[llvm] r270315 - [AVX512] Use update_llc_test_checks to update some tests so we can see all the instruction encodings and ensure everything is with EVEX.
Craig Topper via llvm-commits
llvm-commits at lists.llvm.org
Fri May 20 22:46:59 PDT 2016
Author: ctopper
Date: Sat May 21 00:46:58 2016
New Revision: 270315
URL: http://llvm.org/viewvc/llvm-project?rev=270315&view=rev
Log:
[AVX512] Use update_llc_test_checks to update some tests so we can see all the instruction encodings and ensure everything is with EVEX.
Modified:
llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics.ll
llvm/trunk/test/CodeGen/X86/avx512vl-intrinsics.ll
Modified: llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll?rev=270315&r1=270314&r2=270315&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512bwvl-intrinsics.ll Sat May 21 00:46:58 2016
@@ -1,17 +1,25 @@
+; NOTE: Assertions have been autogenerated by update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl -mattr=+avx512bw -mattr=+avx512vl --show-mc-encoding| FileCheck %s
; 256-bit
define i32 @test_pcmpeq_b_256(<32 x i8> %a, <32 x i8> %b) {
-; CHECK-LABEL: test_pcmpeq_b_256
-; CHECK: vpcmpeqb %ymm1, %ymm0, %k0 ##
+; CHECK-LABEL: test_pcmpeq_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpeqb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x28,0x74,0xc1]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i32 @llvm.x86.avx512.mask.pcmpeq.b.256(<32 x i8> %a, <32 x i8> %b, i32 -1)
ret i32 %res
}
define i32 @test_mask_pcmpeq_b_256(<32 x i8> %a, <32 x i8> %b, i32 %mask) {
-; CHECK-LABEL: test_mask_pcmpeq_b_256
-; CHECK: vpcmpeqb %ymm1, %ymm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_pcmpeq_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpcmpeqb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x74,0xc1]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i32 @llvm.x86.avx512.mask.pcmpeq.b.256(<32 x i8> %a, <32 x i8> %b, i32 %mask)
ret i32 %res
}
@@ -19,15 +27,22 @@ define i32 @test_mask_pcmpeq_b_256(<32 x
declare i32 @llvm.x86.avx512.mask.pcmpeq.b.256(<32 x i8>, <32 x i8>, i32)
define i16 @test_pcmpeq_w_256(<16 x i16> %a, <16 x i16> %b) {
-; CHECK-LABEL: test_pcmpeq_w_256
-; CHECK: vpcmpeqw %ymm1, %ymm0, %k0 ##
+; CHECK-LABEL: test_pcmpeq_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpeqw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x28,0x75,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.mask.pcmpeq.w.256(<16 x i16> %a, <16 x i16> %b, i16 -1)
ret i16 %res
}
define i16 @test_mask_pcmpeq_w_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
-; CHECK-LABEL: test_mask_pcmpeq_w_256
-; CHECK: vpcmpeqw %ymm1, %ymm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_pcmpeq_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpeqw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x75,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.mask.pcmpeq.w.256(<16 x i16> %a, <16 x i16> %b, i16 %mask)
ret i16 %res
}
@@ -35,15 +50,22 @@ define i16 @test_mask_pcmpeq_w_256(<16 x
declare i16 @llvm.x86.avx512.mask.pcmpeq.w.256(<16 x i16>, <16 x i16>, i16)
define i32 @test_pcmpgt_b_256(<32 x i8> %a, <32 x i8> %b) {
-; CHECK-LABEL: test_pcmpgt_b_256
-; CHECK: vpcmpgtb %ymm1, %ymm0, %k0 ##
+; CHECK-LABEL: test_pcmpgt_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpgtb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x28,0x64,0xc1]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i32 @llvm.x86.avx512.mask.pcmpgt.b.256(<32 x i8> %a, <32 x i8> %b, i32 -1)
ret i32 %res
}
define i32 @test_mask_pcmpgt_b_256(<32 x i8> %a, <32 x i8> %b, i32 %mask) {
-; CHECK-LABEL: test_mask_pcmpgt_b_256
-; CHECK: vpcmpgtb %ymm1, %ymm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_pcmpgt_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpcmpgtb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x64,0xc1]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i32 @llvm.x86.avx512.mask.pcmpgt.b.256(<32 x i8> %a, <32 x i8> %b, i32 %mask)
ret i32 %res
}
@@ -51,15 +73,22 @@ define i32 @test_mask_pcmpgt_b_256(<32 x
declare i32 @llvm.x86.avx512.mask.pcmpgt.b.256(<32 x i8>, <32 x i8>, i32)
define i16 @test_pcmpgt_w_256(<16 x i16> %a, <16 x i16> %b) {
-; CHECK-LABEL: test_pcmpgt_w_256
-; CHECK: vpcmpgtw %ymm1, %ymm0, %k0 ##
+; CHECK-LABEL: test_pcmpgt_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpgtw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x28,0x65,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.mask.pcmpgt.w.256(<16 x i16> %a, <16 x i16> %b, i16 -1)
ret i16 %res
}
define i16 @test_mask_pcmpgt_w_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
-; CHECK-LABEL: test_mask_pcmpgt_w_256
-; CHECK: vpcmpgtw %ymm1, %ymm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_pcmpgt_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpgtw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x65,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.mask.pcmpgt.w.256(<16 x i16> %a, <16 x i16> %b, i16 %mask)
ret i16 %res
}
@@ -67,58 +96,97 @@ define i16 @test_mask_pcmpgt_w_256(<16 x
declare i16 @llvm.x86.avx512.mask.pcmpgt.w.256(<16 x i16>, <16 x i16>, i16)
define <8 x i32> @test_cmp_b_256(<32 x i8> %a0, <32 x i8> %a1) {
-; CHECK-LABEL: test_cmp_b_256
-; CHECK: vpcmpeqb %ymm1, %ymm0, %k0 ##
+; CHECK-LABEL: test_cmp_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpeqb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3f,0xc1,0x00]
+; CHECK-NEXT: kmovd %k0, %r8d ## encoding: [0xc5,0x7b,0x93,0xc0]
+; CHECK-NEXT: vpcmpltb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3f,0xc1,0x01]
+; CHECK-NEXT: kmovd %k0, %r9d ## encoding: [0xc5,0x7b,0x93,0xc8]
+; CHECK-NEXT: vpcmpleb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3f,0xc1,0x02]
+; CHECK-NEXT: kmovd %k0, %r10d ## encoding: [0xc5,0x7b,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3f,0xc1,0x03]
+; CHECK-NEXT: kmovd %k0, %esi ## encoding: [0xc5,0xfb,0x93,0xf0]
+; CHECK-NEXT: vpcmpneqb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3f,0xc1,0x04]
+; CHECK-NEXT: kmovd %k0, %edi ## encoding: [0xc5,0xfb,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3f,0xc1,0x05]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3f,0xc1,0x06]
+; CHECK-NEXT: kmovd %k0, %ecx ## encoding: [0xc5,0xfb,0x93,0xc8]
+; CHECK-NEXT: vpcmpordb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3f,0xc1,0x07]
+; CHECK-NEXT: kmovd %k0, %edx ## encoding: [0xc5,0xfb,0x93,0xd0]
+; CHECK-NEXT: vmovd %edi, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6e,0xc7]
+; CHECK-NEXT: vpinsrd $1, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc0,0x01]
+; CHECK-NEXT: vpinsrd $2, %ecx, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc1,0x02]
+; CHECK-NEXT: vpinsrd $3, %edx, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc2,0x03]
+; CHECK-NEXT: vmovd %r8d, %xmm1 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc8]
+; CHECK-NEXT: vpinsrd $1, %r9d, %xmm1, %xmm1 ## encoding: [0xc4,0xc3,0x71,0x22,0xc9,0x01]
+; CHECK-NEXT: vpinsrd $2, %r10d, %xmm1, %xmm1 ## encoding: [0xc4,0xc3,0x71,0x22,0xca,0x02]
+; CHECK-NEXT: vpinsrd $3, %esi, %xmm1, %xmm1 ## encoding: [0xc4,0xe3,0x71,0x22,0xce,0x03]
+; CHECK-NEXT: vinserti32x4 $1, %xmm0, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0x75,0x28,0x38,0xc0,0x01]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 0, i32 -1)
%vec0 = insertelement <8 x i32> undef, i32 %res0, i32 0
-; CHECK: vpcmpltb %ymm1, %ymm0, %k0 ##
%res1 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 1, i32 -1)
%vec1 = insertelement <8 x i32> %vec0, i32 %res1, i32 1
-; CHECK: vpcmpleb %ymm1, %ymm0, %k0 ##
%res2 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 2, i32 -1)
%vec2 = insertelement <8 x i32> %vec1, i32 %res2, i32 2
-; CHECK: vpcmpunordb %ymm1, %ymm0, %k0 ##
%res3 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 3, i32 -1)
%vec3 = insertelement <8 x i32> %vec2, i32 %res3, i32 3
-; CHECK: vpcmpneqb %ymm1, %ymm0, %k0 ##
%res4 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 4, i32 -1)
%vec4 = insertelement <8 x i32> %vec3, i32 %res4, i32 4
-; CHECK: vpcmpnltb %ymm1, %ymm0, %k0 ##
%res5 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 5, i32 -1)
%vec5 = insertelement <8 x i32> %vec4, i32 %res5, i32 5
-; CHECK: vpcmpnleb %ymm1, %ymm0, %k0 ##
%res6 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 6, i32 -1)
%vec6 = insertelement <8 x i32> %vec5, i32 %res6, i32 6
-; CHECK: vpcmpordb %ymm1, %ymm0, %k0 ##
%res7 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 7, i32 -1)
%vec7 = insertelement <8 x i32> %vec6, i32 %res7, i32 7
ret <8 x i32> %vec7
}
define <8 x i32> @test_mask_cmp_b_256(<32 x i8> %a0, <32 x i8> %a1, i32 %mask) {
-; CHECK-LABEL: test_mask_cmp_b_256
-; CHECK: vpcmpeqb %ymm1, %ymm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_cmp_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpcmpeqb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3f,0xc1,0x00]
+; CHECK-NEXT: kmovd %k0, %r8d ## encoding: [0xc5,0x7b,0x93,0xc0]
+; CHECK-NEXT: vpcmpltb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3f,0xc1,0x01]
+; CHECK-NEXT: kmovd %k0, %r9d ## encoding: [0xc5,0x7b,0x93,0xc8]
+; CHECK-NEXT: vpcmpleb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3f,0xc1,0x02]
+; CHECK-NEXT: kmovd %k0, %r10d ## encoding: [0xc5,0x7b,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3f,0xc1,0x03]
+; CHECK-NEXT: kmovd %k0, %esi ## encoding: [0xc5,0xfb,0x93,0xf0]
+; CHECK-NEXT: vpcmpneqb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3f,0xc1,0x04]
+; CHECK-NEXT: kmovd %k0, %edi ## encoding: [0xc5,0xfb,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3f,0xc1,0x05]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3f,0xc1,0x06]
+; CHECK-NEXT: kmovd %k0, %ecx ## encoding: [0xc5,0xfb,0x93,0xc8]
+; CHECK-NEXT: vpcmpordb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3f,0xc1,0x07]
+; CHECK-NEXT: kmovd %k0, %edx ## encoding: [0xc5,0xfb,0x93,0xd0]
+; CHECK-NEXT: vmovd %edi, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6e,0xc7]
+; CHECK-NEXT: vpinsrd $1, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc0,0x01]
+; CHECK-NEXT: vpinsrd $2, %ecx, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc1,0x02]
+; CHECK-NEXT: vpinsrd $3, %edx, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc2,0x03]
+; CHECK-NEXT: vmovd %r8d, %xmm1 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc8]
+; CHECK-NEXT: vpinsrd $1, %r9d, %xmm1, %xmm1 ## encoding: [0xc4,0xc3,0x71,0x22,0xc9,0x01]
+; CHECK-NEXT: vpinsrd $2, %r10d, %xmm1, %xmm1 ## encoding: [0xc4,0xc3,0x71,0x22,0xca,0x02]
+; CHECK-NEXT: vpinsrd $3, %esi, %xmm1, %xmm1 ## encoding: [0xc4,0xe3,0x71,0x22,0xce,0x03]
+; CHECK-NEXT: vinserti32x4 $1, %xmm0, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0x75,0x28,0x38,0xc0,0x01]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 0, i32 %mask)
%vec0 = insertelement <8 x i32> undef, i32 %res0, i32 0
-; CHECK: vpcmpltb %ymm1, %ymm0, %k0 {%k1} ##
%res1 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 1, i32 %mask)
%vec1 = insertelement <8 x i32> %vec0, i32 %res1, i32 1
-; CHECK: vpcmpleb %ymm1, %ymm0, %k0 {%k1} ##
%res2 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 2, i32 %mask)
%vec2 = insertelement <8 x i32> %vec1, i32 %res2, i32 2
-; CHECK: vpcmpunordb %ymm1, %ymm0, %k0 {%k1} ##
%res3 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 3, i32 %mask)
%vec3 = insertelement <8 x i32> %vec2, i32 %res3, i32 3
-; CHECK: vpcmpneqb %ymm1, %ymm0, %k0 {%k1} ##
%res4 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 4, i32 %mask)
%vec4 = insertelement <8 x i32> %vec3, i32 %res4, i32 4
-; CHECK: vpcmpnltb %ymm1, %ymm0, %k0 {%k1} ##
%res5 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 5, i32 %mask)
%vec5 = insertelement <8 x i32> %vec4, i32 %res5, i32 5
-; CHECK: vpcmpnleb %ymm1, %ymm0, %k0 {%k1} ##
%res6 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 6, i32 %mask)
%vec6 = insertelement <8 x i32> %vec5, i32 %res6, i32 6
-; CHECK: vpcmpordb %ymm1, %ymm0, %k0 {%k1} ##
%res7 = call i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 7, i32 %mask)
%vec7 = insertelement <8 x i32> %vec6, i32 %res7, i32 7
ret <8 x i32> %vec7
@@ -127,58 +195,97 @@ define <8 x i32> @test_mask_cmp_b_256(<3
declare i32 @llvm.x86.avx512.mask.cmp.b.256(<32 x i8>, <32 x i8>, i32, i32) nounwind readnone
define <8 x i32> @test_ucmp_b_256(<32 x i8> %a0, <32 x i8> %a1) {
-; CHECK-LABEL: test_ucmp_b_256
-; CHECK: vpcmpequb %ymm1, %ymm0, %k0 ##
+; CHECK-LABEL: test_ucmp_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpequb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3e,0xc1,0x00]
+; CHECK-NEXT: kmovd %k0, %r8d ## encoding: [0xc5,0x7b,0x93,0xc0]
+; CHECK-NEXT: vpcmpltub %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3e,0xc1,0x01]
+; CHECK-NEXT: kmovd %k0, %r9d ## encoding: [0xc5,0x7b,0x93,0xc8]
+; CHECK-NEXT: vpcmpleub %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3e,0xc1,0x02]
