x i64>
@@ -1215,11 +1119,11 @@ define <8 x i64> @test_mm512_shuffle_epi
}
define <8 x i64> @test_mm512_mask_shuffle_epi32(<8 x i64> %a0, i16 %a1, <8 x i64> %a2) {
-; X32-LABEL: test_mm512_mask_shuffle_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpshufd {{.*#+}} zmm0 {%k1} = zmm1[1,0,0,0,5,4,4,4,9,8,8,8,13,12,12,12]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_shuffle_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpshufd {{.*#+}} zmm0 {%k1} = zmm1[1,0,0,0,5,4,4,4,9,8,8,8,13,12,12,12]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_shuffle_epi32:
; X64: # %bb.0:
@@ -1236,11 +1140,11 @@ define <8 x i64> @test_mm512_mask_shuffl
}
define <8 x i64> @test_mm512_maskz_shuffle_epi32(i16 %a0, <8 x i64> %a1) {
-; X32-LABEL: test_mm512_maskz_shuffle_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpshufd {{.*#+}} zmm0 {%k1} {z} = zmm0[1,0,0,0,5,4,4,4,9,8,8,8,13,12,12,12]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_shuffle_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpshufd {{.*#+}} zmm0 {%k1} {z} = zmm0[1,0,0,0,5,4,4,4,9,8,8,8,13,12,12,12]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_shuffle_epi32:
; X64: # %bb.0:
@@ -1256,26 +1160,21 @@ define <8 x i64> @test_mm512_maskz_shuff
}
define <8 x double> @test_mm512_shuffle_pd(<8 x double> %a0, <8 x double> %a1) {
-; X32-LABEL: test_mm512_shuffle_pd:
-; X32: # %bb.0:
-; X32-NEXT: vshufpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[3],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_shuffle_pd:
-; X64: # %bb.0:
-; X64-NEXT: vshufpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[3],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_shuffle_pd:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vshufpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[3],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x double> %a0, <8 x double> %a1, <8 x i32> <i32 0, i32 8, i32 3, i32 10, i32 4, i32 12, i32 6, i32 14>
ret <8 x double> %res
}
define <8 x double> @test_mm512_mask_shuffle_pd(<8 x double> %a0, i8 %a1, <8 x double> %a2, <8 x double> %a3) {
-; X32-LABEL: test_mm512_mask_shuffle_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vshufpd {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[3],zmm2[2],zmm1[4],zmm2[4],zmm1[6],zmm2[6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_shuffle_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vshufpd {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[3],zmm2[2],zmm1[4],zmm2[4],zmm1[6],zmm2[6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_shuffle_pd:
; X64: # %bb.0:
@@ -1289,12 +1188,12 @@ define <8 x double> @test_mm512_mask_shu
}
define <8 x double> @test_mm512_maskz_shuffle_pd(i8 %a0, <8 x double> %a1, <8 x double> %a2) {
-; X32-LABEL: test_mm512_maskz_shuffle_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vshufpd {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[3],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_shuffle_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vshufpd {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[3],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_shuffle_pd:
; X64: # %bb.0:
@@ -1308,15 +1207,10 @@ define <8 x double> @test_mm512_maskz_sh
}
define <8 x i64> @test_mm512_unpackhi_epi32(<8 x i64> %a0, <8 x i64> %a1) {
-; X32-LABEL: test_mm512_unpackhi_epi32:
-; X32: # %bb.0:
-; X32-NEXT: vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_unpackhi_epi32:
-; X64: # %bb.0:
-; X64-NEXT: vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_unpackhi_epi32:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; CHECK-NEXT: ret{{[l|q]}}
%arg0 = bitcast <8 x i64> %a0 to <16 x i32>
%arg1 = bitcast <8 x i64> %a1 to <16 x i32>
%res0 = shufflevector <16 x i32> %arg0, <16 x i32> %arg1, <16 x i32> <i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31>
@@ -1325,11 +1219,11 @@ define <8 x i64> @test_mm512_unpackhi_ep
}
define <8 x i64> @test_mm512_mask_unpackhi_epi32(<8 x i64> %a0, i16 %a1, <8 x i64> %a2, <8 x i64> %a3) {
-; X32-LABEL: test_mm512_mask_unpackhi_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpunpckhdq {{.*#+}} zmm0 {%k1} = zmm1[2],zmm2[2],zmm1[3],zmm2[3],zmm1[6],zmm2[6],zmm1[7],zmm2[7],zmm1[10],zmm2[10],zmm1[11],zmm2[11],zmm1[14],zmm2[14],zmm1[15],zmm2[15]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_unpackhi_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpunpckhdq {{.*#+}} zmm0 {%k1} = zmm1[2],zmm2[2],zmm1[3],zmm2[3],zmm1[6],zmm2[6],zmm1[7],zmm2[7],zmm1[10],zmm2[10],zmm1[11],zmm2[11],zmm1[14],zmm2[14],zmm1[15],zmm2[15]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_unpackhi_epi32:
; X64: # %bb.0:
@@ -1347,11 +1241,11 @@ define <8 x i64> @test_mm512_mask_unpack
}
define <8 x i64> @test_mm512_maskz_unpackhi_epi32(i16 %a0, <8 x i64> %a1, <8 x i64> %a2) {
-; X32-LABEL: test_mm512_maskz_unpackhi_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpunpckhdq {{.*#+}} zmm0 {%k1} {z} = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_unpackhi_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpunpckhdq {{.*#+}} zmm0 {%k1} {z} = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_unpackhi_epi32:
; X64: # %bb.0:
@@ -1368,26 +1262,21 @@ define <8 x i64> @test_mm512_maskz_unpac
}
define <8 x i64> @test_mm512_unpackhi_epi64(<8 x i64> %a0, <8 x i64> %a1) {
-; X32-LABEL: test_mm512_unpackhi_epi64:
-; X32: # %bb.0:
-; X32-NEXT: vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_unpackhi_epi64:
-; X64: # %bb.0:
-; X64-NEXT: vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_unpackhi_epi64:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x i64> %a0, <8 x i64> %a1, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15>
ret <8 x i64> %res
}
define <8 x i64> @test_mm512_mask_unpackhi_epi64(<8 x i64> %a0, i8 %a1, <8 x i64> %a2, <8 x i64> %a3) {
-; X32-LABEL: test_mm512_mask_unpackhi_epi64:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpunpckhqdq {{.*#+}} zmm0 {%k1} = zmm1[1],zmm2[1],zmm1[3],zmm2[3],zmm1[5],zmm2[5],zmm1[7],zmm2[7]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_unpackhi_epi64:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpunpckhqdq {{.*#+}} zmm0 {%k1} = zmm1[1],zmm2[1],zmm1[3],zmm2[3],zmm1[5],zmm2[5],zmm1[7],zmm2[7]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_unpackhi_epi64:
; X64: # %bb.0:
@@ -1401,12 +1290,12 @@ define <8 x i64> @test_mm512_mask_unpack
}
define <8 x i64> @test_mm512_maskz_unpackhi_epi64(i8 %a0, <8 x i64> %a1, <8 x i64> %a2) {
-; X32-LABEL: test_mm512_maskz_unpackhi_epi64:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpunpckhqdq {{.*#+}} zmm0 {%k1} {z} = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_unpackhi_epi64:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpunpckhqdq {{.*#+}} zmm0 {%k1} {z} = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_unpackhi_epi64:
; X64: # %bb.0:
@@ -1420,26 +1309,21 @@ define <8 x i64> @test_mm512_maskz_unpac
}
define <8 x double> @test_mm512_unpackhi_pd(<8 x double> %a0, <8 x double> %a1) {
-; X32-LABEL: test_mm512_unpackhi_pd:
-; X32: # %bb.0:
-; X32-NEXT: vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_unpackhi_pd:
-; X64: # %bb.0:
-; X64-NEXT: vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_unpackhi_pd:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vunpckhpd {{.*#+}} zmm0 = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x double> %a0, <8 x double> %a1, <8 x i32> <i32 1, i32 9, i32 3, i32 11, i32 5, i32 13, i32 7, i32 15>
ret <8 x double> %res
}
define <8 x double> @test_mm512_mask_unpackhi_pd(<8 x double> %a0, i8 %a1, <8 x double> %a2, <8 x double> %a3) {
-; X32-LABEL: test_mm512_mask_unpackhi_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vunpckhpd {{.*#+}} zmm0 {%k1} = zmm1[1],zmm2[1],zmm1[3],zmm2[3],zmm1[5],zmm2[5],zmm1[7],zmm2[7]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_unpackhi_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vunpckhpd {{.*#+}} zmm0 {%k1} = zmm1[1],zmm2[1],zmm1[3],zmm2[3],zmm1[5],zmm2[5],zmm1[7],zmm2[7]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_unpackhi_pd:
; X64: # %bb.0:
@@ -1453,12 +1337,12 @@ define <8 x double> @test_mm512_mask_unp
}
define <8 x double> @test_mm512_maskz_unpackhi_pd(i8 %a0, <8 x double> %a1, <8 x double> %a2) {
-; X32-LABEL: test_mm512_maskz_unpackhi_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vunpckhpd {{.*#+}} zmm0 {%k1} {z} = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_unpackhi_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vunpckhpd {{.*#+}} zmm0 {%k1} {z} = zmm0[1],zmm1[1],zmm0[3],zmm1[3],zmm0[5],zmm1[5],zmm0[7],zmm1[7]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_unpackhi_pd:
; X64: # %bb.0:
@@ -1472,25 +1356,20 @@ define <8 x double> @test_mm512_maskz_un
}
define <16 x float> @test_mm512_unpackhi_ps(<16 x float> %a0, <16 x float> %a1) {
-; X32-LABEL: test_mm512_unpackhi_ps:
-; X32: # %bb.0:
-; X32-NEXT: vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_unpackhi_ps:
-; X64: # %bb.0:
-; X64-NEXT: vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_unpackhi_ps:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vunpckhps {{.*#+}} zmm0 = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <16 x float> %a0, <16 x float> %a1, <16 x i32> <i32 2, i32 18, i32 3, i32 19, i32 6, i32 22, i32 7, i32 23, i32 10, i32 26, i32 11, i32 27, i32 14, i32 30, i32 15, i32 31>
ret <16 x float> %res
}
define <16 x float> @test_mm512_mask_unpackhi_ps(<16 x float> %a0, i16 %a1, <16 x float> %a2, <16 x float> %a3) {
-; X32-LABEL: test_mm512_mask_unpackhi_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vunpckhps {{.*#+}} zmm0 {%k1} = zmm1[2],zmm2[2],zmm1[3],zmm2[3],zmm1[6],zmm2[6],zmm1[7],zmm2[7],zmm1[10],zmm2[10],zmm1[11],zmm2[11],zmm1[14],zmm2[14],zmm1[15],zmm2[15]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_unpackhi_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vunpckhps {{.*#+}} zmm0 {%k1} = zmm1[2],zmm2[2],zmm1[3],zmm2[3],zmm1[6],zmm2[6],zmm1[7],zmm2[7],zmm1[10],zmm2[10],zmm1[11],zmm2[11],zmm1[14],zmm2[14],zmm1[15],zmm2[15]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_unpackhi_ps:
; X64: # %bb.0:
@@ -1504,11 +1383,11 @@ define <16 x float> @test_mm512_mask_unp
}
define <16 x float> @test_mm512_maskz_unpackhi_ps(i16 %a0, <16 x float> %a1, <16 x float> %a2) {
-; X32-LABEL: test_mm512_maskz_unpackhi_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vunpckhps {{.*#+}} zmm0 {%k1} {z} = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_unpackhi_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vunpckhps {{.*#+}} zmm0 {%k1} {z} = zmm0[2],zmm1[2],zmm0[3],zmm1[3],zmm0[6],zmm1[6],zmm0[7],zmm1[7],zmm0[10],zmm1[10],zmm0[11],zmm1[11],zmm0[14],zmm1[14],zmm0[15],zmm1[15]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_unpackhi_ps:
; X64: # %bb.0:
@@ -1522,15 +1401,10 @@ define <16 x float> @test_mm512_maskz_un
}
define <8 x i64> @test_mm512_unpacklo_epi32(<8 x i64> %a0, <8 x i64> %a1) {
-; X32-LABEL: test_mm512_unpacklo_epi32:
-; X32: # %bb.0:
-; X32-NEXT: vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_unpacklo_epi32:
-; X64: # %bb.0:
-; X64-NEXT: vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_unpacklo_epi32:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; CHECK-NEXT: ret{{[l|q]}}
%arg0 = bitcast <8 x i64> %a0 to <16 x i32>
%arg1 = bitcast <8 x i64> %a1 to <16 x i32>
%res0 = shufflevector <16 x i32> %arg0, <16 x i32> %arg1, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29>
@@ -1539,11 +1413,11 @@ define <8 x i64> @test_mm512_unpacklo_ep
}
define <8 x i64> @test_mm512_mask_unpacklo_epi32(<8 x i64> %a0, i16 %a1, <8 x i64> %a2, <8 x i64> %a3) {
-; X32-LABEL: test_mm512_mask_unpacklo_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpunpckldq {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[1],zmm2[1],zmm1[4],zmm2[4],zmm1[5],zmm2[5],zmm1[8],zmm2[8],zmm1[9],zmm2[9],zmm1[12],zmm2[12],zmm1[13],zmm2[13]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_unpacklo_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpunpckldq {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[1],zmm2[1],zmm1[4],zmm2[4],zmm1[5],zmm2[5],zmm1[8],zmm2[8],zmm1[9],zmm2[9],zmm1[12],zmm2[12],zmm1[13],zmm2[13]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_unpacklo_epi32:
; X64: # %bb.0:
@@ -1561,11 +1435,11 @@ define <8 x i64> @test_mm512_mask_unpack
}
define <8 x i64> @test_mm512_maskz_unpacklo_epi32(i16 %a0, <8 x i64> %a1, <8 x i64> %a2) {
-; X32-LABEL: test_mm512_maskz_unpacklo_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpunpckldq {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_unpacklo_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpunpckldq {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_unpacklo_epi32:
; X64: # %bb.0:
@@ -1582,26 +1456,21 @@ define <8 x i64> @test_mm512_maskz_unpac
}
define <8 x i64> @test_mm512_unpacklo_epi64(<8 x i64> %a0, <8 x i64> %a1) {
-; X32-LABEL: test_mm512_unpacklo_epi64:
-; X32: # %bb.0:
-; X32-NEXT: vunpcklpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_unpacklo_epi64:
-; X64: # %bb.0:
-; X64-NEXT: vunpcklpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_unpacklo_epi64:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vunpcklpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x i64> %a0, <8 x i64> %a1, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14>
ret <8 x i64> %res
}
define <8 x i64> @test_mm512_mask_unpacklo_epi64(<8 x i64> %a0, i8 %a1, <8 x i64> %a2, <8 x i64> %a3) {
-; X32-LABEL: test_mm512_mask_unpacklo_epi64:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpunpcklqdq {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[2],zmm2[2],zmm1[4],zmm2[4],zmm1[6],zmm2[6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_unpacklo_epi64:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpunpcklqdq {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[2],zmm2[2],zmm1[4],zmm2[4],zmm1[6],zmm2[6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_unpacklo_epi64:
; X64: # %bb.0:
@@ -1615,12 +1484,12 @@ define <8 x i64> @test_mm512_mask_unpack
}
define <8 x i64> @test_mm512_maskz_unpacklo_epi64(i8 %a0, <8 x i64> %a1, <8 x i64> %a2) {
-; X32-LABEL: test_mm512_maskz_unpacklo_epi64:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpunpcklqdq {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_unpacklo_epi64:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpunpcklqdq {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_unpacklo_epi64:
; X64: # %bb.0:
@@ -1634,26 +1503,21 @@ define <8 x i64> @test_mm512_maskz_unpac
}
define <8 x double> @test_mm512_unpacklo_pd(<8 x double> %a0, <8 x double> %a1) {
-; X32-LABEL: test_mm512_unpacklo_pd:
-; X32: # %bb.0:
-; X32-NEXT: vunpcklpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_unpacklo_pd:
-; X64: # %bb.0:
-; X64-NEXT: vunpcklpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_unpacklo_pd:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vunpcklpd {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x double> %a0, <8 x double> %a1, <8 x i32> <i32 0, i32 8, i32 2, i32 10, i32 4, i32 12, i32 6, i32 14>
ret <8 x double> %res
}
define <8 x double> @test_mm512_mask_unpacklo_pd(<8 x double> %a0, i8 %a1, <8 x double> %a2, <8 x double> %a3) {
-; X32-LABEL: test_mm512_mask_unpacklo_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vunpcklpd {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[2],zmm2[2],zmm1[4],zmm2[4],zmm1[6],zmm2[6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_unpacklo_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vunpcklpd {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[2],zmm2[2],zmm1[4],zmm2[4],zmm1[6],zmm2[6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_unpacklo_pd:
; X64: # %bb.0:
@@ -1667,12 +1531,12 @@ define <8 x double> @test_mm512_mask_unp
}
define <8 x double> @test_mm512_maskz_unpacklo_pd(i8 %a0, <8 x double> %a1, <8 x double> %a2) {
-; X32-LABEL: test_mm512_maskz_unpacklo_pd:
-; X32: # %bb.0:
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vunpcklpd {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_unpacklo_pd:
+; X86: # %bb.0:
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vunpcklpd {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[2],zmm1[2],zmm0[4],zmm1[4],zmm0[6],zmm1[6]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_unpacklo_pd:
; X64: # %bb.0:
@@ -1686,25 +1550,20 @@ define <8 x double> @test_mm512_maskz_un
}
define <16 x float> @test_mm512_unpacklo_ps(<16 x float> %a0, <16 x float> %a1) {
-; X32-LABEL: test_mm512_unpacklo_ps:
-; X32: # %bb.0:
-; X32-NEXT: vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_unpacklo_ps:
-; X64: # %bb.0:
-; X64-NEXT: vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_unpacklo_ps:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vunpcklps {{.*#+}} zmm0 = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <16 x float> %a0, <16 x float> %a1, <16 x i32> <i32 0, i32 16, i32 1, i32 17, i32 4, i32 20, i32 5, i32 21, i32 8, i32 24, i32 9, i32 25, i32 12, i32 28, i32 13, i32 29>
ret <16 x float> %res
}
define <16 x float> @test_mm512_mask_unpacklo_ps(<16 x float> %a0, i16 %a1, <16 x float> %a2, <16 x float> %a3) {
-; X32-LABEL: test_mm512_mask_unpacklo_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vunpcklps {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[1],zmm2[1],zmm1[4],zmm2[4],zmm1[5],zmm2[5],zmm1[8],zmm2[8],zmm1[9],zmm2[9],zmm1[12],zmm2[12],zmm1[13],zmm2[13]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_unpacklo_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vunpcklps {{.*#+}} zmm0 {%k1} = zmm1[0],zmm2[0],zmm1[1],zmm2[1],zmm1[4],zmm2[4],zmm1[5],zmm2[5],zmm1[8],zmm2[8],zmm1[9],zmm2[9],zmm1[12],zmm2[12],zmm1[13],zmm2[13]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_unpacklo_ps:
; X64: # %bb.0:
@@ -1718,11 +1577,11 @@ define <16 x float> @test_mm512_mask_unp
}
define <16 x float> @test_mm512_maskz_unpacklo_ps(i16 %a0, <16 x float> %a1, <16 x float> %a2) {
-; X32-LABEL: test_mm512_maskz_unpacklo_ps:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vunpcklps {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_unpacklo_ps:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vunpcklps {{.*#+}} zmm0 {%k1} {z} = zmm0[0],zmm1[0],zmm0[1],zmm1[1],zmm0[4],zmm1[4],zmm0[5],zmm1[5],zmm0[8],zmm1[8],zmm0[9],zmm1[9],zmm0[12],zmm1[12],zmm0[13],zmm1[13]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_unpacklo_ps:
; X64: # %bb.0:
@@ -1736,107 +1595,68 @@ define <16 x float> @test_mm512_maskz_un
}
define <8 x double> @test_mm512_zextpd128_pd512(<2 x double> %a0) nounwind {
-; X32-LABEL: test_mm512_zextpd128_pd512:
-; X32: # %bb.0:
-; X32-NEXT: vmovaps %xmm0, %xmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_zextpd128_pd512:
-; X64: # %bb.0:
-; X64-NEXT: vmovaps %xmm0, %xmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_zextpd128_pd512:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovaps %xmm0, %xmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <2 x double> %a0, <2 x double> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 2, i32 3, i32 2, i32 3>
ret <8 x double> %res
}
define <8 x double> @test_mm512_zextpd256_pd512(<4 x double> %a0) nounwind {
-; X32-LABEL: test_mm512_zextpd256_pd512:
-; X32: # %bb.0:
-; X32-NEXT: vmovaps %ymm0, %ymm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_zextpd256_pd512:
-; X64: # %bb.0:
-; X64-NEXT: vmovaps %ymm0, %ymm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_zextpd256_pd512:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovaps %ymm0, %ymm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <4 x double> %a0, <4 x double> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
ret <8 x double> %res
}
define <16 x float> @test_mm512_zextps128_ps512(<4 x float> %a0) nounwind {
-; X32-LABEL: test_mm512_zextps128_ps512:
-; X32: # %bb.0:
-; X32-NEXT: vmovaps %xmm0, %xmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_zextps128_ps512:
-; X64: # %bb.0:
-; X64-NEXT: vmovaps %xmm0, %xmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_zextps128_ps512:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovaps %xmm0, %xmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <4 x float> %a0, <4 x float> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 4, i32 5, i32 6, i32 7, i32 4, i32 5, i32 6, i32 7>
ret <16 x float> %res
}
define <16 x float> @test_mm512_zextps256_ps512(<8 x float> %a0) nounwind {
-; X32-LABEL: test_mm512_zextps256_ps512:
-; X32: # %bb.0:
-; X32-NEXT: vmovaps %ymm0, %ymm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_zextps256_ps512:
-; X64: # %bb.0:
-; X64-NEXT: vmovaps %ymm0, %ymm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_zextps256_ps512:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovaps %ymm0, %ymm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <8 x float> %a0, <8 x float> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
ret <16 x float> %res
}
define <8 x i64> @test_mm512_zextsi128_si512(<2 x i64> %a0) nounwind {
-; X32-LABEL: test_mm512_zextsi128_si512:
-; X32: # %bb.0:
-; X32-NEXT: vmovaps %xmm0, %xmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_zextsi128_si512:
-; X64: # %bb.0:
-; X64-NEXT: vmovaps %xmm0, %xmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_zextsi128_si512:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovaps %xmm0, %xmm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <2 x i64> %a0, <2 x i64> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 2, i32 3, i32 2, i32 3>
ret <8 x i64> %res
}
define <8 x i64> @test_mm512_zextsi256_si512(<4 x i64> %a0) nounwind {
-; X32-LABEL: test_mm512_zextsi256_si512:
-; X32: # %bb.0:
-; X32-NEXT: vmovaps %ymm0, %ymm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_zextsi256_si512:
-; X64: # %bb.0:
-; X64-NEXT: vmovaps %ymm0, %ymm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_zextsi256_si512:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vmovaps %ymm0, %ymm0
+; CHECK-NEXT: ret{{[l|q]}}
%res = shufflevector <4 x i64> %a0, <4 x i64> zeroinitializer, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
ret <8 x i64> %res
}
define <8 x i64> @test_mm512_mul_epi32(<8 x i64> %__A, <8 x i64> %__B) nounwind {
-; X32-LABEL: test_mm512_mul_epi32:
-; X32: # %bb.0:
-; X32-NEXT: vpsllq $32, %zmm0, %zmm0
-; X32-NEXT: vpsraq $32, %zmm0, %zmm0
-; X32-NEXT: vpsllq $32, %zmm1, %zmm1
-; X32-NEXT: vpsraq $32, %zmm1, %zmm1
-; X32-NEXT: vpmuldq %zmm0, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_mul_epi32:
-; X64: # %bb.0:
-; X64-NEXT: vpsllq $32, %zmm0, %zmm0
-; X64-NEXT: vpsraq $32, %zmm0, %zmm0
-; X64-NEXT: vpsllq $32, %zmm1, %zmm1
-; X64-NEXT: vpsraq $32, %zmm1, %zmm1
-; X64-NEXT: vpmuldq %zmm0, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_mul_epi32:
+; CHECK: # %bb.0:
+; CHECK-NEXT: vpsllq $32, %zmm0, %zmm0
+; CHECK-NEXT: vpsraq $32, %zmm0, %zmm0
+; CHECK-NEXT: vpsllq $32, %zmm1, %zmm1
+; CHECK-NEXT: vpsraq $32, %zmm1, %zmm1
+; CHECK-NEXT: vpmuldq %zmm0, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
%tmp = shl <8 x i64> %__A, <i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32>
%tmp1 = ashr exact <8 x i64> %tmp, <i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32>
%tmp2 = shl <8 x i64> %__B, <i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32, i64 32>
@@ -1846,11 +1666,11 @@ define <8 x i64> @test_mm512_mul_epi32(<
}
define <8 x i64> @test_mm512_maskz_mul_epi32(i16 zeroext %__k, <8 x i64> %__A, <8 x i64> %__B) nounwind {
-; X32-LABEL: test_mm512_maskz_mul_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpmuldq %zmm0, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_mul_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpmuldq %zmm0, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_mul_epi32:
; X64: # %bb.0:
@@ -1869,12 +1689,12 @@ define <8 x i64> @test_mm512_maskz_mul_e
}
define <8 x i64> @test_mm512_mask_mul_epi32(i16 zeroext %__k, <8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__src) nounwind {
-; X32-LABEL: test_mm512_mask_mul_epi32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpmuldq %zmm0, %zmm1, %zmm2 {%k1}
-; X32-NEXT: vmovdqa64 %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_mul_epi32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpmuldq %zmm0, %zmm1, %zmm2 {%k1}
+; X86-NEXT: vmovdqa64 %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_mul_epi32:
; X64: # %bb.0:
@@ -1894,25 +1714,15 @@ define <8 x i64> @test_mm512_mask_mul_ep
}
define <8 x i64> @test_mm512_mul_epu32(<8 x i64> %__A, <8 x i64> %__B) nounwind {
-; X32-LABEL: test_mm512_mul_epu32:
-; X32: # %bb.0:
-; X32-NEXT: movw $-21846, %ax # imm = 0xAAAA
-; X32-NEXT: kmovw %eax, %k0
-; X32-NEXT: knotw %k0, %k1
-; X32-NEXT: vmovdqa32 %zmm0, %zmm0 {%k1} {z}
-; X32-NEXT: vmovdqa32 %zmm1, %zmm1 {%k1} {z}
-; X32-NEXT: vpmuludq %zmm0, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_mul_epu32:
-; X64: # %bb.0:
-; X64-NEXT: movw $-21846, %ax # imm = 0xAAAA
-; X64-NEXT: kmovw %eax, %k0
-; X64-NEXT: knotw %k0, %k1
-; X64-NEXT: vmovdqa32 %zmm0, %zmm0 {%k1} {z}
-; X64-NEXT: vmovdqa32 %zmm1, %zmm1 {%k1} {z}
-; X64-NEXT: vpmuludq %zmm0, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_mul_epu32:
+; CHECK: # %bb.0:
+; CHECK-NEXT: movw $-21846, %ax # imm = 0xAAAA
+; CHECK-NEXT: kmovw %eax, %k0
+; CHECK-NEXT: knotw %k0, %k1
+; CHECK-NEXT: vmovdqa32 %zmm0, %zmm0 {%k1} {z}
+; CHECK-NEXT: vmovdqa32 %zmm1, %zmm1 {%k1} {z}
+; CHECK-NEXT: vpmuludq %zmm0, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
%tmp = and <8 x i64> %__A, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
%tmp1 = and <8 x i64> %__B, <i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295, i64 4294967295>
%tmp2 = mul nuw <8 x i64> %tmp1, %tmp
@@ -1920,11 +1730,11 @@ define <8 x i64> @test_mm512_mul_epu32(<
}
define <8 x i64> @test_mm512_maskz_mul_epu32(i16 zeroext %__k, <8 x i64> %__A, <8 x i64> %__B) nounwind {
-; X32-LABEL: test_mm512_maskz_mul_epu32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpmuludq %zmm0, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_mul_epu32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpmuludq %zmm0, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_mul_epu32:
; X64: # %bb.0:
@@ -1941,12 +1751,12 @@ define <8 x i64> @test_mm512_maskz_mul_e
}
define <8 x i64> @test_mm512_mask_mul_epu32(i16 zeroext %__k, <8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__src) nounwind {
-; X32-LABEL: test_mm512_mask_mul_epu32:
-; X32: # %bb.0:
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpmuludq %zmm0, %zmm1, %zmm2 {%k1}
-; X32-NEXT: vmovdqa64 %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_mul_epu32:
+; X86: # %bb.0:
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpmuludq %zmm0, %zmm1, %zmm2 {%k1}
+; X86-NEXT: vmovdqa64 %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_mul_epu32:
; X64: # %bb.0:
@@ -1964,13 +1774,13 @@ define <8 x i64> @test_mm512_mask_mul_ep
}
define <8 x double> @test_mm512_set1_epi8(i8 signext %d) nounwind {
-; X32-LABEL: test_mm512_set1_epi8:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vmovd %eax, %xmm0
-; X32-NEXT: vpbroadcastb %xmm0, %ymm0
-; X32-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_set1_epi8:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vmovd %eax, %xmm0
+; X86-NEXT: vpbroadcastb %xmm0, %ymm0
+; X86-NEXT: vinserti64x4 $1, %ymm0, %zmm0, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_set1_epi8:
; X64: # %bb.0: # %entry
@@ -1986,11 +1796,11 @@ entry:
}
define <2 x double> @test_mm_cvtu32_sd(<2 x double> %__A, i32 %__B) {
-; X32-LABEL: test_mm_cvtu32_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: vcvtusi2sdl %eax, %xmm0, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_cvtu32_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: vcvtusi2sdl %eax, %xmm0, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_cvtu32_sd:
; X64: # %bb.0: # %entry
@@ -2003,15 +1813,15 @@ entry:
}
define <2 x double> @test_mm_cvtu64_sd(<2 x double> %__A, i64 %__B) {