+; CHECK-NEXT: kmovd %k0, %r10d ## encoding: [0xc5,0x7b,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordub %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3e,0xc1,0x03]
+; CHECK-NEXT: kmovd %k0, %esi ## encoding: [0xc5,0xfb,0x93,0xf0]
+; CHECK-NEXT: vpcmpnequb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3e,0xc1,0x04]
+; CHECK-NEXT: kmovd %k0, %edi ## encoding: [0xc5,0xfb,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltub %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3e,0xc1,0x05]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleub %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3e,0xc1,0x06]
+; CHECK-NEXT: kmovd %k0, %ecx ## encoding: [0xc5,0xfb,0x93,0xc8]
+; CHECK-NEXT: vpcmpordub %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x3e,0xc1,0x07]
+; CHECK-NEXT: kmovd %k0, %edx ## encoding: [0xc5,0xfb,0x93,0xd0]
+; CHECK-NEXT: vmovd %edi, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6e,0xc7]
+; CHECK-NEXT: vpinsrd $1, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc0,0x01]
+; CHECK-NEXT: vpinsrd $2, %ecx, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc1,0x02]
+; CHECK-NEXT: vpinsrd $3, %edx, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc2,0x03]
+; CHECK-NEXT: vmovd %r8d, %xmm1 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc8]
+; CHECK-NEXT: vpinsrd $1, %r9d, %xmm1, %xmm1 ## encoding: [0xc4,0xc3,0x71,0x22,0xc9,0x01]
+; CHECK-NEXT: vpinsrd $2, %r10d, %xmm1, %xmm1 ## encoding: [0xc4,0xc3,0x71,0x22,0xca,0x02]
+; CHECK-NEXT: vpinsrd $3, %esi, %xmm1, %xmm1 ## encoding: [0xc4,0xe3,0x71,0x22,0xce,0x03]
+; CHECK-NEXT: vinserti32x4 $1, %xmm0, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0x75,0x28,0x38,0xc0,0x01]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 0, i32 -1)
%vec0 = insertelement <8 x i32> undef, i32 %res0, i32 0
-; CHECK: vpcmpltub %ymm1, %ymm0, %k0 ##
%res1 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 1, i32 -1)
%vec1 = insertelement <8 x i32> %vec0, i32 %res1, i32 1
-; CHECK: vpcmpleub %ymm1, %ymm0, %k0 ##
%res2 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 2, i32 -1)
%vec2 = insertelement <8 x i32> %vec1, i32 %res2, i32 2
-; CHECK: vpcmpunordub %ymm1, %ymm0, %k0 ##
%res3 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 3, i32 -1)
%vec3 = insertelement <8 x i32> %vec2, i32 %res3, i32 3
-; CHECK: vpcmpnequb %ymm1, %ymm0, %k0 ##
%res4 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 4, i32 -1)
%vec4 = insertelement <8 x i32> %vec3, i32 %res4, i32 4
-; CHECK: vpcmpnltub %ymm1, %ymm0, %k0 ##
%res5 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 5, i32 -1)
%vec5 = insertelement <8 x i32> %vec4, i32 %res5, i32 5
-; CHECK: vpcmpnleub %ymm1, %ymm0, %k0 ##
%res6 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 6, i32 -1)
%vec6 = insertelement <8 x i32> %vec5, i32 %res6, i32 6
-; CHECK: vpcmpordub %ymm1, %ymm0, %k0 ##
%res7 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 7, i32 -1)
%vec7 = insertelement <8 x i32> %vec6, i32 %res7, i32 7
ret <8 x i32> %vec7
}
define <8 x i32> @test_mask_ucmp_b_256(<32 x i8> %a0, <32 x i8> %a1, i32 %mask) {
-; CHECK-LABEL: test_mask_ucmp_b_256
-; CHECK: vpcmpequb %ymm1, %ymm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_ucmp_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpcmpequb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3e,0xc1,0x00]
+; CHECK-NEXT: kmovd %k0, %r8d ## encoding: [0xc5,0x7b,0x93,0xc0]
+; CHECK-NEXT: vpcmpltub %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3e,0xc1,0x01]
+; CHECK-NEXT: kmovd %k0, %r9d ## encoding: [0xc5,0x7b,0x93,0xc8]
+; CHECK-NEXT: vpcmpleub %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3e,0xc1,0x02]
+; CHECK-NEXT: kmovd %k0, %r10d ## encoding: [0xc5,0x7b,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordub %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3e,0xc1,0x03]
+; CHECK-NEXT: kmovd %k0, %esi ## encoding: [0xc5,0xfb,0x93,0xf0]
+; CHECK-NEXT: vpcmpnequb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3e,0xc1,0x04]
+; CHECK-NEXT: kmovd %k0, %edi ## encoding: [0xc5,0xfb,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltub %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3e,0xc1,0x05]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleub %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3e,0xc1,0x06]
+; CHECK-NEXT: kmovd %k0, %ecx ## encoding: [0xc5,0xfb,0x93,0xc8]
+; CHECK-NEXT: vpcmpordub %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x3e,0xc1,0x07]
+; CHECK-NEXT: kmovd %k0, %edx ## encoding: [0xc5,0xfb,0x93,0xd0]
+; CHECK-NEXT: vmovd %edi, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6e,0xc7]
+; CHECK-NEXT: vpinsrd $1, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc0,0x01]
+; CHECK-NEXT: vpinsrd $2, %ecx, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc1,0x02]
+; CHECK-NEXT: vpinsrd $3, %edx, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x22,0xc2,0x03]
+; CHECK-NEXT: vmovd %r8d, %xmm1 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc8]
+; CHECK-NEXT: vpinsrd $1, %r9d, %xmm1, %xmm1 ## encoding: [0xc4,0xc3,0x71,0x22,0xc9,0x01]
+; CHECK-NEXT: vpinsrd $2, %r10d, %xmm1, %xmm1 ## encoding: [0xc4,0xc3,0x71,0x22,0xca,0x02]
+; CHECK-NEXT: vpinsrd $3, %esi, %xmm1, %xmm1 ## encoding: [0xc4,0xe3,0x71,0x22,0xce,0x03]
+; CHECK-NEXT: vinserti32x4 $1, %xmm0, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0x75,0x28,0x38,0xc0,0x01]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 0, i32 %mask)
%vec0 = insertelement <8 x i32> undef, i32 %res0, i32 0
-; CHECK: vpcmpltub %ymm1, %ymm0, %k0 {%k1} ##
%res1 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 1, i32 %mask)
%vec1 = insertelement <8 x i32> %vec0, i32 %res1, i32 1
-; CHECK: vpcmpleub %ymm1, %ymm0, %k0 {%k1} ##
%res2 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 2, i32 %mask)
%vec2 = insertelement <8 x i32> %vec1, i32 %res2, i32 2
-; CHECK: vpcmpunordub %ymm1, %ymm0, %k0 {%k1} ##
%res3 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 3, i32 %mask)
%vec3 = insertelement <8 x i32> %vec2, i32 %res3, i32 3
-; CHECK: vpcmpnequb %ymm1, %ymm0, %k0 {%k1} ##
%res4 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 4, i32 %mask)
%vec4 = insertelement <8 x i32> %vec3, i32 %res4, i32 4
-; CHECK: vpcmpnltub %ymm1, %ymm0, %k0 {%k1} ##
%res5 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 5, i32 %mask)
%vec5 = insertelement <8 x i32> %vec4, i32 %res5, i32 5
-; CHECK: vpcmpnleub %ymm1, %ymm0, %k0 {%k1} ##
%res6 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 6, i32 %mask)
%vec6 = insertelement <8 x i32> %vec5, i32 %res6, i32 6
-; CHECK: vpcmpordub %ymm1, %ymm0, %k0 {%k1} ##
%res7 = call i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8> %a0, <32 x i8> %a1, i32 7, i32 %mask)
%vec7 = insertelement <8 x i32> %vec6, i32 %res7, i32 7
ret <8 x i32> %vec7
@@ -187,58 +294,95 @@ define <8 x i32> @test_mask_ucmp_b_256(<
declare i32 @llvm.x86.avx512.mask.ucmp.b.256(<32 x i8>, <32 x i8>, i32, i32) nounwind readnone
define <8 x i16> @test_cmp_w_256(<16 x i16> %a0, <16 x i16> %a1) {
-; CHECK-LABEL: test_cmp_w_256
-; CHECK: vpcmpeqw %ymm1, %ymm0, %k0 ##
+; CHECK-LABEL: test_cmp_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpeqw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3f,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3f,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmplew %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3f,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3f,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %esi ## encoding: [0xc5,0xf8,0x93,0xf0]
+; CHECK-NEXT: vpcmpneqw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3f,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3f,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnlew %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3f,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmpordw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3f,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: vmovd %r8d, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc0]
+; CHECK-NEXT: vpinsrw $1, %r9d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc1,0x01]
+; CHECK-NEXT: vpinsrw $2, %r10d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc2,0x02]
+; CHECK-NEXT: vpinsrw $3, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc6,0x03]
+; CHECK-NEXT: vpinsrw $4, %edi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc7,0x04]
+; CHECK-NEXT: vpinsrw $5, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc0,0x05]
+; CHECK-NEXT: vpinsrw $6, %ecx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc1,0x06]
+; CHECK-NEXT: vpinsrw $7, %edx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc2,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 0, i16 -1)
%vec0 = insertelement <8 x i16> undef, i16 %res0, i32 0
-; CHECK: vpcmpltw %ymm1, %ymm0, %k0 ##
%res1 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 1, i16 -1)
%vec1 = insertelement <8 x i16> %vec0, i16 %res1, i32 1
-; CHECK: vpcmplew %ymm1, %ymm0, %k0 ##
%res2 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 2, i16 -1)
%vec2 = insertelement <8 x i16> %vec1, i16 %res2, i32 2
-; CHECK: vpcmpunordw %ymm1, %ymm0, %k0 ##
%res3 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 3, i16 -1)
%vec3 = insertelement <8 x i16> %vec2, i16 %res3, i32 3
-; CHECK: vpcmpneqw %ymm1, %ymm0, %k0 ##
%res4 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 4, i16 -1)
%vec4 = insertelement <8 x i16> %vec3, i16 %res4, i32 4
-; CHECK: vpcmpnltw %ymm1, %ymm0, %k0 ##
%res5 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 5, i16 -1)
%vec5 = insertelement <8 x i16> %vec4, i16 %res5, i32 5
-; CHECK: vpcmpnlew %ymm1, %ymm0, %k0 ##
%res6 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 6, i16 -1)
%vec6 = insertelement <8 x i16> %vec5, i16 %res6, i32 6
-; CHECK: vpcmpordw %ymm1, %ymm0, %k0 ##
%res7 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 7, i16 -1)
%vec7 = insertelement <8 x i16> %vec6, i16 %res7, i32 7
ret <8 x i16> %vec7
}
define <8 x i16> @test_mask_cmp_w_256(<16 x i16> %a0, <16 x i16> %a1, i16 %mask) {
-; CHECK-LABEL: test_mask_cmp_w_256
-; CHECK: vpcmpeqw %ymm1, %ymm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_cmp_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpeqw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3f,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3f,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmplew %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3f,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3f,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %esi ## encoding: [0xc5,0xf8,0x93,0xf0]
+; CHECK-NEXT: vpcmpneqw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3f,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3f,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnlew %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3f,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmpordw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3f,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: vmovd %r8d, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc0]
+; CHECK-NEXT: vpinsrw $1, %r9d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc1,0x01]
+; CHECK-NEXT: vpinsrw $2, %r10d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc2,0x02]
+; CHECK-NEXT: vpinsrw $3, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc6,0x03]
+; CHECK-NEXT: vpinsrw $4, %edi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc7,0x04]
+; CHECK-NEXT: vpinsrw $5, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc0,0x05]
+; CHECK-NEXT: vpinsrw $6, %ecx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc1,0x06]
+; CHECK-NEXT: vpinsrw $7, %edx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc2,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 0, i16 %mask)
%vec0 = insertelement <8 x i16> undef, i16 %res0, i32 0
-; CHECK: vpcmpltw %ymm1, %ymm0, %k0 {%k1} ##
%res1 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 1, i16 %mask)
%vec1 = insertelement <8 x i16> %vec0, i16 %res1, i32 1
-; CHECK: vpcmplew %ymm1, %ymm0, %k0 {%k1} ##
%res2 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 2, i16 %mask)
%vec2 = insertelement <8 x i16> %vec1, i16 %res2, i32 2
-; CHECK: vpcmpunordw %ymm1, %ymm0, %k0 {%k1} ##
%res3 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 3, i16 %mask)
%vec3 = insertelement <8 x i16> %vec2, i16 %res3, i32 3
-; CHECK: vpcmpneqw %ymm1, %ymm0, %k0 {%k1} ##
%res4 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 4, i16 %mask)
%vec4 = insertelement <8 x i16> %vec3, i16 %res4, i32 4
-; CHECK: vpcmpnltw %ymm1, %ymm0, %k0 {%k1} ##
%res5 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 5, i16 %mask)
%vec5 = insertelement <8 x i16> %vec4, i16 %res5, i32 5
-; CHECK: vpcmpnlew %ymm1, %ymm0, %k0 {%k1} ##
%res6 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 6, i16 %mask)
%vec6 = insertelement <8 x i16> %vec5, i16 %res6, i32 6
-; CHECK: vpcmpordw %ymm1, %ymm0, %k0 {%k1} ##