-; X32-LABEL: test_mm_cvtu64_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; X32-NEXT: vpinsrd $1, {{[0-9]+}}(%esp), %xmm1, %xmm1
-; X32-NEXT: vpunpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1]
-; X32-NEXT: vsubpd {{\.LCPI.*}}, %xmm1, %xmm1
-; X32-NEXT: vhaddpd %xmm1, %xmm1, %xmm1
-; X32-NEXT: vmovsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
-; X32-NEXT: retl
+; X86-LABEL: test_mm_cvtu64_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X86-NEXT: vpinsrd $1, {{[0-9]+}}(%esp), %xmm1, %xmm1
+; X86-NEXT: vpunpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1]
+; X86-NEXT: vsubpd {{\.LCPI.*}}, %xmm1, %xmm1
+; X86-NEXT: vhaddpd %xmm1, %xmm1, %xmm1
+; X86-NEXT: vmovsd {{.*#+}} xmm0 = xmm1[0],xmm0[1]
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_cvtu64_sd:
; X64: # %bb.0: # %entry
@@ -2024,11 +1834,11 @@ entry:
}
define <4 x float> @test_mm_cvtu32_ss(<4 x float> %__A, i32 %__B) {
-; X32-LABEL: test_mm_cvtu32_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movl {{[0-9]+}}(%esp), %eax
-; X32-NEXT: vcvtusi2ssl %eax, %xmm0, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_cvtu32_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movl {{[0-9]+}}(%esp), %eax
+; X86-NEXT: vcvtusi2ssl %eax, %xmm0, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_cvtu32_ss:
; X64: # %bb.0: # %entry
@@ -2041,31 +1851,31 @@ entry:
}
define <4 x float> @test_mm_cvtu64_ss(<4 x float> %__A, i64 %__B) {
-; X32-LABEL: test_mm_cvtu64_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: pushl %ebp
-; X32-NEXT: .cfi_def_cfa_offset 8
-; X32-NEXT: .cfi_offset %ebp, -8
-; X32-NEXT: movl %esp, %ebp
-; X32-NEXT: .cfi_def_cfa_register %ebp
-; X32-NEXT: andl $-8, %esp
-; X32-NEXT: subl $16, %esp
-; X32-NEXT: movl 12(%ebp), %eax
-; X32-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; X32-NEXT: vpinsrd $1, %eax, %xmm1, %xmm1
-; X32-NEXT: vmovq %xmm1, {{[0-9]+}}(%esp)
-; X32-NEXT: xorl %ecx, %ecx
-; X32-NEXT: testl %eax, %eax
-; X32-NEXT: setns %cl
-; X32-NEXT: fildll {{[0-9]+}}(%esp)
-; X32-NEXT: fadds {{\.LCPI.*}}(,%ecx,4)
-; X32-NEXT: fstps {{[0-9]+}}(%esp)
-; X32-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
-; X32-NEXT: vmovss {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
-; X32-NEXT: movl %ebp, %esp
-; X32-NEXT: popl %ebp
-; X32-NEXT: .cfi_def_cfa %esp, 4
-; X32-NEXT: retl
+; X86-LABEL: test_mm_cvtu64_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: pushl %ebp
+; X86-NEXT: .cfi_def_cfa_offset 8
+; X86-NEXT: .cfi_offset %ebp, -8
+; X86-NEXT: movl %esp, %ebp
+; X86-NEXT: .cfi_def_cfa_register %ebp
+; X86-NEXT: andl $-8, %esp
+; X86-NEXT: subl $16, %esp
+; X86-NEXT: movl 12(%ebp), %eax
+; X86-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X86-NEXT: vpinsrd $1, %eax, %xmm1, %xmm1
+; X86-NEXT: vmovq %xmm1, {{[0-9]+}}(%esp)
+; X86-NEXT: xorl %ecx, %ecx
+; X86-NEXT: testl %eax, %eax
+; X86-NEXT: setns %cl
+; X86-NEXT: fildll {{[0-9]+}}(%esp)
+; X86-NEXT: fadds {{\.LCPI.*}}(,%ecx,4)
+; X86-NEXT: fstps {{[0-9]+}}(%esp)
+; X86-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero
+; X86-NEXT: vmovss {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3]
+; X86-NEXT: movl %ebp, %esp
+; X86-NEXT: popl %ebp
+; X86-NEXT: .cfi_def_cfa %esp, 4
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_cvtu64_ss:
; X64: # %bb.0: # %entry
@@ -2078,30 +1888,20 @@ entry:
}
define <8 x double> @test_mm512_cvtps_pd(<8 x float> %__A) {
-; X32-LABEL: test_mm512_cvtps_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vcvtps2pd %ymm0, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_cvtps_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vcvtps2pd %ymm0, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_cvtps_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vcvtps2pd %ymm0, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%conv.i = fpext <8 x float> %__A to <8 x double>
ret <8 x double> %conv.i
}
define <8 x double> @test_mm512_cvtpslo_pd(<16 x float> %__A) {
-; X32-LABEL: test_mm512_cvtpslo_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vcvtps2pd %ymm0, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_cvtpslo_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vcvtps2pd %ymm0, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_cvtpslo_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vcvtps2pd %ymm0, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%shuffle.i.i = shufflevector <16 x float> %__A, <16 x float> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
%conv.i.i = fpext <8 x float> %shuffle.i.i to <8 x double>
@@ -2109,12 +1909,12 @@ entry:
}
define <8 x double> @test_mm512_mask_cvtps_pd(<8 x double> %__W, i8 zeroext %__U, <8 x float> %__A) {
-; X32-LABEL: test_mm512_mask_cvtps_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vcvtps2pd %ymm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_cvtps_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vcvtps2pd %ymm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_cvtps_pd:
; X64: # %bb.0: # %entry
@@ -2129,12 +1929,12 @@ entry:
}
define <8 x double> @test_mm512_mask_cvtpslo_pd(<8 x double> %__W, i8 zeroext %__U, <16 x float> %__A) {
-; X32-LABEL: test_mm512_mask_cvtpslo_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vcvtps2pd %ymm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_cvtpslo_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vcvtps2pd %ymm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_cvtpslo_pd:
; X64: # %bb.0: # %entry
@@ -2150,12 +1950,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_cvtps_pd(i8 zeroext %__U, <8 x float> %__A) {
-; X32-LABEL: test_mm512_maskz_cvtps_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vcvtps2pd %ymm0, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_cvtps_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vcvtps2pd %ymm0, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_cvtps_pd:
; X64: # %bb.0: # %entry
@@ -2170,17 +1970,11 @@ entry:
}
define <2 x i64> @test_mm512_cvtepi32_epi8(<8 x i64> %__A) {
-; X32-LABEL: test_mm512_cvtepi32_epi8:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpmovdb %zmm0, %xmm0
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_cvtepi32_epi8:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpmovdb %zmm0, %xmm0
-; X64-NEXT: vzeroupper
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_cvtepi32_epi8:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpmovdb %zmm0, %xmm0
+; CHECK-NEXT: vzeroupper
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = bitcast <8 x i64> %__A to <16 x i32>
%conv.i = trunc <16 x i32> %0 to <16 x i8>
@@ -2189,12 +1983,12 @@ entry:
}
define <2 x i64> @test_mm512_mask_cvtepi32_epi8(<2 x i64> %__O, i16 zeroext %__M, <8 x i64> %__A) {
-; X32-LABEL: test_mm512_mask_cvtepi32_epi8:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpmovdb %zmm1, %xmm0 {%k1}
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_cvtepi32_epi8:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpmovdb %zmm1, %xmm0 {%k1}
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_cvtepi32_epi8:
; X64: # %bb.0: # %entry
@@ -2211,12 +2005,12 @@ entry:
}
define <2 x i64> @test_mm512_maskz_cvtepi32_epi8(i16 zeroext %__M, <8 x i64> %__A) {
-; X32-LABEL: test_mm512_maskz_cvtepi32_epi8:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpmovdb %zmm0, %xmm0 {%k1} {z}
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_cvtepi32_epi8:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpmovdb %zmm0, %xmm0 {%k1} {z}
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_cvtepi32_epi8:
; X64: # %bb.0: # %entry
@@ -2232,15 +2026,10 @@ entry:
}
define <4 x i64> @test_mm512_cvtepi64_epi32(<8 x i64> %__A) {
-; X32-LABEL: test_mm512_cvtepi64_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpmovqd %zmm0, %ymm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_cvtepi64_epi32:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpmovqd %zmm0, %ymm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_cvtepi64_epi32:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpmovqd %zmm0, %ymm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%conv.i = trunc <8 x i64> %__A to <8 x i32>
%0 = bitcast <8 x i32> %conv.i to <4 x i64>
@@ -2248,12 +2037,12 @@ entry:
}
define <4 x i64> @test_mm512_mask_cvtepi64_epi32(<4 x i64> %__O, i8 zeroext %__M, <8 x i64> %__A) {
-; X32-LABEL: test_mm512_mask_cvtepi64_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpmovqd %zmm1, %ymm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_cvtepi64_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpmovqd %zmm1, %ymm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_cvtepi64_epi32:
; X64: # %bb.0: # %entry
@@ -2270,12 +2059,12 @@ entry:
}
define <4 x i64> @test_mm512_maskz_cvtepi64_epi32(i8 zeroext %__M, <8 x i64> %__A) {
-; X32-LABEL: test_mm512_maskz_cvtepi64_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpmovqd %zmm0, %ymm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_cvtepi64_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpmovqd %zmm0, %ymm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_cvtepi64_epi32:
; X64: # %bb.0: # %entry
@@ -2291,17 +2080,11 @@ entry:
}
define <2 x i64> @test_mm512_cvtepi64_epi16(<8 x i64> %__A) {
-; X32-LABEL: test_mm512_cvtepi64_epi16:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpmovqw %zmm0, %xmm0
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_cvtepi64_epi16:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpmovqw %zmm0, %xmm0
-; X64-NEXT: vzeroupper
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_cvtepi64_epi16:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpmovqw %zmm0, %xmm0
+; CHECK-NEXT: vzeroupper
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%conv.i = trunc <8 x i64> %__A to <8 x i16>
%0 = bitcast <8 x i16> %conv.i to <2 x i64>
@@ -2309,13 +2092,13 @@ entry:
}
define <2 x i64> @test_mm512_mask_cvtepi64_epi16(<2 x i64> %__O, i8 zeroext %__M, <8 x i64> %__A) {
-; X32-LABEL: test_mm512_mask_cvtepi64_epi16:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpmovqw %zmm1, %xmm0 {%k1}
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_cvtepi64_epi16:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpmovqw %zmm1, %xmm0 {%k1}
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_cvtepi64_epi16:
; X64: # %bb.0: # %entry
@@ -2331,13 +2114,13 @@ entry:
}
define <2 x i64> @test_mm512_maskz_cvtepi64_epi16(i8 zeroext %__M, <8 x i64> %__A) {
-; X32-LABEL: test_mm512_maskz_cvtepi64_epi16:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpmovqw %zmm0, %xmm0 {%k1} {z}
-; X32-NEXT: vzeroupper
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_cvtepi64_epi16:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpmovqw %zmm0, %xmm0 {%k1} {z}
+; X86-NEXT: vzeroupper
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_cvtepi64_epi16:
; X64: # %bb.0: # %entry
@@ -2355,15 +2138,10 @@ declare <16 x i8> @llvm.x86.avx512.mask.
declare <8 x i16> @llvm.x86.avx512.mask.pmov.qw.512(<8 x i64>, <8 x i16>, i8)
define <8 x i64> @test_mm512_ternarylogic_epi32(<8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__C) {
-; X32-LABEL: test_mm512_ternarylogic_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpternlogd $4, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_ternarylogic_epi32:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpternlogd $4, %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_ternarylogic_epi32:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpternlogd $4, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = bitcast <8 x i64> %__A to <16 x i32>
%1 = bitcast <8 x i64> %__B to <16 x i32>
@@ -2376,11 +2154,11 @@ entry:
declare <16 x i32> @llvm.x86.avx512.pternlog.d.512(<16 x i32>, <16 x i32>, <16 x i32>, i32) #1
define <8 x i64> @test_mm512_mask_ternarylogic_epi32(<8 x i64> %__A, i16 zeroext %__U, <8 x i64> %__B, <8 x i64> %__C) {
-; X32-LABEL: test_mm512_mask_ternarylogic_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpternlogd $4, %zmm2, %zmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_ternarylogic_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpternlogd $4, %zmm2, %zmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_ternarylogic_epi32:
; X64: # %bb.0: # %entry
@@ -2399,11 +2177,11 @@ entry:
}
define <8 x i64> @test_mm512_maskz_ternarylogic_epi32(i16 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__C) {
-; X32-LABEL: test_mm512_maskz_ternarylogic_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpternlogd $4, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_ternarylogic_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpternlogd $4, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_ternarylogic_epi32:
; X64: # %bb.0: # %entry
@@ -2422,15 +2200,10 @@ entry:
}
define <8 x i64> @test_mm512_ternarylogic_epi64(<8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__C) {
-; X32-LABEL: test_mm512_ternarylogic_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpternlogq $4, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_ternarylogic_epi64:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpternlogq $4, %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_ternarylogic_epi64:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpternlogq $4, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <8 x i64> @llvm.x86.avx512.pternlog.q.512(<8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__C, i32 4)
ret <8 x i64> %0
@@ -2439,12 +2212,12 @@ entry:
declare <8 x i64> @llvm.x86.avx512.pternlog.q.512(<8 x i64>, <8 x i64>, <8 x i64>, i32) #1
define <8 x i64> @test_mm512_mask_ternarylogic_epi64(<8 x i64> %__A, i8 zeroext %__U, <8 x i64> %__B, <8 x i64> %__C) {
-; X32-LABEL: test_mm512_mask_ternarylogic_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpternlogq $4, %zmm2, %zmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_ternarylogic_epi64:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpternlogq $4, %zmm2, %zmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_ternarylogic_epi64:
; X64: # %bb.0: # %entry
@@ -2459,12 +2232,12 @@ entry:
}
define <8 x i64> @test_mm512_maskz_ternarylogic_epi64(i8 zeroext %__U, <8 x i64> %__A, <8 x i64> %__B, <8 x i64> %__C) {
-; X32-LABEL: test_mm512_maskz_ternarylogic_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpternlogq $4, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_ternarylogic_epi64:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpternlogq $4, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_ternarylogic_epi64:
; X64: # %bb.0: # %entry
@@ -2481,12 +2254,12 @@ entry:
declare <16 x i32> @llvm.x86.avx512.vpermi2var.d.512(<16 x i32>, <16 x i32>, <16 x i32>)
define <8 x i64> @test_mm512_mask2_permutex2var_epi32(<8 x i64> %__A, <8 x i64> %__I, i16 zeroext %__U, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_mask2_permutex2var_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpermi2d %zmm2, %zmm0, %zmm1 {%k1}
-; X32-NEXT: vmovdqa64 %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask2_permutex2var_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpermi2d %zmm2, %zmm0, %zmm1 {%k1}
+; X86-NEXT: vmovdqa64 %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask2_permutex2var_epi32:
; X64: # %bb.0: # %entry
@@ -2508,13 +2281,13 @@ entry:
declare <8 x double> @llvm.x86.avx512.vpermi2var.pd.512(<8 x double>, <8 x i64>, <8 x double>)
define <8 x double> @test_mm512_mask2_permutex2var_pd(<8 x double> %__A, <8 x i64> %__I, i8 zeroext %__U, <8 x double> %__B) {
-; X32-LABEL: test_mm512_mask2_permutex2var_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermi2pd %zmm2, %zmm0, %zmm1 {%k1}
-; X32-NEXT: vmovapd %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask2_permutex2var_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermi2pd %zmm2, %zmm0, %zmm1 {%k1}
+; X86-NEXT: vmovapd %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask2_permutex2var_pd:
; X64: # %bb.0: # %entry
@@ -2533,12 +2306,12 @@ entry:
declare <16 x float> @llvm.x86.avx512.vpermi2var.ps.512(<16 x float>, <16 x i32>, <16 x float>)
define <16 x float> @test_mm512_mask2_permutex2var_ps(<16 x float> %__A, <8 x i64> %__I, i16 zeroext %__U, <16 x float> %__B) {
-; X32-LABEL: test_mm512_mask2_permutex2var_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpermi2ps %zmm2, %zmm0, %zmm1 {%k1}
-; X32-NEXT: vmovaps %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask2_permutex2var_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpermi2ps %zmm2, %zmm0, %zmm1 {%k1}
+; X86-NEXT: vmovaps %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask2_permutex2var_ps:
; X64: # %bb.0: # %entry
@@ -2558,13 +2331,13 @@ entry:
declare <8 x i64> @llvm.x86.avx512.vpermi2var.q.512(<8 x i64>, <8 x i64>, <8 x i64>)
define <8 x i64> @test_mm512_mask2_permutex2var_epi64(<8 x i64> %__A, <8 x i64> %__I, i8 zeroext %__U, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_mask2_permutex2var_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermi2q %zmm2, %zmm0, %zmm1 {%k1}
-; X32-NEXT: vmovdqa64 %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask2_permutex2var_epi64:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermi2q %zmm2, %zmm0, %zmm1 {%k1}
+; X86-NEXT: vmovdqa64 %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask2_permutex2var_epi64:
; X64: # %bb.0: # %entry
@@ -2580,15 +2353,10 @@ entry:
}
define <8 x i64> @test_mm512_permutex2var_epi32(<8 x i64> %__A, <8 x i64> %__I, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_permutex2var_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpermt2d %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_permutex2var_epi32:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpermt2d %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_permutex2var_epi32:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpermt2d %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = bitcast <8 x i64> %__A to <16 x i32>
%1 = bitcast <8 x i64> %__I to <16 x i32>
@@ -2599,11 +2367,11 @@ entry:
}
define <8 x i64> @test_mm512_maskz_permutex2var_epi32(i16 zeroext %__U, <8 x i64> %__A, <8 x i64> %__I, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_maskz_permutex2var_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpermt2d %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_permutex2var_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpermt2d %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_permutex2var_epi32:
; X64: # %bb.0: # %entry
@@ -2622,11 +2390,11 @@ entry:
}
define <8 x i64> @test_mm512_mask_permutex2var_epi32(<8 x i64> %__A, i16 zeroext %__U, <8 x i64> %__I, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_mask_permutex2var_epi32:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpermt2d %zmm2, %zmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_permutex2var_epi32:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpermt2d %zmm2, %zmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_permutex2var_epi32:
; X64: # %bb.0: # %entry
@@ -2645,27 +2413,22 @@ entry:
}
define <8 x double> @test_mm512_permutex2var_pd(<8 x double> %__A, <8 x i64> %__I, <8 x double> %__B) {
-; X32-LABEL: test_mm512_permutex2var_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpermt2pd %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_permutex2var_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpermt2pd %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_permutex2var_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpermt2pd %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vpermi2var.pd.512(<8 x double> %__A, <8 x i64> %__I, <8 x double> %__B)
ret <8 x double> %0
}
define <8 x double> @test_mm512_mask_permutex2var_pd(<8 x double> %__A, i8 zeroext %__U, <8 x i64> %__I, <8 x double> %__B) {
-; X32-LABEL: test_mm512_mask_permutex2var_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermt2pd %zmm2, %zmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_permutex2var_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermt2pd %zmm2, %zmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_permutex2var_pd:
; X64: # %bb.0: # %entry
@@ -2680,12 +2443,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_permutex2var_pd(i8 zeroext %__U, <8 x double> %__A, <8 x i64> %__I, <8 x double> %__B) {
-; X32-LABEL: test_mm512_maskz_permutex2var_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermt2pd %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_permutex2var_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermt2pd %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_permutex2var_pd:
; X64: # %bb.0: # %entry
@@ -2700,15 +2463,10 @@ entry:
}
define <16 x float> @test_mm512_permutex2var_ps(<16 x float> %__A, <8 x i64> %__I, <16 x float> %__B) {
-; X32-LABEL: test_mm512_permutex2var_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpermt2ps %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_permutex2var_ps:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpermt2ps %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_permutex2var_ps:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpermt2ps %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = bitcast <8 x i64> %__I to <16 x i32>
%1 = tail call <16 x float> @llvm.x86.avx512.vpermi2var.ps.512(<16 x float> %__A, <16 x i32> %0, <16 x float> %__B)
@@ -2716,11 +2474,11 @@ entry:
}
define <16 x float> @test_mm512_mask_permutex2var_ps(<16 x float> %__A, i16 zeroext %__U, <8 x i64> %__I, <16 x float> %__B) {
-; X32-LABEL: test_mm512_mask_permutex2var_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpermt2ps %zmm2, %zmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_permutex2var_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpermt2ps %zmm2, %zmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_permutex2var_ps:
; X64: # %bb.0: # %entry
@@ -2736,11 +2494,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_permutex2var_ps(i16 zeroext %__U, <16 x float> %__A, <8 x i64> %__I, <16 x float> %__B) {
-; X32-LABEL: test_mm512_maskz_permutex2var_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vpermt2ps %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_permutex2var_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vpermt2ps %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_permutex2var_ps:
; X64: # %bb.0: # %entry
@@ -2756,27 +2514,22 @@ entry:
}
define <8 x i64> @test_mm512_permutex2var_epi64(<8 x i64> %__A, <8 x i64> %__I, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_permutex2var_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpermt2q %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_permutex2var_epi64:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpermt2q %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_permutex2var_epi64:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpermt2q %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <8 x i64> @llvm.x86.avx512.vpermi2var.q.512(<8 x i64> %__A, <8 x i64> %__I, <8 x i64> %__B)
ret <8 x i64> %0
}
define <8 x i64> @test_mm512_mask_permutex2var_epi64(<8 x i64> %__A, i8 zeroext %__U, <8 x i64> %__I, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_mask_permutex2var_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermt2q %zmm2, %zmm1, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_permutex2var_epi64:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermt2q %zmm2, %zmm1, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_permutex2var_epi64:
; X64: # %bb.0: # %entry
@@ -2791,12 +2544,12 @@ entry:
}
define <8 x i64> @test_mm512_maskz_permutex2var_epi64(i8 zeroext %__U, <8 x i64> %__A, <8 x i64> %__I, <8 x i64> %__B) {
-; X32-LABEL: test_mm512_maskz_permutex2var_epi64:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vpermt2q %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_permutex2var_epi64:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vpermt2q %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_permutex2var_epi64:
; X64: # %bb.0: # %entry
@@ -2812,15 +2565,10 @@ entry:
define <8 x double> @test_mm512_fmadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fmadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmadd_round_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmadd_round_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
ret <8 x double> %0
@@ -2829,12 +2577,12 @@ entry:
declare <8 x double> @llvm.x86.avx512.vfmadd.pd.512(<8 x double>, <8 x double>, <8 x double>, i32) #1
define <8 x double> @test_mm512_mask_fmadd_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fmadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmadd_round_pd:
; X64: # %bb.0: # %entry
@@ -2849,13 +2597,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmadd_round_pd:
; X64: # %bb.0: # %entry
@@ -2871,12 +2619,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmadd_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fmadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmadd_round_pd:
; X64: # %bb.0: # %entry
@@ -2891,11 +2639,11 @@ entry:
}
define <8 x double> @test_mm512_fmsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fmsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm2, %zmm2
-; X32-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fmsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm2, %zmm2
+; X86-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fmsub_round_pd:
; X64: # %bb.0: # %entry
@@ -2909,12 +2657,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fmsub_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fmsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsub_round_pd:
; X64: # %bb.0: # %entry
@@ -2930,12 +2678,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmsub_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fmsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsub_round_pd:
; X64: # %bb.0: # %entry
@@ -2951,11 +2699,11 @@ entry:
}
define <8 x double> @test_mm512_fnmadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fnmadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm0, %zmm0
-; X32-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fnmadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm0, %zmm0
+; X86-NEXT: vfmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fnmadd_round_pd:
; X64: # %bb.0: # %entry
@@ -2969,13 +2717,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fnmadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fnmadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fnmadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmadd_round_pd:
; X64: # %bb.0: # %entry
@@ -2992,12 +2740,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fnmadd_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fnmadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fnmadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmadd_round_pd:
; X64: # %bb.0: # %entry
@@ -3013,21 +2761,13 @@ entry:
}
define <8 x double> @test_mm512_fnmsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fnmsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpbroadcastq {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0]
-; X32-NEXT: vpxorq %zmm3, %zmm0, %zmm4
-; X32-NEXT: vpxorq %zmm3, %zmm2, %zmm0
-; X32-NEXT: vfmadd231pd {rn-sae}, %zmm4, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fnmsub_round_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpbroadcastq {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0]
-; X64-NEXT: vpxorq %zmm3, %zmm0, %zmm4
-; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0
-; X64-NEXT: vfmadd231pd {rn-sae}, %zmm4, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fnmsub_round_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpbroadcastq {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0]
+; CHECK-NEXT: vpxorq %zmm3, %zmm0, %zmm4
+; CHECK-NEXT: vpxorq %zmm3, %zmm2, %zmm0
+; CHECK-NEXT: vfmadd231pd {rn-sae}, %zmm4, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%sub = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%sub1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -3036,12 +2776,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fnmsub_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fnmsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fnmsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmsub_round_pd:
; X64: # %bb.0: # %entry
@@ -3058,27 +2798,22 @@ entry:
}
define <8 x double> @test_mm512_fmadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fmadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmadd_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmadd_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
ret <8 x double> %0
}
define <8 x double> @test_mm512_mask_fmadd_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fmadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmadd_pd:
; X64: # %bb.0: # %entry
@@ -3093,13 +2828,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmadd_pd:
; X64: # %bb.0: # %entry
@@ -3115,12 +2850,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmadd_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fmadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmadd_pd:
; X64: # %bb.0: # %entry
@@ -3135,11 +2870,11 @@ entry:
}
define <8 x double> @test_mm512_fmsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fmsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm2, %zmm2
-; X32-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fmsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm2, %zmm2
+; X86-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fmsub_pd:
; X64: # %bb.0: # %entry
@@ -3153,12 +2888,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fmsub_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fmsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) - zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsub_pd:
; X64: # %bb.0: # %entry
@@ -3174,12 +2909,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmsub_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fmsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) - zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) - zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsub_pd:
; X64: # %bb.0: # %entry
@@ -3195,11 +2930,11 @@ entry:
}
define <8 x double> @test_mm512_fnmadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fnmadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm0, %zmm0
-; X32-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fnmadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm0, %zmm0
+; X86-NEXT: vfmadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fnmadd_pd:
; X64: # %bb.0: # %entry
@@ -3213,13 +2948,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fnmadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fnmadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) + zmm2
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fnmadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmadd_pd:
; X64: # %bb.0: # %entry
@@ -3236,12 +2971,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fnmadd_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fnmadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd213pd {{.*#+}} zmm0 = -(zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fnmadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd213pd {{.*#+}} zmm0 = -(zmm1 * zmm0) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmadd_pd:
; X64: # %bb.0: # %entry
@@ -3257,21 +2992,13 @@ entry:
}
define <8 x double> @test_mm512_fnmsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fnmsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpbroadcastq {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0]
-; X32-NEXT: vpxorq %zmm3, %zmm0, %zmm4
-; X32-NEXT: vpxorq %zmm3, %zmm2, %zmm0
-; X32-NEXT: vfmadd231pd {{.*#+}} zmm0 = (zmm1 * zmm4) + zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fnmsub_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpbroadcastq {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0]
-; X64-NEXT: vpxorq %zmm3, %zmm0, %zmm4
-; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0
-; X64-NEXT: vfmadd231pd {{.*#+}} zmm0 = (zmm1 * zmm4) + zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fnmsub_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpbroadcastq {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0]
+; CHECK-NEXT: vpxorq %zmm3, %zmm0, %zmm4
+; CHECK-NEXT: vpxorq %zmm3, %zmm2, %zmm0
+; CHECK-NEXT: vfmadd231pd {{.*#+}} zmm0 = (zmm1 * zmm4) + zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__A
%sub1.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -3280,12 +3007,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fnmsub_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fnmsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub213pd {{.*#+}} zmm0 = -(zmm1 * zmm0) - zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fnmsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub213pd {{.*#+}} zmm0 = -(zmm1 * zmm0) - zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmsub_pd:
; X64: # %bb.0: # %entry
@@ -3302,15 +3029,10 @@ entry:
}
define <16 x float> @test_mm512_fmadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fmadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmadd_round_ps:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmadd_round_ps:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
ret <16 x float> %0
@@ -3319,11 +3041,11 @@ entry:
declare <16 x float> @llvm.x86.avx512.vfmadd.ps.512(<16 x float>, <16 x float>, <16 x float>, i32) #1
define <16 x float> @test_mm512_mask_fmadd_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fmadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmadd_round_ps:
; X64: # %bb.0: # %entry
@@ -3338,12 +3060,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmadd_round_ps:
; X64: # %bb.0: # %entry
@@ -3359,11 +3081,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmadd_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fmadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmadd_round_ps:
; X64: # %bb.0: # %entry
@@ -3378,11 +3100,11 @@ entry:
}
define <16 x float> @test_mm512_fmsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fmsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm2, %zmm2
-; X32-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fmsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm2, %zmm2
+; X86-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fmsub_round_ps:
; X64: # %bb.0: # %entry
@@ -3396,11 +3118,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fmsub_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fmsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsub_round_ps:
; X64: # %bb.0: # %entry
@@ -3416,11 +3138,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmsub_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fmsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsub_round_ps:
; X64: # %bb.0: # %entry
@@ -3436,11 +3158,11 @@ entry:
}
define <16 x float> @test_mm512_fnmadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fnmadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm0, %zmm0
-; X32-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fnmadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm0, %zmm0
+; X86-NEXT: vfmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fnmadd_round_ps:
; X64: # %bb.0: # %entry
@@ -3454,12 +3176,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fnmadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fnmadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fnmadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmadd_round_ps:
; X64: # %bb.0: # %entry
@@ -3476,11 +3198,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fnmadd_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fnmadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fnmadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmadd_round_ps:
; X64: # %bb.0: # %entry
@@ -3496,21 +3218,13 @@ entry:
}
define <16 x float> @test_mm512_fnmsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fnmsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpbroadcastd {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0]
-; X32-NEXT: vpxorq %zmm3, %zmm0, %zmm4
-; X32-NEXT: vpxorq %zmm3, %zmm2, %zmm0
-; X32-NEXT: vfmadd231ps {rn-sae}, %zmm4, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fnmsub_round_ps:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpbroadcastd {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0]
-; X64-NEXT: vpxorq %zmm3, %zmm0, %zmm4
-; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0
-; X64-NEXT: vfmadd231ps {rn-sae}, %zmm4, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fnmsub_round_ps:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpbroadcastd {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0]
+; CHECK-NEXT: vpxorq %zmm3, %zmm0, %zmm4
+; CHECK-NEXT: vpxorq %zmm3, %zmm2, %zmm0
+; CHECK-NEXT: vfmadd231ps {rn-sae}, %zmm4, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%sub = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%sub1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -3519,11 +3233,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fnmsub_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fnmsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fnmsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmsub_round_ps:
; X64: # %bb.0: # %entry
@@ -3540,26 +3254,21 @@ entry:
}
define <16 x float> @test_mm512_fmadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fmadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmadd_ps:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmadd_ps:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
ret <16 x float> %0
}
define <16 x float> @test_mm512_mask_fmadd_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fmadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmadd_ps:
; X64: # %bb.0: # %entry
@@ -3574,12 +3283,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) + zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmadd_ps:
; X64: # %bb.0: # %entry
@@ -3595,11 +3304,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmadd_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fmadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmadd_ps:
; X64: # %bb.0: # %entry
@@ -3614,11 +3323,11 @@ entry:
}
define <16 x float> @test_mm512_fmsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fmsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm2, %zmm2
-; X32-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fmsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm2, %zmm2
+; X86-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fmsub_ps:
; X64: # %bb.0: # %entry
@@ -3632,11 +3341,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fmsub_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fmsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) - zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsub_ps:
; X64: # %bb.0: # %entry
@@ -3652,11 +3361,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmsub_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fmsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) - zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) - zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsub_ps:
; X64: # %bb.0: # %entry
@@ -3672,11 +3381,11 @@ entry:
}
define <16 x float> @test_mm512_fnmadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fnmadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm0, %zmm0
-; X32-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fnmadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm0, %zmm0
+; X86-NEXT: vfmadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fnmadd_ps:
; X64: # %bb.0: # %entry
@@ -3690,12 +3399,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fnmadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fnmadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmadd231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) + zmm2
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fnmadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmadd231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmadd_ps:
; X64: # %bb.0: # %entry
@@ -3712,11 +3421,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fnmadd_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fnmadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmadd213ps {{.*#+}} zmm0 = -(zmm1 * zmm0) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fnmadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmadd213ps {{.*#+}} zmm0 = -(zmm1 * zmm0) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmadd_ps:
; X64: # %bb.0: # %entry
@@ -3732,21 +3441,13 @@ entry:
}
define <16 x float> @test_mm512_fnmsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fnmsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpbroadcastd {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0]
-; X32-NEXT: vpxorq %zmm3, %zmm0, %zmm4
-; X32-NEXT: vpxorq %zmm3, %zmm2, %zmm0
-; X32-NEXT: vfmadd231ps {{.*#+}} zmm0 = (zmm1 * zmm4) + zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fnmsub_ps:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vpbroadcastd {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0]
-; X64-NEXT: vpxorq %zmm3, %zmm0, %zmm4
-; X64-NEXT: vpxorq %zmm3, %zmm2, %zmm0
-; X64-NEXT: vfmadd231ps {{.*#+}} zmm0 = (zmm1 * zmm4) + zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fnmsub_ps:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vpbroadcastd {{.*#+}} zmm3 = [-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0,-0]
+; CHECK-NEXT: vpxorq %zmm3, %zmm0, %zmm4
+; CHECK-NEXT: vpxorq %zmm3, %zmm2, %zmm0
+; CHECK-NEXT: vfmadd231ps {{.*#+}} zmm0 = (zmm1 * zmm4) + zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__A
%sub1.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -3755,11 +3456,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fnmsub_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fnmsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmsub213ps {{.*#+}} zmm0 = -(zmm1 * zmm0) - zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fnmsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmsub213ps {{.*#+}} zmm0 = -(zmm1 * zmm0) - zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fnmsub_ps:
; X64: # %bb.0: # %entry
@@ -3776,15 +3477,10 @@ entry:
}
define <8 x double> @test_mm512_fmaddsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fmaddsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmaddsub_round_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmaddsub_round_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i32 8)
ret <8 x double> %0
@@ -3793,12 +3489,12 @@ entry:
declare <8 x double> @llvm.x86.avx512.vfmaddsub.pd.512(<8 x double>, <8 x double>, <8 x double>, i32) #1
define <8 x double> @test_mm512_mask_fmaddsub_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fmaddsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmaddsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmaddsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmaddsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmaddsub_round_pd:
; X64: # %bb.0: # %entry
@@ -3813,13 +3509,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmaddsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmaddsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmaddsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmaddsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmaddsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmaddsub_round_pd:
; X64: # %bb.0: # %entry
@@ -3835,12 +3531,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmaddsub_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fmaddsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmaddsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmaddsub_round_pd:
; X64: # %bb.0: # %entry
@@ -3855,11 +3551,11 @@ entry:
}
define <8 x double> @test_mm512_fmsubadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fmsubadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm2, %zmm2
-; X32-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fmsubadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxorq {{\.LCPI.*}}{1to8}, %zmm2, %zmm2
+; X86-NEXT: vfmaddsub213pd {rn-sae}, %zmm2, %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fmsubadd_round_pd:
; X64: # %bb.0: # %entry
@@ -3873,12 +3569,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fmsubadd_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fmsubadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsubadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmsubadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsubadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsubadd_round_pd:
; X64: # %bb.0: # %entry
@@ -3894,12 +3590,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmsubadd_round_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fmsubadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmsubadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsubadd213pd {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsubadd_round_pd:
; X64: # %bb.0: # %entry
@@ -3915,15 +3611,10 @@ entry:
}
define <8 x double> @test_mm512_fmaddsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fmaddsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmaddsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmaddsub_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmaddsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmaddsub_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmaddsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) #10
%1 = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
@@ -3933,12 +3624,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fmaddsub_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fmaddsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmaddsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmaddsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmaddsub132pd {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmaddsub_pd:
; X64: # %bb.0: # %entry
@@ -3956,13 +3647,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmaddsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmaddsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmaddsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmaddsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmaddsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmaddsub_pd:
; X64: # %bb.0: # %entry
@@ -3981,12 +3672,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmaddsub_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fmaddsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmaddsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmaddsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmaddsub213pd {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmaddsub_pd:
; X64: # %bb.0: # %entry
@@ -4004,15 +3695,10 @@ entry:
}
define <8 x double> @test_mm512_fmsubadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_fmsubadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmsubadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmsubadd_pd:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmsubadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmsubadd_pd:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmsubadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%sub.i = fsub <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %__C
%0 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %__A, <8 x double> %__B, <8 x double> %sub.i) #10
@@ -4022,12 +3708,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fmsubadd_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fmsubadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsubadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmsubadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsubadd132pd {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsubadd_pd:
; X64: # %bb.0: # %entry
@@ -4045,12 +3731,12 @@ entry:
}
define <8 x double> @test_mm512_maskz_fmsubadd_pd(i8 zeroext %__U, <8 x double> %__A, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_maskz_fmsubadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsubadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmsubadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsubadd213pd {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsubadd_pd:
; X64: # %bb.0: # %entry
@@ -4068,15 +3754,10 @@ entry:
}
define <16 x float> @test_mm512_fmaddsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fmaddsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmaddsub_round_ps:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmaddsub_round_ps:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i32 8)
ret <16 x float> %0
@@ -4085,11 +3766,11 @@ entry:
declare <16 x float> @llvm.x86.avx512.vfmaddsub.ps.512(<16 x float>, <16 x float>, <16 x float>, i32) #1
define <16 x float> @test_mm512_mask_fmaddsub_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fmaddsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmaddsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmaddsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmaddsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmaddsub_round_ps:
; X64: # %bb.0: # %entry
@@ -4104,12 +3785,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmaddsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmaddsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmaddsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmaddsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmaddsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmaddsub_round_ps:
; X64: # %bb.0: # %entry
@@ -4125,11 +3806,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmaddsub_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fmaddsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmaddsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmaddsub_round_ps:
; X64: # %bb.0: # %entry
@@ -4144,11 +3825,11 @@ entry:
}
define <16 x float> @test_mm512_fmsubadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fmsubadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm2, %zmm2
-; X32-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_fmsubadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: vpxord {{\.LCPI.*}}{1to16}, %zmm2, %zmm2
+; X86-NEXT: vfmaddsub213ps {rn-sae}, %zmm2, %zmm1, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_fmsubadd_round_ps:
; X64: # %bb.0: # %entry
@@ -4162,11 +3843,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fmsubadd_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fmsubadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsubadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmsubadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsubadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsubadd_round_ps:
; X64: # %bb.0: # %entry
@@ -4182,11 +3863,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmsubadd_round_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fmsubadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmsubadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsubadd213ps {rn-sae}, %zmm2, %zmm1, %zmm0 {%k1} {z}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsubadd_round_ps:
; X64: # %bb.0: # %entry
@@ -4202,15 +3883,10 @@ entry:
}
define <16 x float> @test_mm512_fmaddsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fmaddsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmaddsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmaddsub_ps:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmaddsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmaddsub_ps:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmaddsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) #10
%1 = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
@@ -4220,11 +3896,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fmaddsub_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fmaddsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmaddsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmaddsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmaddsub132ps {{.*#+}} zmm0 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmaddsub_ps:
; X64: # %bb.0: # %entry
@@ -4242,12 +3918,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmaddsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmaddsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmaddsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmaddsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmaddsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) +/- zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmaddsub_ps:
; X64: # %bb.0: # %entry
@@ -4266,11 +3942,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmaddsub_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fmaddsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmaddsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmaddsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmaddsub213ps {{.*#+}} zmm0 = (zmm1 * zmm0) +/- zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmaddsub_ps:
; X64: # %bb.0: # %entry
@@ -4288,15 +3964,10 @@ entry:
}
define <16 x float> @test_mm512_fmsubadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_fmsubadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: vfmsubadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
-; X32-NEXT: retl
-;
-; X64-LABEL: test_mm512_fmsubadd_ps:
-; X64: # %bb.0: # %entry
-; X64-NEXT: vfmsubadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
-; X64-NEXT: retq
+; CHECK-LABEL: test_mm512_fmsubadd_ps:
+; CHECK: # %bb.0: # %entry
+; CHECK-NEXT: vfmsubadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
+; CHECK-NEXT: ret{{[l|q]}}
entry:
%sub.i = fsub <16 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %__C
%0 = tail call <16 x float> @llvm.fma.v16f32(<16 x float> %__A, <16 x float> %__B, <16 x float> %sub.i) #10
@@ -4306,11 +3977,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fmsubadd_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fmsubadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsubadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fmsubadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsubadd132ps {{.*#+}} zmm0 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fmsubadd_ps:
; X64: # %bb.0: # %entry
@@ -4328,11 +3999,11 @@ entry:
}
define <16 x float> @test_mm512_maskz_fmsubadd_ps(i16 zeroext %__U, <16 x float> %__A, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_maskz_fmsubadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsubadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_maskz_fmsubadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsubadd213ps {{.*#+}} zmm0 = (zmm1 * zmm0) -/+ zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_maskz_fmsubadd_ps:
; X64: # %bb.0: # %entry
@@ -4350,13 +4021,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsub_round_pd:
; X64: # %bb.0: # %entry
@@ -4373,13 +4044,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub231pd {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsub_pd:
; X64: # %bb.0: # %entry
@@ -4396,12 +4067,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsub_round_ps:
; X64: # %bb.0: # %entry
@@ -4418,12 +4089,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsub231ps {{.*#+}} zmm2 = (zmm0 * zmm1) - zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsub_ps:
; X64: # %bb.0: # %entry
@@ -4440,13 +4111,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmsubadd_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmsubadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsubadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmsubadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsubadd231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsubadd_round_pd:
; X64: # %bb.0: # %entry
@@ -4463,13 +4134,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fmsubadd_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmsubadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsubadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmsubadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsubadd231pd {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsubadd_pd:
; X64: # %bb.0: # %entry
@@ -4488,12 +4159,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmsubadd_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmsubadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsubadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmsubadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsubadd231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsubadd_round_ps:
; X64: # %bb.0: # %entry
@@ -4510,12 +4181,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fmsubadd_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fmsubadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfmsubadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fmsubadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfmsubadd231ps {{.*#+}} zmm2 = (zmm0 * zmm1) -/+ zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fmsubadd_ps:
; X64: # %bb.0: # %entry
@@ -4534,12 +4205,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fnmadd_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fnmadd_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fnmadd_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmadd_round_pd:
; X64: # %bb.0: # %entry
@@ -4555,12 +4226,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fnmadd_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fnmadd_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fnmadd_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmadd_pd:
; X64: # %bb.0: # %entry
@@ -4576,11 +4247,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fnmadd_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fnmadd_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fnmadd_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmadd132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmadd_round_ps:
; X64: # %bb.0: # %entry
@@ -4596,11 +4267,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fnmadd_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fnmadd_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmadd132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fnmadd_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmadd132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) + zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmadd_ps:
; X64: # %bb.0: # %entry
@@ -4616,12 +4287,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fnmsub_round_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fnmsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fnmsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub132pd {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmsub_round_pd:
; X64: # %bb.0: # %entry
@@ -4638,13 +4309,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fnmsub_round_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fnmsub_round_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fnmsub_round_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub231pd {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmsub_round_pd:
; X64: # %bb.0: # %entry
@@ -4662,12 +4333,12 @@ entry:
}
define <8 x double> @test_mm512_mask_fnmsub_pd(<8 x double> %__A, i8 zeroext %__U, <8 x double> %__B, <8 x double> %__C) {
-; X32-LABEL: test_mm512_mask_fnmsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fnmsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub132pd {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmsub_pd:
; X64: # %bb.0: # %entry
@@ -4684,13 +4355,13 @@ entry:
}
define <8 x double> @test_mm512_mask3_fnmsub_pd(<8 x double> %__A, <8 x double> %__B, <8 x double> %__C, i8 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fnmsub_pd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
-; X32-NEXT: vmovapd %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fnmsub_pd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub231pd {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: vmovapd %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmsub_pd:
; X64: # %bb.0: # %entry
@@ -4708,11 +4379,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fnmsub_round_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fnmsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fnmsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmsub132ps {rn-sae}, %zmm1, %zmm2, %zmm0 {%k1}
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmsub_round_ps:
; X64: # %bb.0: # %entry
@@ -4729,12 +4400,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fnmsub_round_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fnmsub_round_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fnmsub_round_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmsub231ps {rn-sae}, %zmm1, %zmm0, %zmm2 {%k1}
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmsub_round_ps:
; X64: # %bb.0: # %entry
@@ -4752,11 +4423,11 @@ entry:
}
define <16 x float> @test_mm512_mask_fnmsub_ps(<16 x float> %__A, i16 zeroext %__U, <16 x float> %__B, <16 x float> %__C) {
-; X32-LABEL: test_mm512_mask_fnmsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmsub132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask_fnmsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmsub132ps {{.*#+}} zmm0 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask_fnmsub_ps:
; X64: # %bb.0: # %entry
@@ -4773,12 +4444,12 @@ entry:
}
define <16 x float> @test_mm512_mask3_fnmsub_ps(<16 x float> %__A, <16 x float> %__B, <16 x float> %__C, i16 zeroext %__U) {
-; X32-LABEL: test_mm512_mask3_fnmsub_ps:
-; X32: # %bb.0: # %entry
-; X32-NEXT: kmovw {{[0-9]+}}(%esp), %k1
-; X32-NEXT: vfnmsub231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
-; X32-NEXT: vmovaps %zmm2, %zmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm512_mask3_fnmsub_ps:
+; X86: # %bb.0: # %entry
+; X86-NEXT: kmovw {{[0-9]+}}(%esp), %k1
+; X86-NEXT: vfnmsub231ps {{.*#+}} zmm2 = -(zmm0 * zmm1) - zmm2
+; X86-NEXT: vmovaps %zmm2, %zmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm512_mask3_fnmsub_ps:
; X64: # %bb.0: # %entry
@@ -4796,12 +4467,12 @@ entry:
}
define <4 x float> @test_mm_mask_fmadd_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; X32-LABEL: test_mm_mask_fmadd_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fmadd_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmadd_ss:
; X64: # %bb.0: # %entry
@@ -4822,12 +4493,12 @@ entry:
}
define <4 x float> @test_mm_mask_fmadd_round_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; X32-LABEL: test_mm_mask_fmadd_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fmadd_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmadd_round_ss:
; X64: # %bb.0: # %entry
@@ -4842,12 +4513,12 @@ entry:
declare <4 x float> @llvm.x86.avx512.mask.vfmadd.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) #1
define <4 x float> @test_mm_maskz_fmadd_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; X32-LABEL: test_mm_maskz_fmadd_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fmadd_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmadd_ss:
; X64: # %bb.0: # %entry
@@ -4867,12 +4538,12 @@ entry:
}
define <4 x float> @test_mm_maskz_fmadd_round_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; X32-LABEL: test_mm_maskz_fmadd_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fmadd_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmadd_round_ss:
; X64: # %bb.0: # %entry
@@ -4887,13 +4558,13 @@ entry:
declare <4 x float> @llvm.x86.avx512.maskz.vfmadd.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) #1
define <4 x float> @test_mm_mask3_fmadd_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fmadd_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
-; X32-NEXT: vmovaps %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fmadd_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmadd_ss:
; X64: # %bb.0: # %entry
@@ -4915,13 +4586,13 @@ entry:
}
define <4 x float> @test_mm_mask3_fmadd_round_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fmadd_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
-; X32-NEXT: vmovaps %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fmadd_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmadd_round_ss:
; X64: # %bb.0: # %entry
@@ -4937,12 +4608,12 @@ entry:
declare <4 x float> @llvm.x86.avx512.mask3.vfmadd.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) #1
define <4 x float> @test_mm_mask_fmsub_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; X32-LABEL: test_mm_mask_fmsub_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fmsub_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsub_ss:
; X64: # %bb.0: # %entry
@@ -4964,14 +4635,14 @@ entry:
}
define <4 x float> @test_mm_mask_fmsub_round_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; X32-LABEL: test_mm_mask_fmsub_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; X32-NEXT: vxorps %xmm3, %xmm2, %xmm2
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fmsub_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: vxorps %xmm3, %xmm2, %xmm2
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsub_round_ss:
; X64: # %bb.0: # %entry
@@ -4987,12 +4658,12 @@ entry:
}
define <4 x float> @test_mm_maskz_fmsub_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; X32-LABEL: test_mm_maskz_fmsub_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fmsub_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsub_ss:
; X64: # %bb.0: # %entry
@@ -5013,14 +4684,14 @@ entry:
}
define <4 x float> @test_mm_maskz_fmsub_round_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; X32-LABEL: test_mm_maskz_fmsub_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; X32-NEXT: vxorps %xmm3, %xmm2, %xmm2
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fmsub_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: vxorps %xmm3, %xmm2, %xmm2
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsub_round_ss:
; X64: # %bb.0: # %entry
@@ -5036,13 +4707,13 @@ entry:
}
define <4 x float> @test_mm_mask3_fmsub_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fmsub_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub231ss {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
-; X32-NEXT: vmovaps %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fmsub_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub231ss {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsub_ss:
; X64: # %bb.0: # %entry
@@ -5065,13 +4736,13 @@ entry:
}
define <4 x float> @test_mm_mask3_fmsub_round_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fmsub_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub231ss {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
-; X32-NEXT: vmovaps %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fmsub_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub231ss {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsub_round_ss:
; X64: # %bb.0: # %entry
@@ -5087,12 +4758,12 @@ entry:
declare <4 x float> @llvm.x86.avx512.mask3.vfmsub.ss(<4 x float>, <4 x float>, <4 x float>, i8, i32) #1
define <4 x float> @test_mm_mask_fnmadd_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; X32-LABEL: test_mm_mask_fnmadd_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fnmadd_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmadd_ss:
; X64: # %bb.0: # %entry
@@ -5114,14 +4785,14 @@ entry:
}
define <4 x float> @test_mm_mask_fnmadd_round_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; X32-LABEL: test_mm_mask_fnmadd_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; X32-NEXT: vxorps %xmm3, %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fnmadd_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: vxorps %xmm3, %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmadd_round_ss:
; X64: # %bb.0: # %entry
@@ -5137,12 +4808,12 @@ entry:
}
define <4 x float> @test_mm_maskz_fnmadd_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; X32-LABEL: test_mm_maskz_fnmadd_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fnmadd_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmadd_ss:
; X64: # %bb.0: # %entry
@@ -5163,14 +4834,14 @@ entry:
}
define <4 x float> @test_mm_maskz_fnmadd_round_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; X32-LABEL: test_mm_maskz_fnmadd_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; X32-NEXT: vxorps %xmm3, %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fnmadd_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: vxorps %xmm3, %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmadd_round_ss:
; X64: # %bb.0: # %entry
@@ -5186,13 +4857,13 @@ entry:
}
define <4 x float> @test_mm_mask3_fnmadd_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fnmadd_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
-; X32-NEXT: vmovaps %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fnmadd_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmadd_ss:
; X64: # %bb.0: # %entry
@@ -5215,15 +4886,15 @@ entry:
}
define <4 x float> @test_mm_mask3_fnmadd_round_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fnmadd_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; X32-NEXT: vxorps %xmm3, %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
-; X32-NEXT: vmovaps %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fnmadd_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: vxorps %xmm3, %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmadd_round_ss:
; X64: # %bb.0: # %entry
@@ -5240,12 +4911,12 @@ entry:
}
define <4 x float> @test_mm_mask_fnmsub_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; X32-LABEL: test_mm_mask_fnmsub_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fnmsub_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmsub_ss:
; X64: # %bb.0: # %entry
@@ -5268,15 +4939,15 @@ entry:
}
define <4 x float> @test_mm_mask_fnmsub_round_ss(<4 x float> %__W, i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B) {