%res7 = call i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 7, i16 %mask)
%vec7 = insertelement <8 x i16> %vec6, i16 %res7, i32 7
ret <8 x i16> %vec7
@@ -247,58 +391,95 @@ define <8 x i16> @test_mask_cmp_w_256(<1
declare i16 @llvm.x86.avx512.mask.cmp.w.256(<16 x i16>, <16 x i16>, i32, i16) nounwind readnone
define <8 x i16> @test_ucmp_w_256(<16 x i16> %a0, <16 x i16> %a1) {
-; CHECK-LABEL: test_ucmp_w_256
-; CHECK: vpcmpequw %ymm1, %ymm0, %k0 ##
+; CHECK-LABEL: test_ucmp_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpequw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3e,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltuw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3e,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmpleuw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3e,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunorduw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3e,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %esi ## encoding: [0xc5,0xf8,0x93,0xf0]
+; CHECK-NEXT: vpcmpnequw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3e,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltuw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3e,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleuw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3e,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmporduw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x3e,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: vmovd %r8d, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc0]
+; CHECK-NEXT: vpinsrw $1, %r9d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc1,0x01]
+; CHECK-NEXT: vpinsrw $2, %r10d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc2,0x02]
+; CHECK-NEXT: vpinsrw $3, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc6,0x03]
+; CHECK-NEXT: vpinsrw $4, %edi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc7,0x04]
+; CHECK-NEXT: vpinsrw $5, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc0,0x05]
+; CHECK-NEXT: vpinsrw $6, %ecx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc1,0x06]
+; CHECK-NEXT: vpinsrw $7, %edx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc2,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 0, i16 -1)
%vec0 = insertelement <8 x i16> undef, i16 %res0, i32 0
-; CHECK: vpcmpltuw %ymm1, %ymm0, %k0 ##
%res1 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 1, i16 -1)
%vec1 = insertelement <8 x i16> %vec0, i16 %res1, i32 1
-; CHECK: vpcmpleuw %ymm1, %ymm0, %k0 ##
%res2 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 2, i16 -1)
%vec2 = insertelement <8 x i16> %vec1, i16 %res2, i32 2
-; CHECK: vpcmpunorduw %ymm1, %ymm0, %k0 ##
%res3 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 3, i16 -1)
%vec3 = insertelement <8 x i16> %vec2, i16 %res3, i32 3
-; CHECK: vpcmpnequw %ymm1, %ymm0, %k0 ##
%res4 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 4, i16 -1)
%vec4 = insertelement <8 x i16> %vec3, i16 %res4, i32 4
-; CHECK: vpcmpnltuw %ymm1, %ymm0, %k0 ##
%res5 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 5, i16 -1)
%vec5 = insertelement <8 x i16> %vec4, i16 %res5, i32 5
-; CHECK: vpcmpnleuw %ymm1, %ymm0, %k0 ##
%res6 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 6, i16 -1)
%vec6 = insertelement <8 x i16> %vec5, i16 %res6, i32 6
-; CHECK: vpcmporduw %ymm1, %ymm0, %k0 ##
%res7 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 7, i16 -1)
%vec7 = insertelement <8 x i16> %vec6, i16 %res7, i32 7
ret <8 x i16> %vec7
}
define <8 x i16> @test_mask_ucmp_w_256(<16 x i16> %a0, <16 x i16> %a1, i16 %mask) {
-; CHECK-LABEL: test_mask_ucmp_w_256
-; CHECK: vpcmpequw %ymm1, %ymm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_ucmp_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpequw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3e,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltuw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3e,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmpleuw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3e,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunorduw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3e,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %esi ## encoding: [0xc5,0xf8,0x93,0xf0]
+; CHECK-NEXT: vpcmpnequw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3e,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltuw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3e,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleuw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3e,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmporduw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x3e,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: vmovd %r8d, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc0]
+; CHECK-NEXT: vpinsrw $1, %r9d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc1,0x01]
+; CHECK-NEXT: vpinsrw $2, %r10d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc2,0x02]
+; CHECK-NEXT: vpinsrw $3, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc6,0x03]
+; CHECK-NEXT: vpinsrw $4, %edi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc7,0x04]
+; CHECK-NEXT: vpinsrw $5, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc0,0x05]
+; CHECK-NEXT: vpinsrw $6, %ecx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc1,0x06]
+; CHECK-NEXT: vpinsrw $7, %edx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc2,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 0, i16 %mask)
%vec0 = insertelement <8 x i16> undef, i16 %res0, i32 0
-; CHECK: vpcmpltuw %ymm1, %ymm0, %k0 {%k1} ##
%res1 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 1, i16 %mask)
%vec1 = insertelement <8 x i16> %vec0, i16 %res1, i32 1
-; CHECK: vpcmpleuw %ymm1, %ymm0, %k0 {%k1} ##
%res2 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 2, i16 %mask)
%vec2 = insertelement <8 x i16> %vec1, i16 %res2, i32 2
-; CHECK: vpcmpunorduw %ymm1, %ymm0, %k0 {%k1} ##
%res3 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 3, i16 %mask)
%vec3 = insertelement <8 x i16> %vec2, i16 %res3, i32 3
-; CHECK: vpcmpnequw %ymm1, %ymm0, %k0 {%k1} ##
%res4 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 4, i16 %mask)
%vec4 = insertelement <8 x i16> %vec3, i16 %res4, i32 4
-; CHECK: vpcmpnltuw %ymm1, %ymm0, %k0 {%k1} ##
%res5 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 5, i16 %mask)
%vec5 = insertelement <8 x i16> %vec4, i16 %res5, i32 5
-; CHECK: vpcmpnleuw %ymm1, %ymm0, %k0 {%k1} ##
%res6 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 6, i16 %mask)
%vec6 = insertelement <8 x i16> %vec5, i16 %res6, i32 6
-; CHECK: vpcmporduw %ymm1, %ymm0, %k0 {%k1} ##
%res7 = call i16 @llvm.x86.avx512.mask.ucmp.w.256(<16 x i16> %a0, <16 x i16> %a1, i32 7, i16 %mask)
%vec7 = insertelement <8 x i16> %vec6, i16 %res7, i32 7
ret <8 x i16> %vec7
@@ -309,15 +490,22 @@ declare i16 @llvm.x86.avx512.mask.ucmp.w
; 128-bit
define i16 @test_pcmpeq_b_128(<16 x i8> %a, <16 x i8> %b) {
-; CHECK-LABEL: test_pcmpeq_b_128
-; CHECK: vpcmpeqb %xmm1, %xmm0, %k0 ##
+; CHECK-LABEL: test_pcmpeq_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpeqb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x08,0x74,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.mask.pcmpeq.b.128(<16 x i8> %a, <16 x i8> %b, i16 -1)
ret i16 %res
}
define i16 @test_mask_pcmpeq_b_128(<16 x i8> %a, <16 x i8> %b, i16 %mask) {
-; CHECK-LABEL: test_mask_pcmpeq_b_128
-; CHECK: vpcmpeqb %xmm1, %xmm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_pcmpeq_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpeqb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x74,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.mask.pcmpeq.b.128(<16 x i8> %a, <16 x i8> %b, i16 %mask)
ret i16 %res
}
@@ -325,15 +513,22 @@ define i16 @test_mask_pcmpeq_b_128(<16 x
declare i16 @llvm.x86.avx512.mask.pcmpeq.b.128(<16 x i8>, <16 x i8>, i16)
define i8 @test_pcmpeq_w_128(<8 x i16> %a, <8 x i16> %b) {
-; CHECK-LABEL: test_pcmpeq_w_128
-; CHECK: vpcmpeqw %xmm1, %xmm0, %k0 ##
+; CHECK-LABEL: test_pcmpeq_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpeqw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x08,0x75,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.mask.pcmpeq.w.128(<8 x i16> %a, <8 x i16> %b, i8 -1)
ret i8 %res
}
define i8 @test_mask_pcmpeq_w_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_pcmpeq_w_128
-; CHECK: vpcmpeqw %xmm1, %xmm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_pcmpeq_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpeqw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x75,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.mask.pcmpeq.w.128(<8 x i16> %a, <8 x i16> %b, i8 %mask)
ret i8 %res
}
@@ -341,15 +536,22 @@ define i8 @test_mask_pcmpeq_w_128(<8 x i
declare i8 @llvm.x86.avx512.mask.pcmpeq.w.128(<8 x i16>, <8 x i16>, i8)
define i16 @test_pcmpgt_b_128(<16 x i8> %a, <16 x i8> %b) {
-; CHECK-LABEL: test_pcmpgt_b_128
-; CHECK: vpcmpgtb %xmm1, %xmm0, %k0 ##
+; CHECK-LABEL: test_pcmpgt_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpgtb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x08,0x64,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.mask.pcmpgt.b.128(<16 x i8> %a, <16 x i8> %b, i16 -1)
ret i16 %res
}
define i16 @test_mask_pcmpgt_b_128(<16 x i8> %a, <16 x i8> %b, i16 %mask) {
-; CHECK-LABEL: test_mask_pcmpgt_b_128
-; CHECK: vpcmpgtb %xmm1, %xmm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_pcmpgt_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpgtb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x64,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.mask.pcmpgt.b.128(<16 x i8> %a, <16 x i8> %b, i16 %mask)
ret i16 %res
}
@@ -357,15 +559,22 @@ define i16 @test_mask_pcmpgt_b_128(<16 x
declare i16 @llvm.x86.avx512.mask.pcmpgt.b.128(<16 x i8>, <16 x i8>, i16)
define i8 @test_pcmpgt_w_128(<8 x i16> %a, <8 x i16> %b) {
-; CHECK-LABEL: test_pcmpgt_w_128
-; CHECK: vpcmpgtw %xmm1, %xmm0, %k0 ##
+; CHECK-LABEL: test_pcmpgt_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpgtw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x08,0x65,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.mask.pcmpgt.w.128(<8 x i16> %a, <8 x i16> %b, i8 -1)
ret i8 %res
}
define i8 @test_mask_pcmpgt_w_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
-; CHECK-LABEL: test_mask_pcmpgt_w_128
-; CHECK: vpcmpgtw %xmm1, %xmm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_pcmpgt_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpgtw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x65,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.mask.pcmpgt.w.128(<8 x i16> %a, <8 x i16> %b, i8 %mask)
ret i8 %res
}
@@ -373,58 +582,95 @@ define i8 @test_mask_pcmpgt_w_128(<8 x i
declare i8 @llvm.x86.avx512.mask.pcmpgt.w.128(<8 x i16>, <8 x i16>, i8)
define <8 x i16> @test_cmp_b_128(<16 x i8> %a0, <16 x i8> %a1) {
-; CHECK-LABEL: test_cmp_b_128
-; CHECK: vpcmpeqb %xmm1, %xmm0, %k0 ##
+; CHECK-LABEL: test_cmp_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpeqb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3f,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3f,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmpleb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3f,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3f,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %esi ## encoding: [0xc5,0xf8,0x93,0xf0]
+; CHECK-NEXT: vpcmpneqb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3f,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3f,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3f,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmpordb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3f,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: vmovd %r8d, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc0]
+; CHECK-NEXT: vpinsrw $1, %r9d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc1,0x01]
+; CHECK-NEXT: vpinsrw $2, %r10d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc2,0x02]
+; CHECK-NEXT: vpinsrw $3, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc6,0x03]
+; CHECK-NEXT: vpinsrw $4, %edi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc7,0x04]
+; CHECK-NEXT: vpinsrw $5, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc0,0x05]
+; CHECK-NEXT: vpinsrw $6, %ecx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc1,0x06]
+; CHECK-NEXT: vpinsrw $7, %edx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc2,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 0, i16 -1)
%vec0 = insertelement <8 x i16> undef, i16 %res0, i32 0
-; CHECK: vpcmpltb %xmm1, %xmm0, %k0 ##
%res1 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 1, i16 -1)
%vec1 = insertelement <8 x i16> %vec0, i16 %res1, i32 1
-; CHECK: vpcmpleb %xmm1, %xmm0, %k0 ##
%res2 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 2, i16 -1)
%vec2 = insertelement <8 x i16> %vec1, i16 %res2, i32 2
-; CHECK: vpcmpunordb %xmm1, %xmm0, %k0 ##
%res3 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 3, i16 -1)
%vec3 = insertelement <8 x i16> %vec2, i16 %res3, i32 3
-; CHECK: vpcmpneqb %xmm1, %xmm0, %k0 ##
%res4 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 4, i16 -1)
%vec4 = insertelement <8 x i16> %vec3, i16 %res4, i32 4
-; CHECK: vpcmpnltb %xmm1, %xmm0, %k0 ##
%res5 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 5, i16 -1)
%vec5 = insertelement <8 x i16> %vec4, i16 %res5, i32 5
-; CHECK: vpcmpnleb %xmm1, %xmm0, %k0 ##
%res6 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 6, i16 -1)
%vec6 = insertelement <8 x i16> %vec5, i16 %res6, i32 6
-; CHECK: vpcmpordb %xmm1, %xmm0, %k0 ##
%res7 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 7, i16 -1)
%vec7 = insertelement <8 x i16> %vec6, i16 %res7, i32 7
ret <8 x i16> %vec7
}
define <8 x i16> @test_mask_cmp_b_128(<16 x i8> %a0, <16 x i8> %a1, i16 %mask) {
-; CHECK-LABEL: test_mask_cmp_b_128
-; CHECK: vpcmpeqb %xmm1, %xmm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_cmp_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpeqb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3f,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3f,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmpleb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3f,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3f,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %esi ## encoding: [0xc5,0xf8,0x93,0xf0]
+; CHECK-NEXT: vpcmpneqb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3f,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3f,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3f,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmpordb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3f,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: vmovd %r8d, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc0]
+; CHECK-NEXT: vpinsrw $1, %r9d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc1,0x01]
+; CHECK-NEXT: vpinsrw $2, %r10d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc2,0x02]
+; CHECK-NEXT: vpinsrw $3, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc6,0x03]
+; CHECK-NEXT: vpinsrw $4, %edi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc7,0x04]
+; CHECK-NEXT: vpinsrw $5, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc0,0x05]
+; CHECK-NEXT: vpinsrw $6, %ecx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc1,0x06]
+; CHECK-NEXT: vpinsrw $7, %edx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc2,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 0, i16 %mask)
%vec0 = insertelement <8 x i16> undef, i16 %res0, i32 0
-; CHECK: vpcmpltb %xmm1, %xmm0, %k0 {%k1} ##
%res1 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 1, i16 %mask)
%vec1 = insertelement <8 x i16> %vec0, i16 %res1, i32 1
-; CHECK: vpcmpleb %xmm1, %xmm0, %k0 {%k1} ##
%res2 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 2, i16 %mask)
%vec2 = insertelement <8 x i16> %vec1, i16 %res2, i32 2
-; CHECK: vpcmpunordb %xmm1, %xmm0, %k0 {%k1} ##
%res3 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 3, i16 %mask)
%vec3 = insertelement <8 x i16> %vec2, i16 %res3, i32 3
-; CHECK: vpcmpneqb %xmm1, %xmm0, %k0 {%k1} ##
%res4 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 4, i16 %mask)
%vec4 = insertelement <8 x i16> %vec3, i16 %res4, i32 4
-; CHECK: vpcmpnltb %xmm1, %xmm0, %k0 {%k1} ##
%res5 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 5, i16 %mask)
%vec5 = insertelement <8 x i16> %vec4, i16 %res5, i32 5
-; CHECK: vpcmpnleb %xmm1, %xmm0, %k0 {%k1} ##
%res6 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 6, i16 %mask)
%vec6 = insertelement <8 x i16> %vec5, i16 %res6, i32 6
-; CHECK: vpcmpordb %xmm1, %xmm0, %k0 {%k1} ##
%res7 = call i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 7, i16 %mask)
%vec7 = insertelement <8 x i16> %vec6, i16 %res7, i32 7
ret <8 x i16> %vec7
@@ -433,58 +679,95 @@ define <8 x i16> @test_mask_cmp_b_128(<1
declare i16 @llvm.x86.avx512.mask.cmp.b.128(<16 x i8>, <16 x i8>, i32, i16) nounwind readnone
define <8 x i16> @test_ucmp_b_128(<16 x i8> %a0, <16 x i8> %a1) {
-; CHECK-LABEL: test_ucmp_b_128
-; CHECK: vpcmpequb %xmm1, %xmm0, %k0 ##
+; CHECK-LABEL: test_ucmp_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpequb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3e,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltub %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3e,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmpleub %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3e,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordub %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3e,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %esi ## encoding: [0xc5,0xf8,0x93,0xf0]
+; CHECK-NEXT: vpcmpnequb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3e,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltub %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3e,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleub %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3e,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmpordub %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x3e,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: vmovd %r8d, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc0]
+; CHECK-NEXT: vpinsrw $1, %r9d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc1,0x01]
+; CHECK-NEXT: vpinsrw $2, %r10d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc2,0x02]
+; CHECK-NEXT: vpinsrw $3, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc6,0x03]
+; CHECK-NEXT: vpinsrw $4, %edi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc7,0x04]
+; CHECK-NEXT: vpinsrw $5, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc0,0x05]
+; CHECK-NEXT: vpinsrw $6, %ecx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc1,0x06]
+; CHECK-NEXT: vpinsrw $7, %edx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc2,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 0, i16 -1)
%vec0 = insertelement <8 x i16> undef, i16 %res0, i32 0
-; CHECK: vpcmpltub %xmm1, %xmm0, %k0 ##
%res1 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 1, i16 -1)
%vec1 = insertelement <8 x i16> %vec0, i16 %res1, i32 1
-; CHECK: vpcmpleub %xmm1, %xmm0, %k0 ##
%res2 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 2, i16 -1)
%vec2 = insertelement <8 x i16> %vec1, i16 %res2, i32 2
-; CHECK: vpcmpunordub %xmm1, %xmm0, %k0 ##
%res3 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 3, i16 -1)
%vec3 = insertelement <8 x i16> %vec2, i16 %res3, i32 3
-; CHECK: vpcmpnequb %xmm1, %xmm0, %k0 ##
%res4 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 4, i16 -1)
%vec4 = insertelement <8 x i16> %vec3, i16 %res4, i32 4
-; CHECK: vpcmpnltub %xmm1, %xmm0, %k0 ##
%res5 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 5, i16 -1)
%vec5 = insertelement <8 x i16> %vec4, i16 %res5, i32 5
-; CHECK: vpcmpnleub %xmm1, %xmm0, %k0 ##
%res6 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 6, i16 -1)
%vec6 = insertelement <8 x i16> %vec5, i16 %res6, i32 6
-; CHECK: vpcmpordub %xmm1, %xmm0, %k0 ##
%res7 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 7, i16 -1)
%vec7 = insertelement <8 x i16> %vec6, i16 %res7, i32 7
ret <8 x i16> %vec7
}
define <8 x i16> @test_mask_ucmp_b_128(<16 x i8> %a0, <16 x i8> %a1, i16 %mask) {
-; CHECK-LABEL: test_mask_ucmp_b_128
-; CHECK: vpcmpequb %xmm1, %xmm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_ucmp_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpequb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3e,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltub %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3e,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmpleub %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3e,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordub %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3e,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %esi ## encoding: [0xc5,0xf8,0x93,0xf0]
+; CHECK-NEXT: vpcmpnequb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3e,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltub %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3e,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleub %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3e,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmpordub %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x3e,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: vmovd %r8d, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0x6e,0xc0]
+; CHECK-NEXT: vpinsrw $1, %r9d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc1,0x01]
+; CHECK-NEXT: vpinsrw $2, %r10d, %xmm0, %xmm0 ## encoding: [0x62,0xd1,0x7d,0x08,0xc4,0xc2,0x02]
+; CHECK-NEXT: vpinsrw $3, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc6,0x03]
+; CHECK-NEXT: vpinsrw $4, %edi, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc7,0x04]
+; CHECK-NEXT: vpinsrw $5, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc0,0x05]
+; CHECK-NEXT: vpinsrw $6, %ecx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc1,0x06]
+; CHECK-NEXT: vpinsrw $7, %edx, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xc4,0xc2,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 0, i16 %mask)
%vec0 = insertelement <8 x i16> undef, i16 %res0, i32 0
-; CHECK: vpcmpltub %xmm1, %xmm0, %k0 {%k1} ##
%res1 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 1, i16 %mask)
%vec1 = insertelement <8 x i16> %vec0, i16 %res1, i32 1
-; CHECK: vpcmpleub %xmm1, %xmm0, %k0 {%k1} ##
%res2 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 2, i16 %mask)
%vec2 = insertelement <8 x i16> %vec1, i16 %res2, i32 2
-; CHECK: vpcmpunordub %xmm1, %xmm0, %k0 {%k1} ##
%res3 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 3, i16 %mask)
%vec3 = insertelement <8 x i16> %vec2, i16 %res3, i32 3
-; CHECK: vpcmpnequb %xmm1, %xmm0, %k0 {%k1} ##
%res4 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 4, i16 %mask)
%vec4 = insertelement <8 x i16> %vec3, i16 %res4, i32 4
-; CHECK: vpcmpnltub %xmm1, %xmm0, %k0 {%k1} ##
%res5 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 5, i16 %mask)
%vec5 = insertelement <8 x i16> %vec4, i16 %res5, i32 5
-; CHECK: vpcmpnleub %xmm1, %xmm0, %k0 {%k1} ##
%res6 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 6, i16 %mask)
%vec6 = insertelement <8 x i16> %vec5, i16 %res6, i32 6
-; CHECK: vpcmpordub %xmm1, %xmm0, %k0 {%k1} ##
%res7 = call i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8> %a0, <16 x i8> %a1, i32 7, i16 %mask)
%vec7 = insertelement <8 x i16> %vec6, i16 %res7, i32 7
ret <8 x i16> %vec7
@@ -493,58 +776,111 @@ define <8 x i16> @test_mask_ucmp_b_128(<
declare i16 @llvm.x86.avx512.mask.ucmp.b.128(<16 x i8>, <16 x i8>, i32, i16) nounwind readnone
define <8 x i8> @test_cmp_w_128(<8 x i16> %a0, <8 x i16> %a1) {
-; CHECK-LABEL: test_cmp_w_128
-; CHECK: vpcmpeqw %xmm1, %xmm0, %k0 ##
+; CHECK-LABEL: test_cmp_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpeqw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3f,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3f,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmplew %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3f,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3f,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %r11d ## encoding: [0xc5,0x78,0x93,0xd8]
+; CHECK-NEXT: vpcmpneqw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3f,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3f,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnlew %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3f,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmpordw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3f,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: movzbl %r8b, %esi ## encoding: [0x41,0x0f,0xb6,0xf0]
+; CHECK-NEXT: vpinsrb $0, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x00]
+; CHECK-NEXT: movzbl %r9b, %esi ## encoding: [0x41,0x0f,0xb6,0xf1]
+; CHECK-NEXT: vpinsrb $2, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x02]
+; CHECK-NEXT: movzbl %r10b, %esi ## encoding: [0x41,0x0f,0xb6,0xf2]
+; CHECK-NEXT: vpinsrb $4, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x04]
+; CHECK-NEXT: movzbl %r11b, %esi ## encoding: [0x41,0x0f,0xb6,0xf3]
+; CHECK-NEXT: vpinsrb $6, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x06]
+; CHECK-NEXT: movzbl %dil, %esi ## encoding: [0x40,0x0f,0xb6,0xf7]
+; CHECK-NEXT: vpinsrb $8, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x08]
+; CHECK-NEXT: movzbl %al, %eax ## encoding: [0x0f,0xb6,0xc0]
+; CHECK-NEXT: vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0a]
+; CHECK-NEXT: movzbl %cl, %eax ## encoding: [0x0f,0xb6,0xc1]
+; CHECK-NEXT: vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0c]