-; X32-LABEL: test_mm_mask_fnmsub_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; X32-NEXT: vxorps %xmm3, %xmm1, %xmm1
-; X32-NEXT: vxorps %xmm3, %xmm2, %xmm2
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fnmsub_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: vxorps %xmm3, %xmm1, %xmm1
+; X86-NEXT: vxorps %xmm3, %xmm2, %xmm2
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmsub_round_ss:
; X64: # %bb.0: # %entry
@@ -5294,12 +4965,12 @@ entry:
}
define <4 x float> @test_mm_maskz_fnmsub_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; X32-LABEL: test_mm_maskz_fnmsub_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fnmsub_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmsub_ss:
; X64: # %bb.0: # %entry
@@ -5321,15 +4992,15 @@ entry:
}
define <4 x float> @test_mm_maskz_fnmsub_round_ss(i8 zeroext %__U, <4 x float> %__A, <4 x float> %__B, <4 x float> %__C) {
-; X32-LABEL: test_mm_maskz_fnmsub_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; X32-NEXT: vxorps %xmm3, %xmm1, %xmm1
-; X32-NEXT: vxorps %xmm3, %xmm2, %xmm2
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fnmsub_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: vxorps %xmm3, %xmm1, %xmm1
+; X86-NEXT: vxorps %xmm3, %xmm2, %xmm2
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmsub_round_ss:
; X64: # %bb.0: # %entry
@@ -5347,13 +5018,13 @@ entry:
}
define <4 x float> @test_mm_mask3_fnmsub_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fnmsub_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
-; X32-NEXT: vmovaps %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fnmsub_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub231ss {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmsub_ss:
; X64: # %bb.0: # %entry
@@ -5377,15 +5048,15 @@ entry:
}
define <4 x float> @test_mm_mask3_fnmsub_round_ss(<4 x float> %__W, <4 x float> %__X, <4 x float> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fnmsub_round_ss:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
-; X32-NEXT: vxorps %xmm3, %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub231ss {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
-; X32-NEXT: vmovaps %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fnmsub_round_ss:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vbroadcastss {{.*#+}} xmm3 = [-0,-0,-0,-0]
+; X86-NEXT: vxorps %xmm3, %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub231ss {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovaps %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmsub_round_ss:
; X64: # %bb.0: # %entry
@@ -5402,12 +5073,12 @@ entry:
}
define <2 x double> @test_mm_mask_fmadd_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; X32-LABEL: test_mm_mask_fmadd_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fmadd_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmadd_sd:
; X64: # %bb.0: # %entry
@@ -5428,12 +5099,12 @@ entry:
}
define <2 x double> @test_mm_mask_fmadd_round_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; X32-LABEL: test_mm_mask_fmadd_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fmadd_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmadd_round_sd:
; X64: # %bb.0: # %entry
@@ -5448,12 +5119,12 @@ entry:
declare <2 x double> @llvm.x86.avx512.mask.vfmadd.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) #1
define <2 x double> @test_mm_maskz_fmadd_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; X32-LABEL: test_mm_maskz_fmadd_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fmadd_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmadd_sd:
; X64: # %bb.0: # %entry
@@ -5473,12 +5144,12 @@ entry:
}
define <2 x double> @test_mm_maskz_fmadd_round_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; X32-LABEL: test_mm_maskz_fmadd_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fmadd_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmadd_round_sd:
; X64: # %bb.0: # %entry
@@ -5493,13 +5164,13 @@ entry:
declare <2 x double> @llvm.x86.avx512.maskz.vfmadd.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) #1
define <2 x double> @test_mm_mask3_fmadd_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fmadd_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd231sd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
-; X32-NEXT: vmovapd %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fmadd_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd231sd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmadd_sd:
; X64: # %bb.0: # %entry
@@ -5521,13 +5192,13 @@ entry:
}
define <2 x double> @test_mm_mask3_fmadd_round_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fmadd_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd231sd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
-; X32-NEXT: vmovapd %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fmadd_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd231sd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmadd_round_sd:
; X64: # %bb.0: # %entry
@@ -5543,12 +5214,12 @@ entry:
declare <2 x double> @llvm.x86.avx512.mask3.vfmadd.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) #1
define <2 x double> @test_mm_mask_fmsub_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; X32-LABEL: test_mm_mask_fmsub_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fmsub_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsub_sd:
; X64: # %bb.0: # %entry
@@ -5570,13 +5241,13 @@ entry:
}
define <2 x double> @test_mm_mask_fmsub_round_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; X32-LABEL: test_mm_mask_fmsub_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vxorpd {{\.LCPI.*}}, %xmm2, %xmm2
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fmsub_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vxorpd {{\.LCPI.*}}, %xmm2, %xmm2
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fmsub_round_sd:
; X64: # %bb.0: # %entry
@@ -5591,12 +5262,12 @@ entry:
}
define <2 x double> @test_mm_maskz_fmsub_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; X32-LABEL: test_mm_maskz_fmsub_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fmsub_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsub_sd:
; X64: # %bb.0: # %entry
@@ -5617,13 +5288,13 @@ entry:
}
define <2 x double> @test_mm_maskz_fmsub_round_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; X32-LABEL: test_mm_maskz_fmsub_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vxorpd {{\.LCPI.*}}, %xmm2, %xmm2
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fmsub_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vxorpd {{\.LCPI.*}}, %xmm2, %xmm2
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fmsub_round_sd:
; X64: # %bb.0: # %entry
@@ -5638,13 +5309,13 @@ entry:
}
define <2 x double> @test_mm_mask3_fmsub_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fmsub_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub231sd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
-; X32-NEXT: vmovapd %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fmsub_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub231sd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsub_sd:
; X64: # %bb.0: # %entry
@@ -5667,13 +5338,13 @@ entry:
}
define <2 x double> @test_mm_mask3_fmsub_round_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fmsub_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub231sd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
-; X32-NEXT: vmovapd %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fmsub_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub231sd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fmsub_round_sd:
; X64: # %bb.0: # %entry
@@ -5689,12 +5360,12 @@ entry:
declare <2 x double> @llvm.x86.avx512.mask3.vfmsub.sd(<2 x double>, <2 x double>, <2 x double>, i8, i32) #1
define <2 x double> @test_mm_mask_fnmadd_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; X32-LABEL: test_mm_mask_fnmadd_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fnmadd_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmadd_sd:
; X64: # %bb.0: # %entry
@@ -5716,13 +5387,13 @@ entry:
}
define <2 x double> @test_mm_mask_fnmadd_round_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; X32-LABEL: test_mm_mask_fnmadd_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vxorpd {{\.LCPI.*}}, %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fnmadd_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vxorpd {{\.LCPI.*}}, %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmadd_round_sd:
; X64: # %bb.0: # %entry
@@ -5737,12 +5408,12 @@ entry:
}
define <2 x double> @test_mm_maskz_fnmadd_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; X32-LABEL: test_mm_maskz_fnmadd_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fnmadd_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmadd_sd:
; X64: # %bb.0: # %entry
@@ -5763,13 +5434,13 @@ entry:
}
define <2 x double> @test_mm_maskz_fnmadd_round_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; X32-LABEL: test_mm_maskz_fnmadd_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vxorpd {{\.LCPI.*}}, %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fnmadd_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vxorpd {{\.LCPI.*}}, %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmadd_round_sd:
; X64: # %bb.0: # %entry
@@ -5784,13 +5455,13 @@ entry:
}
define <2 x double> @test_mm_mask3_fnmadd_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fnmadd_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmadd231sd {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
-; X32-NEXT: vmovapd %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fnmadd_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmadd231sd {{.*#+}} xmm2 = -(xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmadd_sd:
; X64: # %bb.0: # %entry
@@ -5813,14 +5484,14 @@ entry:
}
define <2 x double> @test_mm_mask3_fnmadd_round_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fnmadd_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vxorpd {{\.LCPI.*}}, %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd231sd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
-; X32-NEXT: vmovapd %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fnmadd_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vxorpd {{\.LCPI.*}}, %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd231sd {{.*#+}} xmm2 = (xmm0 * xmm1) + xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmadd_round_sd:
; X64: # %bb.0: # %entry
@@ -5836,12 +5507,12 @@ entry:
}
define <2 x double> @test_mm_mask_fnmsub_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; X32-LABEL: test_mm_mask_fnmsub_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fnmsub_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmsub_sd:
; X64: # %bb.0: # %entry
@@ -5864,15 +5535,15 @@ entry:
}
define <2 x double> @test_mm_mask_fnmsub_round_sd(<2 x double> %__W, i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B) {
-; X32-LABEL: test_mm_mask_fnmsub_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vmovapd {{.*#+}} xmm3 = [-0.000000e+00,-0.000000e+00]
-; X32-NEXT: vxorpd %xmm3, %xmm1, %xmm1
-; X32-NEXT: vxorpd %xmm3, %xmm2, %xmm2
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask_fnmsub_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vmovapd {{.*#+}} xmm3 = [-0.000000e+00,-0.000000e+00]
+; X86-NEXT: vxorpd %xmm3, %xmm1, %xmm1
+; X86-NEXT: vxorpd %xmm3, %xmm2, %xmm2
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask_fnmsub_round_sd:
; X64: # %bb.0: # %entry
@@ -5890,12 +5561,12 @@ entry:
}
define <2 x double> @test_mm_maskz_fnmsub_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; X32-LABEL: test_mm_maskz_fnmsub_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fnmsub_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmsub_sd:
; X64: # %bb.0: # %entry
@@ -5917,15 +5588,15 @@ entry:
}
define <2 x double> @test_mm_maskz_fnmsub_round_sd(i8 zeroext %__U, <2 x double> %__A, <2 x double> %__B, <2 x double> %__C) {
-; X32-LABEL: test_mm_maskz_fnmsub_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vmovapd {{.*#+}} xmm3 = [-0.000000e+00,-0.000000e+00]
-; X32-NEXT: vxorpd %xmm3, %xmm1, %xmm1
-; X32-NEXT: vxorpd %xmm3, %xmm2, %xmm2
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
-; X32-NEXT: retl
+; X86-LABEL: test_mm_maskz_fnmsub_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vmovapd {{.*#+}} xmm3 = [-0.000000e+00,-0.000000e+00]
+; X86-NEXT: vxorpd %xmm3, %xmm1, %xmm1
+; X86-NEXT: vxorpd %xmm3, %xmm2, %xmm2
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_maskz_fnmsub_round_sd:
; X64: # %bb.0: # %entry
@@ -5943,13 +5614,13 @@ entry:
}
define <2 x double> @test_mm_mask3_fnmsub_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fnmsub_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfnmsub231sd {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
-; X32-NEXT: vmovapd %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fnmsub_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfnmsub231sd {{.*#+}} xmm2 = -(xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmsub_sd:
; X64: # %bb.0: # %entry
@@ -5973,14 +5644,14 @@ entry:
}
define <2 x double> @test_mm_mask3_fnmsub_round_sd(<2 x double> %__W, <2 x double> %__X, <2 x double> %__Y, i8 zeroext %__U) {
-; X32-LABEL: test_mm_mask3_fnmsub_round_sd:
-; X32: # %bb.0: # %entry
-; X32-NEXT: movb {{[0-9]+}}(%esp), %al
-; X32-NEXT: vxorpd {{\.LCPI.*}}, %xmm1, %xmm1
-; X32-NEXT: kmovw %eax, %k1
-; X32-NEXT: vfmsub231sd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
-; X32-NEXT: vmovapd %xmm2, %xmm0
-; X32-NEXT: retl
+; X86-LABEL: test_mm_mask3_fnmsub_round_sd:
+; X86: # %bb.0: # %entry
+; X86-NEXT: movb {{[0-9]+}}(%esp), %al
+; X86-NEXT: vxorpd {{\.LCPI.*}}, %xmm1, %xmm1
+; X86-NEXT: kmovw %eax, %k1
+; X86-NEXT: vfmsub231sd {{.*#+}} xmm2 = (xmm0 * xmm1) - xmm2
+; X86-NEXT: vmovapd %xmm2, %xmm0
+; X86-NEXT: retl
;
; X64-LABEL: test_mm_mask3_fnmsub_round_sd:
; X64: # %bb.0: # %entry
More information about the llvm-commits
mailing list