+; CHECK-NEXT: movzbl %dl, %eax ## encoding: [0x0f,0xb6,0xc2]
+; CHECK-NEXT: vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0e]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 0, i8 -1)
%vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
-; CHECK: vpcmpltw %xmm1, %xmm0, %k0 ##
%res1 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 1, i8 -1)
%vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
-; CHECK: vpcmplew %xmm1, %xmm0, %k0 ##
%res2 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 2, i8 -1)
%vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
-; CHECK: vpcmpunordw %xmm1, %xmm0, %k0 ##
%res3 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 3, i8 -1)
%vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
-; CHECK: vpcmpneqw %xmm1, %xmm0, %k0 ##
%res4 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 4, i8 -1)
%vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
-; CHECK: vpcmpnltw %xmm1, %xmm0, %k0 ##
%res5 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 5, i8 -1)
%vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
-; CHECK: vpcmpnlew %xmm1, %xmm0, %k0 ##
%res6 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 6, i8 -1)
%vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
-; CHECK: vpcmpordw %xmm1, %xmm0, %k0 ##
%res7 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 7, i8 -1)
%vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
ret <8 x i8> %vec7
}
define <8 x i8> @test_mask_cmp_w_128(<8 x i16> %a0, <8 x i16> %a1, i8 %mask) {
-; CHECK-LABEL: test_mask_cmp_w_128
-; CHECK: vpcmpeqw %xmm1, %xmm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_cmp_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpeqw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3f,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3f,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmplew %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3f,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunordw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3f,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %r11d ## encoding: [0xc5,0x78,0x93,0xd8]
+; CHECK-NEXT: vpcmpneqw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3f,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3f,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnlew %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3f,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmpordw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3f,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: movzbl %r8b, %esi ## encoding: [0x41,0x0f,0xb6,0xf0]
+; CHECK-NEXT: vpinsrb $0, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x00]
+; CHECK-NEXT: movzbl %r9b, %esi ## encoding: [0x41,0x0f,0xb6,0xf1]
+; CHECK-NEXT: vpinsrb $2, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x02]
+; CHECK-NEXT: movzbl %r10b, %esi ## encoding: [0x41,0x0f,0xb6,0xf2]
+; CHECK-NEXT: vpinsrb $4, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x04]
+; CHECK-NEXT: movzbl %r11b, %esi ## encoding: [0x41,0x0f,0xb6,0xf3]
+; CHECK-NEXT: vpinsrb $6, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x06]
+; CHECK-NEXT: movzbl %dil, %esi ## encoding: [0x40,0x0f,0xb6,0xf7]
+; CHECK-NEXT: vpinsrb $8, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x08]
+; CHECK-NEXT: movzbl %al, %eax ## encoding: [0x0f,0xb6,0xc0]
+; CHECK-NEXT: vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0a]
+; CHECK-NEXT: movzbl %cl, %eax ## encoding: [0x0f,0xb6,0xc1]
+; CHECK-NEXT: vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0c]
+; CHECK-NEXT: movzbl %dl, %eax ## encoding: [0x0f,0xb6,0xc2]
+; CHECK-NEXT: vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0e]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 0, i8 %mask)
%vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
-; CHECK: vpcmpltw %xmm1, %xmm0, %k0 {%k1} ##
%res1 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 1, i8 %mask)
%vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
-; CHECK: vpcmplew %xmm1, %xmm0, %k0 {%k1} ##
%res2 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 2, i8 %mask)
%vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
-; CHECK: vpcmpunordw %xmm1, %xmm0, %k0 {%k1} ##
%res3 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 3, i8 %mask)
%vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
-; CHECK: vpcmpneqw %xmm1, %xmm0, %k0 {%k1} ##
%res4 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 4, i8 %mask)
%vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
-; CHECK: vpcmpnltw %xmm1, %xmm0, %k0 {%k1} ##
%res5 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 5, i8 %mask)
%vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
-; CHECK: vpcmpnlew %xmm1, %xmm0, %k0 {%k1} ##
%res6 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 6, i8 %mask)
%vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
-; CHECK: vpcmpordw %xmm1, %xmm0, %k0 {%k1} ##
%res7 = call i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 7, i8 %mask)
%vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
ret <8 x i8> %vec7
@@ -553,58 +889,111 @@ define <8 x i8> @test_mask_cmp_w_128(<8
declare i8 @llvm.x86.avx512.mask.cmp.w.128(<8 x i16>, <8 x i16>, i32, i8) nounwind readnone
define <8 x i8> @test_ucmp_w_128(<8 x i16> %a0, <8 x i16> %a1) {
-; CHECK-LABEL: test_ucmp_w_128
-; CHECK: vpcmpequw %xmm1, %xmm0, %k0 ##
+; CHECK-LABEL: test_ucmp_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpcmpequw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3e,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltuw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3e,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmpleuw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3e,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunorduw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3e,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %r11d ## encoding: [0xc5,0x78,0x93,0xd8]
+; CHECK-NEXT: vpcmpnequw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3e,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltuw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3e,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleuw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3e,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmporduw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x3e,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: movzbl %r8b, %esi ## encoding: [0x41,0x0f,0xb6,0xf0]
+; CHECK-NEXT: vpinsrb $0, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x00]
+; CHECK-NEXT: movzbl %r9b, %esi ## encoding: [0x41,0x0f,0xb6,0xf1]
+; CHECK-NEXT: vpinsrb $2, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x02]
+; CHECK-NEXT: movzbl %r10b, %esi ## encoding: [0x41,0x0f,0xb6,0xf2]
+; CHECK-NEXT: vpinsrb $4, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x04]
+; CHECK-NEXT: movzbl %r11b, %esi ## encoding: [0x41,0x0f,0xb6,0xf3]
+; CHECK-NEXT: vpinsrb $6, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x06]
+; CHECK-NEXT: movzbl %dil, %esi ## encoding: [0x40,0x0f,0xb6,0xf7]
+; CHECK-NEXT: vpinsrb $8, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x08]
+; CHECK-NEXT: movzbl %al, %eax ## encoding: [0x0f,0xb6,0xc0]
+; CHECK-NEXT: vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0a]
+; CHECK-NEXT: movzbl %cl, %eax ## encoding: [0x0f,0xb6,0xc1]
+; CHECK-NEXT: vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0c]
+; CHECK-NEXT: movzbl %dl, %eax ## encoding: [0x0f,0xb6,0xc2]
+; CHECK-NEXT: vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0e]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 0, i8 -1)
%vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
-; CHECK: vpcmpltuw %xmm1, %xmm0, %k0 ##
%res1 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 1, i8 -1)
%vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
-; CHECK: vpcmpleuw %xmm1, %xmm0, %k0 ##
%res2 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 2, i8 -1)
%vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
-; CHECK: vpcmpunorduw %xmm1, %xmm0, %k0 ##
%res3 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 3, i8 -1)
%vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
-; CHECK: vpcmpnequw %xmm1, %xmm0, %k0 ##
%res4 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 4, i8 -1)
%vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
-; CHECK: vpcmpnltuw %xmm1, %xmm0, %k0 ##
%res5 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 5, i8 -1)
%vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
-; CHECK: vpcmpnleuw %xmm1, %xmm0, %k0 ##
%res6 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 6, i8 -1)
%vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
-; CHECK: vpcmporduw %xmm1, %xmm0, %k0 ##
%res7 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 7, i8 -1)
%vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
ret <8 x i8> %vec7
}
define <8 x i8> @test_mask_ucmp_w_128(<8 x i16> %a0, <8 x i16> %a1, i8 %mask) {
-; CHECK-LABEL: test_mask_ucmp_w_128
-; CHECK: vpcmpequw %xmm1, %xmm0, %k0 {%k1} ##
+; CHECK-LABEL: test_mask_ucmp_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpcmpequw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3e,0xc1,0x00]
+; CHECK-NEXT: kmovw %k0, %r8d ## encoding: [0xc5,0x78,0x93,0xc0]
+; CHECK-NEXT: vpcmpltuw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3e,0xc1,0x01]
+; CHECK-NEXT: kmovw %k0, %r9d ## encoding: [0xc5,0x78,0x93,0xc8]
+; CHECK-NEXT: vpcmpleuw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3e,0xc1,0x02]
+; CHECK-NEXT: kmovw %k0, %r10d ## encoding: [0xc5,0x78,0x93,0xd0]
+; CHECK-NEXT: vpcmpunorduw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3e,0xc1,0x03]
+; CHECK-NEXT: kmovw %k0, %r11d ## encoding: [0xc5,0x78,0x93,0xd8]
+; CHECK-NEXT: vpcmpnequw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3e,0xc1,0x04]
+; CHECK-NEXT: kmovw %k0, %edi ## encoding: [0xc5,0xf8,0x93,0xf8]
+; CHECK-NEXT: vpcmpnltuw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3e,0xc1,0x05]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: vpcmpnleuw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3e,0xc1,0x06]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vpcmporduw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x3e,0xc1,0x07]
+; CHECK-NEXT: kmovw %k0, %edx ## encoding: [0xc5,0xf8,0x93,0xd0]
+; CHECK-NEXT: movzbl %r8b, %esi ## encoding: [0x41,0x0f,0xb6,0xf0]
+; CHECK-NEXT: vpinsrb $0, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x00]
+; CHECK-NEXT: movzbl %r9b, %esi ## encoding: [0x41,0x0f,0xb6,0xf1]
+; CHECK-NEXT: vpinsrb $2, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x02]
+; CHECK-NEXT: movzbl %r10b, %esi ## encoding: [0x41,0x0f,0xb6,0xf2]
+; CHECK-NEXT: vpinsrb $4, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x04]
+; CHECK-NEXT: movzbl %r11b, %esi ## encoding: [0x41,0x0f,0xb6,0xf3]
+; CHECK-NEXT: vpinsrb $6, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x06]
+; CHECK-NEXT: movzbl %dil, %esi ## encoding: [0x40,0x0f,0xb6,0xf7]
+; CHECK-NEXT: vpinsrb $8, %esi, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc6,0x08]
+; CHECK-NEXT: movzbl %al, %eax ## encoding: [0x0f,0xb6,0xc0]
+; CHECK-NEXT: vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0a]
+; CHECK-NEXT: movzbl %cl, %eax ## encoding: [0x0f,0xb6,0xc1]
+; CHECK-NEXT: vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0c]
+; CHECK-NEXT: movzbl %dl, %eax ## encoding: [0x0f,0xb6,0xc2]
+; CHECK-NEXT: vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x20,0xc0,0x0e]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 0, i8 %mask)
%vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
-; CHECK: vpcmpltuw %xmm1, %xmm0, %k0 {%k1} ##
%res1 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 1, i8 %mask)
%vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
-; CHECK: vpcmpleuw %xmm1, %xmm0, %k0 {%k1} ##
%res2 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 2, i8 %mask)
%vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
-; CHECK: vpcmpunorduw %xmm1, %xmm0, %k0 {%k1} ##
%res3 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 3, i8 %mask)
%vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
-; CHECK: vpcmpnequw %xmm1, %xmm0, %k0 {%k1} ##
%res4 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 4, i8 %mask)
%vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
-; CHECK: vpcmpnltuw %xmm1, %xmm0, %k0 {%k1} ##
%res5 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 5, i8 %mask)
%vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
-; CHECK: vpcmpnleuw %xmm1, %xmm0, %k0 {%k1} ##
%res6 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 6, i8 %mask)
%vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
-; CHECK: vpcmporduw %xmm1, %xmm0, %k0 {%k1} ##
%res7 = call i8 @llvm.x86.avx512.mask.ucmp.w.128(<8 x i16> %a0, <8 x i16> %a1, i32 7, i8 %mask)
%vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
ret <8 x i8> %vec7
@@ -615,8 +1004,11 @@ declare i8 @llvm.x86.avx512.mask.ucmp.w.
declare <8 x float> @llvm.x86.avx512.mask.vfmadd.ps.256(<8 x float>, <8 x float>, <8 x float>, i8) nounwind readnone
define <8 x float> @test_mask_vfmadd256_ps(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd256_ps
- ; CHECK: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa8,0xc2]
+; CHECK-LABEL: test_mask_vfmadd256_ps:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfmadd.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) nounwind
ret <8 x float> %res
}
@@ -624,8 +1016,11 @@ define <8 x float> @test_mask_vfmadd256_
declare <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float>, <4 x float>, <4 x float>, i8) nounwind readnone
define <4 x float> @test_mask_vfmadd128_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps
- ; CHECK: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0xc2]
+; CHECK-LABEL: test_mask_vfmadd128_ps:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
}
@@ -634,7 +1029,10 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double> @test_mask_fmadd256_pd(<4 x double> %a, <4 x double> %b, <4 x double> %c, i8 %mask) {
; CHECK-LABEL: test_mask_fmadd256_pd:
-; CHECK: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0xc2]
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %a, <4 x double> %b, <4 x double> %c, i8 %mask)
ret <4 x double> %res
}
@@ -643,7 +1041,10 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double> @test_mask_fmadd128_pd(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
; CHECK-LABEL: test_mask_fmadd128_pd:
-; CHECK: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0xc2]
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask)
ret <2 x double> %res
}
@@ -651,12 +1052,12 @@ define <2 x double> @test_mask_fmadd128_
define <2 x double>@test_int_x86_avx512_mask_vfmadd_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0xda]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -668,12 +1069,12 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_mask3_vfmadd_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmadd_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmadd231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb8,0xd9]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -685,12 +1086,12 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_maskz_vfmadd_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm3 {%k1} {z}
-; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa8,0xda]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.maskz.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.maskz.vfmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -700,12 +1101,12 @@ define <2 x double>@test_int_x86_avx512_
define <4 x double>@test_int_x86_avx512_mask_vfmadd_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm3 {%k1}
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0xda]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -717,12 +1118,12 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask3_vfmadd_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmadd_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmadd231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xb8,0xd9]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -734,12 +1135,12 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_maskz_vfmadd_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm3 {%k1} {z}
-; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xa9,0xa8,0xda]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.maskz.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.maskz.vfmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -749,12 +1150,12 @@ define <4 x double>@test_int_x86_avx512_
define <4 x float>@test_int_x86_avx512_mask_vfmadd_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0xda]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -766,12 +1167,12 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask3_vfmadd_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmadd_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmadd231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb8,0xd9]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -783,12 +1184,12 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_maskz_vfmadd_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm3 {%k1} {z}
-; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa8,0xda]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.maskz.vfmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -798,12 +1199,12 @@ define <4 x float>@test_int_x86_avx512_m
define <8 x float>@test_int_x86_avx512_mask_vfmadd_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfmadd_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm3 {%k1}
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa8,0xda]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa8,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -815,12 +1216,12 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float>@test_int_x86_avx512_mask3_vfmadd_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmadd_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmadd231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xb8,0xd9]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa8,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -832,12 +1233,12 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float>@test_int_x86_avx512_maskz_vfmadd_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmadd_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm3 {%k1} {z}
-; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xa9,0xa8,0xda]
+; CHECK-NEXT: vfmadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa8,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.maskz.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.maskz.vfmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -850,12 +1251,12 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_mask3_vfmsub_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsub_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfmsub213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmsub231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xba,0xd9]
+; CHECK-NEXT: vfmsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xaa,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -868,12 +1269,12 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask3_vfmsub_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsub_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfmsub213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmsub231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xba,0xd9]
+; CHECK-NEXT: vfmsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xaa,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -885,12 +1286,12 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask3_vfmsub_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsub_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfmsub213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmsub231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xba,0xd9]
+; CHECK-NEXT: vfmsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xaa,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -902,12 +1303,12 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float>@test_int_x86_avx512_mask3_vfmsub_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsub_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfmsub213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmsub231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xba,0xd9]
+; CHECK-NEXT: vfmsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xaa,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -917,8 +1318,11 @@ define <8 x float>@test_int_x86_avx512_m
declare <8 x float> @llvm.x86.avx512.mask.vfnmadd.ps.256(<8 x float>, <8 x float>, <8 x float>, i8) nounwind readnone
define <8 x float> @test_mask_vfnmadd256_ps(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfnmadd256_ps
- ; CHECK: vfnmadd213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xac,0xc2]
+; CHECK-LABEL: test_mask_vfnmadd256_ps:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xac,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfnmadd.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) nounwind
ret <8 x float> %res
}
@@ -926,8 +1330,11 @@ define <8 x float> @test_mask_vfnmadd256
declare <4 x float> @llvm.x86.avx512.mask.vfnmadd.ps.128(<4 x float>, <4 x float>, <4 x float>, i8) nounwind readnone
define <4 x float> @test_mask_vfnmadd128_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfnmadd128_ps
- ; CHECK: vfnmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xac,0xc2]
+; CHECK-LABEL: test_mask_vfnmadd128_ps:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xac,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfnmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
}
@@ -935,8 +1342,11 @@ define <4 x float> @test_mask_vfnmadd128
declare <4 x double> @llvm.x86.avx512.mask.vfnmadd.pd.256(<4 x double>, <4 x double>, <4 x double>, i8) nounwind readnone
define <4 x double> @test_mask_vfnmadd256_pd(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfnmadd256_pd
- ; CHECK: vfnmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xac,0xc2]
+; CHECK-LABEL: test_mask_vfnmadd256_pd:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xac,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfnmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
ret <4 x double> %res
}
@@ -944,8 +1354,11 @@ define <4 x double> @test_mask_vfnmadd25
declare <2 x double> @llvm.x86.avx512.mask.vfnmadd.pd.128(<2 x double>, <2 x double>, <2 x double>, i8) nounwind readnone
define <2 x double> @test_mask_vfnmadd128_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfnmadd128_pd
- ; CHECK: vfnmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xac,0xc2]
+; CHECK-LABEL: test_mask_vfnmadd128_pd:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xac,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfnmadd.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
ret <2 x double> %res
}
@@ -953,8 +1366,11 @@ define <2 x double> @test_mask_vfnmadd12
declare <8 x float> @llvm.x86.avx512.mask.vfnmsub.ps.256(<8 x float>, <8 x float>, <8 x float>, i8) nounwind readnone
define <8 x float> @test_mask_vfnmsub256_ps(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfnmsub256_ps
- ; CHECK: vfnmsub213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xae,0xc2]
+; CHECK-LABEL: test_mask_vfnmsub256_ps:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xae,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfnmsub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2, i8 %mask) nounwind
ret <8 x float> %res
}
@@ -962,8 +1378,11 @@ define <8 x float> @test_mask_vfnmsub256
declare <4 x float> @llvm.x86.avx512.mask.vfnmsub.ps.128(<4 x float>, <4 x float>, <4 x float>, i8) nounwind readnone
define <4 x float> @test_mask_vfnmsub128_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfnmsub128_ps
- ; CHECK: vfnmsub213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xae,0xc2]
+; CHECK-LABEL: test_mask_vfnmsub128_ps:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xae,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfnmsub.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
}
@@ -971,8 +1390,11 @@ define <4 x float> @test_mask_vfnmsub128
declare <4 x double> @llvm.x86.avx512.mask.vfnmsub.pd.256(<4 x double>, <4 x double>, <4 x double>, i8) nounwind readnone
define <4 x double> @test_mask_vfnmsub256_pd(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfnmsub256_pd
- ; CHECK: vfnmsub213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xae,0xc2]
+; CHECK-LABEL: test_mask_vfnmsub256_pd:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xae,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfnmsub.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
ret <4 x double> %res
}
@@ -980,8 +1402,11 @@ define <4 x double> @test_mask_vfnmsub25
declare <2 x double> @llvm.x86.avx512.mask.vfnmsub.pd.128(<2 x double>, <2 x double>, <2 x double>, i8) nounwind readnone
define <2 x double> @test_mask_vfnmsub128_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfnmsub128_pd
- ; CHECK: vfnmsub213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xae,0xc2]
+; CHECK-LABEL: test_mask_vfnmsub128_pd:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xae,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfnmsub.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
ret <2 x double> %res
}
@@ -990,12 +1415,12 @@ define <2 x double> @test_mask_vfnmsub12
define <2 x double>@test_int_x86_avx512_mask_vfnmsub_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmsub_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xae,0xda]
+; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xae,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfnmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vfnmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1007,12 +1432,12 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfnmsub_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfnmsub231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xbe,0xd9]
+; CHECK-NEXT: vfnmsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xae,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1022,12 +1447,12 @@ define <2 x double>@test_int_x86_avx512_
define <4 x double>@test_int_x86_avx512_mask_vfnmsub_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmsub_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm3 {%k1}
-; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xae,0xda]
+; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xae,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfnmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.vfnmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1039,12 +1464,12 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask3_vfnmsub_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfnmsub_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfnmsub231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xbe,0xd9]
+; CHECK-NEXT: vfnmsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xae,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfnmsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1054,12 +1479,12 @@ define <4 x double>@test_int_x86_avx512_
define <4 x float>@test_int_x86_avx512_mask_vfnmsub_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmsub_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xae,0xda]
+; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xae,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfnmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.vfnmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1071,12 +1496,12 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfnmsub_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfnmsub231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xbe,0xd9]
+; CHECK-NEXT: vfnmsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xae,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1086,12 +1511,12 @@ define <4 x float>@test_int_x86_avx512_m
define <8 x float>@test_int_x86_avx512_mask_vfnmsub_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmsub_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm3 {%k1}
-; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xae,0xda]
+; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xae,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfnmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.vfnmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -1103,12 +1528,12 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float>@test_int_x86_avx512_mask3_vfnmsub_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfnmsub_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfnmsub231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xbe,0xd9]
+; CHECK-NEXT: vfnmsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xae,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfnmsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -1118,12 +1543,12 @@ define <8 x float>@test_int_x86_avx512_m
define <2 x double>@test_int_x86_avx512_mask_vfnmadd_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmadd_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xac,0xda]
+; CHECK-NEXT: vfnmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xac,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfnmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vfnmadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1133,12 +1558,12 @@ define <2 x double>@test_int_x86_avx512_
define <4 x double>@test_int_x86_avx512_mask_vfnmadd_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmadd_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm3 {%k1}
-; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xac,0xda]
+; CHECK-NEXT: vfnmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xac,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfnmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.vfnmadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1148,12 +1573,12 @@ define <4 x double>@test_int_x86_avx512_
define <4 x float>@test_int_x86_avx512_mask_vfnmadd_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmadd_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xac,0xda]
+; CHECK-NEXT: vfnmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xac,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfnmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.vfnmadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1163,12 +1588,12 @@ define <4 x float>@test_int_x86_avx512_m
define <8 x float>@test_int_x86_avx512_mask_vfnmadd_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfnmadd_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm3 {%k1}
-; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xac,0xda]
+; CHECK-NEXT: vfnmadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xac,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfnmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.vfnmadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -1179,7 +1604,10 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float> @test_mask_fmaddsub256_ps(<8 x float> %a, <8 x float> %b, <8 x float> %c, i8 %mask) {
; CHECK-LABEL: test_mask_fmaddsub256_ps:
-; CHECK: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa6,0xc2]
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa6,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %c, i8 %mask)
ret <8 x float> %res
}
@@ -1188,7 +1616,10 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float> @test_mask_fmaddsub128_ps(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
; CHECK-LABEL: test_mask_fmaddsub128_ps:
-; CHECK: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa6,0xc2]
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa6,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask)
ret <4 x float> %res
}
@@ -1196,8 +1627,11 @@ define <4 x float> @test_mask_fmaddsub12
declare <4 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.256(<4 x double>, <4 x double>, <4 x double>, i8) nounwind readnone
define <4 x double> @test_mask_vfmaddsub256_pd(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmaddsub256_pd
- ; CHECK: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa6,0xc2]
+; CHECK-LABEL: test_mask_vfmaddsub256_pd:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa6,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
ret <4 x double> %res
}
@@ -1205,8 +1639,11 @@ define <4 x double> @test_mask_vfmaddsub
declare <2 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.128(<2 x double>, <2 x double>, <2 x double>, i8) nounwind readnone
define <2 x double> @test_mask_vfmaddsub128_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmaddsub128_pd
- ; CHECK: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa6,0xc2]
+; CHECK-LABEL: test_mask_vfmaddsub128_pd:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa6,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
ret <2 x double> %res
}
@@ -1214,12 +1651,12 @@ define <2 x double> @test_mask_vfmaddsub
define <2 x double>@test_int_x86_avx512_mask_vfmaddsub_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfmaddsub_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa6,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1231,12 +1668,12 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmaddsub_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmaddsub231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb6,0xd9]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa6,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1248,12 +1685,12 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm3 {%k1} {z}
-; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0x89,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa6,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1263,12 +1700,12 @@ define <2 x double>@test_int_x86_avx512_
define <4 x double>@test_int_x86_avx512_mask_vfmaddsub_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfmaddsub_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm3 {%k1}
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa6,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1280,12 +1717,12 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask3_vfmaddsub_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmaddsub_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmaddsub231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xb6,0xd9]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa6,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1297,12 +1734,12 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_maskz_vfmaddsub_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm3 {%k1} {z}
-; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xa9,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa6,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.maskz.vfmaddsub.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1312,12 +1749,12 @@ define <4 x double>@test_int_x86_avx512_
define <4 x float>@test_int_x86_avx512_mask_vfmaddsub_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfmaddsub_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa6,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1329,12 +1766,12 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmaddsub_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmaddsub231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb6,0xd9]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa6,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1346,12 +1783,12 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm3 {%k1} {z}
-; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0x89,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa6,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1361,12 +1798,12 @@ define <4 x float>@test_int_x86_avx512_m
define <8 x float>@test_int_x86_avx512_mask_vfmaddsub_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vfmaddsub_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm3 {%k1}
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa6,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -1378,12 +1815,12 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float>@test_int_x86_avx512_mask3_vfmaddsub_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmaddsub_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmaddsub231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xb6,0xd9]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa6,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -1395,12 +1832,12 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float>@test_int_x86_avx512_maskz_vfmaddsub_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vfmaddsub_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm0, %zmm3
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm3 {%k1} {z}
-; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm0, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd8]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xa9,0xa6,0xda]
+; CHECK-NEXT: vfmaddsub213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa6,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.maskz.vfmaddsub.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -1412,12 +1849,12 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsubadd_pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfmsubadd213pd %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmsubadd231pd %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0xb7,0xd9]
+; CHECK-NEXT: vfmsubadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa7,0xc2]
+; CHECK-NEXT: vaddpd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
%res2=fadd <2 x double> %res, %res1
@@ -1429,12 +1866,12 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask3_vfmsubadd_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsubadd_pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfmsubadd213pd %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmsubadd231pd %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0xb7,0xd9]
+; CHECK-NEXT: vfmsubadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa7,0xc2]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask3.vfmsubadd.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
%res2=fadd <4 x double> %res, %res1
@@ -1446,12 +1883,12 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsubadd_ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm3 {%k1}
-; CHECK-NEXT: vfmsubadd213ps %xmm2, %xmm1, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmsubadd231ps %xmm1, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0xb7,0xd9]
+; CHECK-NEXT: vfmsubadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa7,0xc2]
+; CHECK-NEXT: vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
%res2=fadd <4 x float> %res, %res1
@@ -1463,12 +1900,12 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float>@test_int_x86_avx512_mask3_vfmsubadd_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask3_vfmsubadd_ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovaps %zmm2, %zmm3
-; CHECK-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm3 {%k1}
-; CHECK-NEXT: vfmsubadd213ps %ymm2, %ymm1, %ymm0
-; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm2, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xda]
+; CHECK-NEXT: vfmsubadd231ps %ymm1, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0xb7,0xd9]
+; CHECK-NEXT: vfmsubadd213ps %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0xa7,0xc2]
+; CHECK-NEXT: vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask3.vfmsubadd.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
%res2=fadd <8 x float> %res, %res1
@@ -1477,54 +1914,72 @@ define <8 x float>@test_int_x86_avx512_m
define <4 x float> @test_mask_vfmadd128_ps_r(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_r
- ; CHECK: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0xc2]
+; CHECK-LABEL: test_mask_vfmadd128_ps_r:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
}
define <4 x float> @test_mask_vfmadd128_ps_rz(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rz
- ; CHECK: vfmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0xc2]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rz:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213ps %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 -1) nounwind
ret <4 x float> %res
}
define <4 x float> @test_mask_vfmadd128_ps_rmk(<4 x float> %a0, <4 x float> %a1, <4 x float>* %ptr_a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rmk
- ; CHECK: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rmk:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%a2 = load <4 x float>, <4 x float>* %ptr_a2
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
}
define <4 x float> @test_mask_vfmadd128_ps_rmka(<4 x float> %a0, <4 x float> %a1, <4 x float>* %ptr_a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rmka
- ; CHECK: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rmka:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x09,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%a2 = load <4 x float>, <4 x float>* %ptr_a2, align 8
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 %mask) nounwind
ret <4 x float> %res
}
define <4 x float> @test_mask_vfmadd128_ps_rmkz(<4 x float> %a0, <4 x float> %a1, <4 x float>* %ptr_a2) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rmkz
- ; CHECK: vfmadd213ps (%rdi), %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rmkz:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%a2 = load <4 x float>, <4 x float>* %ptr_a2
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 -1) nounwind
ret <4 x float> %res
}
define <4 x float> @test_mask_vfmadd128_ps_rmkza(<4 x float> %a0, <4 x float> %a1, <4 x float>* %ptr_a2) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rmkza
- ; CHECK: vfmadd213ps (%rdi), %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rmkza:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213ps (%rdi), %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x08,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%a2 = load <4 x float>, <4 x float>* %ptr_a2, align 4
%res = call <4 x float> @llvm.x86.avx512.mask.vfmadd.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2, i8 -1) nounwind
ret <4 x float> %res
}
define <4 x float> @test_mask_vfmadd128_ps_rmb(<4 x float> %a0, <4 x float> %a1, float* %ptr_a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rmb
- ; CHECK: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rmb:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_a2
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%vecinit2.i = insertelement <4 x float> %vecinit.i, float %q, i32 1
@@ -1535,8 +1990,11 @@ define <4 x float> @test_mask_vfmadd128_
}
define <4 x float> @test_mask_vfmadd128_ps_rmba(<4 x float> %a0, <4 x float> %a1, float* %ptr_a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rmba
- ; CHECK: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rmba:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x75,0x19,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_a2, align 4
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%vecinit2.i = insertelement <4 x float> %vecinit.i, float %q, i32 1
@@ -1547,8 +2005,10 @@ define <4 x float> @test_mask_vfmadd128_
}
define <4 x float> @test_mask_vfmadd128_ps_rmbz(<4 x float> %a0, <4 x float> %a1, float* %ptr_a2) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rmbz
- ; CHECK: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rmbz:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_a2
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%vecinit2.i = insertelement <4 x float> %vecinit.i, float %q, i32 1
@@ -1559,8 +2019,10 @@ define <4 x float> @test_mask_vfmadd128_
}
define <4 x float> @test_mask_vfmadd128_ps_rmbza(<4 x float> %a0, <4 x float> %a1, float* %ptr_a2) {
- ; CHECK-LABEL: test_mask_vfmadd128_ps_rmbza
- ; CHECK: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_ps_rmbza:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213ps (%rdi){1to4}, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0x75,0x18,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_a2, align 4
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%vecinit2.i = insertelement <4 x float> %vecinit.i, float %q, i32 1
@@ -1571,104 +2033,142 @@ define <4 x float> @test_mask_vfmadd128_
}
define <2 x double> @test_mask_vfmadd128_pd_r(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd128_pd_r
- ; CHECK: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0xc2]
+; CHECK-LABEL: test_mask_vfmadd128_pd_r:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
ret <2 x double> %res
}
define <2 x double> @test_mask_vfmadd128_pd_rz(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) {
- ; CHECK-LABEL: test_mask_vfmadd128_pd_rz
- ; CHECK: vfmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0xc2]
+; CHECK-LABEL: test_mask_vfmadd128_pd_rz:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213pd %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 -1) nounwind
ret <2 x double> %res
}
define <2 x double> @test_mask_vfmadd128_pd_rmk(<2 x double> %a0, <2 x double> %a1, <2 x double>* %ptr_a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd128_pd_rmk
- ; CHECK: vfmadd213pd (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_pd_rmk:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vfmadd213pd (%rdi), %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x09,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%a2 = load <2 x double>, <2 x double>* %ptr_a2
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 %mask) nounwind
ret <2 x double> %res
}
define <2 x double> @test_mask_vfmadd128_pd_rmkz(<2 x double> %a0, <2 x double> %a1, <2 x double>* %ptr_a2) {
- ; CHECK-LABEL: test_mask_vfmadd128_pd_rmkz
- ; CHECK: vfmadd213pd (%rdi), %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd128_pd_rmkz:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213pd (%rdi), %xmm1, %xmm0 ## encoding: [0x62,0xf2,0xf5,0x08,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%a2 = load <2 x double>, <2 x double>* %ptr_a2
%res = call <2 x double> @llvm.x86.avx512.mask.vfmadd.pd.128(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2, i8 -1) nounwind
ret <2 x double> %res
}
define <4 x double> @test_mask_vfmadd256_pd_r(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd256_pd_r
- ; CHECK: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0xc2]
+; CHECK-LABEL: test_mask_vfmadd256_pd_r:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
ret <4 x double> %res
}
define <4 x double> @test_mask_vfmadd256_pd_rz(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) {
- ; CHECK-LABEL: test_mask_vfmadd256_pd_rz
- ; CHECK: vfmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0xc2]
+; CHECK-LABEL: test_mask_vfmadd256_pd_rz:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213pd %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 -1) nounwind
ret <4 x double> %res
}
define <4 x double> @test_mask_vfmadd256_pd_rmk(<4 x double> %a0, <4 x double> %a1, <4 x double>* %ptr_a2, i8 %mask) {
- ; CHECK-LABEL: test_mask_vfmadd256_pd_rmk
- ; CHECK: vfmadd213pd (%rdi), %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd256_pd_rmk:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vfmadd213pd (%rdi), %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%a2 = load <4 x double>, <4 x double>* %ptr_a2
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 %mask) nounwind
ret <4 x double> %res
}
define <4 x double> @test_mask_vfmadd256_pd_rmkz(<4 x double> %a0, <4 x double> %a1, <4 x double>* %ptr_a2) {
- ; CHECK-LABEL: test_mask_vfmadd256_pd_rmkz
- ; CHECK: vfmadd213pd (%rdi), %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0x07]
+; CHECK-LABEL: test_mask_vfmadd256_pd_rmkz:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vfmadd213pd (%rdi), %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0xa8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%a2 = load <4 x double>, <4 x double>* %ptr_a2
%res = call <4 x double> @llvm.x86.avx512.mask.vfmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2, i8 -1) nounwind
ret <4 x double> %res
}
define <8 x i16> @test_mask_add_epi16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_add_epi16_rr_128
- ;CHECK: vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfd,0xc1]
+; CHECK-LABEL: test_mask_add_epi16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.padd.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_add_epi16_rrk_128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rrk_128
- ;CHECK: vpaddw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xfd,0xd1]
+; CHECK-LABEL: test_mask_add_epi16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xfd,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.padd.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_add_epi16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rrkz_128
- ;CHECK: vpaddw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xfd,0xc1]
+; CHECK-LABEL: test_mask_add_epi16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.padd.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_add_epi16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_add_epi16_rm_128
- ;CHECK: vpaddw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfd,0x07]
+; CHECK-LABEL: test_mask_add_epi16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.padd.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_add_epi16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rmk_128
- ;CHECK: vpaddw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xfd,0x0f]
+; CHECK-LABEL: test_mask_add_epi16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xfd,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.padd.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_add_epi16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rmkz_128
- ;CHECK: vpaddw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xfd,0x07]
+; CHECK-LABEL: test_mask_add_epi16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xfd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.padd.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
@@ -1677,45 +2177,63 @@ define <8 x i16> @test_mask_add_epi16_rm
declare <8 x i16> @llvm.x86.avx512.mask.padd.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <16 x i16> @test_mask_add_epi16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_add_epi16_rr_256
- ;CHECK: vpaddw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfd,0xc1]
+; CHECK-LABEL: test_mask_add_epi16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.padd.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_add_epi16_rrk_256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rrk_256
- ;CHECK: vpaddw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xfd,0xd1]
+; CHECK-LABEL: test_mask_add_epi16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xfd,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.padd.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_add_epi16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rrkz_256
- ;CHECK: vpaddw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xfd,0xc1]
+; CHECK-LABEL: test_mask_add_epi16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.padd.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_add_epi16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_add_epi16_rm_256
- ;CHECK: vpaddw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfd,0x07]
+; CHECK-LABEL: test_mask_add_epi16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.padd.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_add_epi16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rmk_256
- ;CHECK: vpaddw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xfd,0x0f]
+; CHECK-LABEL: test_mask_add_epi16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xfd,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.padd.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_add_epi16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rmkz_256
- ;CHECK: vpaddw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xfd,0x07]
+; CHECK-LABEL: test_mask_add_epi16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xfd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.padd.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
@@ -1724,45 +2242,63 @@ define <16 x i16> @test_mask_add_epi16_r
declare <16 x i16> @llvm.x86.avx512.mask.padd.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <8 x i16> @test_mask_sub_epi16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_sub_epi16_rr_128
- ;CHECK: vpsubw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xf9,0xc1]
+; CHECK-LABEL: test_mask_sub_epi16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xf9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psub.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_sub_epi16_rrk_128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rrk_128
- ;CHECK: vpsubw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xf9,0xd1]
+; CHECK-LABEL: test_mask_sub_epi16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xf9,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psub.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_sub_epi16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rrkz_128
- ;CHECK: vpsubw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xf9,0xc1]
+; CHECK-LABEL: test_mask_sub_epi16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xf9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psub.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_sub_epi16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_sub_epi16_rm_128
- ;CHECK: vpsubw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xf9,0x07]
+; CHECK-LABEL: test_mask_sub_epi16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xf9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psub.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_sub_epi16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rmk_128
- ;CHECK: vpsubw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xf9,0x0f]
+; CHECK-LABEL: test_mask_sub_epi16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xf9,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psub.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_sub_epi16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rmkz_128
- ;CHECK: vpsubw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xf9,0x07]
+; CHECK-LABEL: test_mask_sub_epi16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xf9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psub.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
@@ -1771,45 +2307,63 @@ define <8 x i16> @test_mask_sub_epi16_rm
declare <8 x i16> @llvm.x86.avx512.mask.psub.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <16 x i16> @test_mask_sub_epi16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_sub_epi16_rr_256
- ;CHECK: vpsubw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xf9,0xc1]
+; CHECK-LABEL: test_mask_sub_epi16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xf9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psub.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_sub_epi16_rrk_256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rrk_256
- ;CHECK: vpsubw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xf9,0xd1]
+; CHECK-LABEL: test_mask_sub_epi16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xf9,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psub.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_sub_epi16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rrkz_256
- ;CHECK: vpsubw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xf9,0xc1]
+; CHECK-LABEL: test_mask_sub_epi16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xf9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psub.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_sub_epi16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_sub_epi16_rm_256
- ;CHECK: vpsubw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xf9,0x07]
+; CHECK-LABEL: test_mask_sub_epi16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xf9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psub.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_sub_epi16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rmk_256
- ;CHECK: vpsubw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xf9,0x0f]
+; CHECK-LABEL: test_mask_sub_epi16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xf9,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psub.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_sub_epi16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rmkz_256
- ;CHECK: vpsubw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xf9,0x07]
+; CHECK-LABEL: test_mask_sub_epi16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xf9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psub.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
@@ -1818,45 +2372,63 @@ define <16 x i16> @test_mask_sub_epi16_r
declare <16 x i16> @llvm.x86.avx512.mask.psub.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <32 x i16> @test_mask_add_epi16_rr_512(<32 x i16> %a, <32 x i16> %b) {
- ;CHECK-LABEL: test_mask_add_epi16_rr_512
- ;CHECK: vpaddw %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc1]
+; CHECK-LABEL: test_mask_add_epi16_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddw %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.padd.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 -1)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_add_epi16_rrk_512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rrk_512
- ;CHECK: vpaddw %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xfd,0xd1]
+; CHECK-LABEL: test_mask_add_epi16_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpaddw %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xfd,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.padd.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_add_epi16_rrkz_512(<32 x i16> %a, <32 x i16> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rrkz_512
- ;CHECK: vpaddw %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xfd,0xc1]
+; CHECK-LABEL: test_mask_add_epi16_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpaddw %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.padd.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_add_epi16_rm_512(<32 x i16> %a, <32 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_add_epi16_rm_512
- ;CHECK: vpaddw (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xfd,0x07]
+; CHECK-LABEL: test_mask_add_epi16_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddw (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xfd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.padd.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 -1)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_add_epi16_rmk_512(<32 x i16> %a, <32 x i16>* %ptr_b, <32 x i16> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rmk_512
- ;CHECK: vpaddw (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xfd,0x0f]
+; CHECK-LABEL: test_mask_add_epi16_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpaddw (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xfd,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.padd.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_add_epi16_rmkz_512(<32 x i16> %a, <32 x i16>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_add_epi16_rmkz_512
- ;CHECK: vpaddw (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xfd,0x07]
+; CHECK-LABEL: test_mask_add_epi16_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpaddw (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xfd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.padd.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 %mask)
ret <32 x i16> %res
@@ -1865,45 +2437,63 @@ define <32 x i16> @test_mask_add_epi16_r
declare <32 x i16> @llvm.x86.avx512.mask.padd.w.512(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <32 x i16> @test_mask_sub_epi16_rr_512(<32 x i16> %a, <32 x i16> %b) {
- ;CHECK-LABEL: test_mask_sub_epi16_rr_512
- ;CHECK: vpsubw %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xf9,0xc1]
+; CHECK-LABEL: test_mask_sub_epi16_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubw %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xf9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psub.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 -1)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_sub_epi16_rrk_512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rrk_512
- ;CHECK: vpsubw %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xf9,0xd1]
+; CHECK-LABEL: test_mask_sub_epi16_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpsubw %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xf9,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psub.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_sub_epi16_rrkz_512(<32 x i16> %a, <32 x i16> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rrkz_512
- ;CHECK: vpsubw %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xf9,0xc1]
+; CHECK-LABEL: test_mask_sub_epi16_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpsubw %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xf9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.psub.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_sub_epi16_rm_512(<32 x i16> %a, <32 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_sub_epi16_rm_512
- ;CHECK: vpsubw (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xf9,0x07]
+; CHECK-LABEL: test_mask_sub_epi16_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubw (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xf9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.psub.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 -1)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_sub_epi16_rmk_512(<32 x i16> %a, <32 x i16>* %ptr_b, <32 x i16> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rmk_512
- ;CHECK: vpsubw (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xf9,0x0f]
+; CHECK-LABEL: test_mask_sub_epi16_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpsubw (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xf9,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.psub.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_sub_epi16_rmkz_512(<32 x i16> %a, <32 x i16>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_sub_epi16_rmkz_512
- ;CHECK: vpsubw (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xf9,0x07]
+; CHECK-LABEL: test_mask_sub_epi16_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpsubw (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xf9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.psub.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 %mask)
ret <32 x i16> %res
@@ -1912,45 +2502,63 @@ define <32 x i16> @test_mask_sub_epi16_r
declare <32 x i16> @llvm.x86.avx512.mask.psub.w.512(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <32 x i16> @test_mask_mullo_epi16_rr_512(<32 x i16> %a, <32 x i16> %b) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rr_512
- ;CHECK: vpmullw %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xd5,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi16_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullw %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xd5,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.pmull.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 -1)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_mullo_epi16_rrk_512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rrk_512
- ;CHECK: vpmullw %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xd5,0xd1]
+; CHECK-LABEL: test_mask_mullo_epi16_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpmullw %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xd5,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.pmull.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_mullo_epi16_rrkz_512(<32 x i16> %a, <32 x i16> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rrkz_512
- ;CHECK: vpmullw %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xd5,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi16_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpmullw %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xd5,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i16> @llvm.x86.avx512.mask.pmull.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_mullo_epi16_rm_512(<32 x i16> %a, <32 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rm_512
- ;CHECK: vpmullw (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xd5,0x07]
+; CHECK-LABEL: test_mask_mullo_epi16_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullw (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0xd5,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.pmull.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 -1)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_mullo_epi16_rmk_512(<32 x i16> %a, <32 x i16>* %ptr_b, <32 x i16> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rmk_512
- ;CHECK: vpmullw (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xd5,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi16_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpmullw (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x49,0xd5,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.pmull.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> %passThru, i32 %mask)
ret <32 x i16> %res
}
define <32 x i16> @test_mask_mullo_epi16_rmkz_512(<32 x i16> %a, <32 x i16>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rmkz_512
- ;CHECK: vpmullw (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xd5,0x07]
+; CHECK-LABEL: test_mask_mullo_epi16_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpmullw (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xc9,0xd5,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i16>, <32 x i16>* %ptr_b
%res = call <32 x i16> @llvm.x86.avx512.mask.pmull.w.512(<32 x i16> %a, <32 x i16> %b, <32 x i16> zeroinitializer, i32 %mask)
ret <32 x i16> %res
@@ -1959,45 +2567,63 @@ define <32 x i16> @test_mask_mullo_epi16
declare <32 x i16> @llvm.x86.avx512.mask.pmull.w.512(<32 x i16>, <32 x i16>, <32 x i16>, i32)
define <8 x i16> @test_mask_mullo_epi16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rr_128
- ;CHECK: vpmullw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd5,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd5,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmull.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_mullo_epi16_rrk_128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rrk_128
- ;CHECK: vpmullw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd5,0xd1]
+; CHECK-LABEL: test_mask_mullo_epi16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmullw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd5,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmull.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_mullo_epi16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rrkz_128
- ;CHECK: vpmullw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd5,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmullw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd5,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmull.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_mullo_epi16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rm_128
- ;CHECK: vpmullw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd5,0x07]
+; CHECK-LABEL: test_mask_mullo_epi16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd5,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.pmull.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_mullo_epi16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rmk_128
- ;CHECK: vpmullw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd5,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmullw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd5,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.pmull.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_mullo_epi16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rmkz_128
- ;CHECK: vpmullw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd5,0x07]
+; CHECK-LABEL: test_mask_mullo_epi16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmullw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd5,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.pmull.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
@@ -2006,45 +2632,63 @@ define <8 x i16> @test_mask_mullo_epi16_
declare <8 x i16> @llvm.x86.avx512.mask.pmull.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <16 x i16> @test_mask_mullo_epi16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rr_256
- ;CHECK: vpmullw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd5,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd5,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmull.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_mullo_epi16_rrk_256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rrk_256
- ;CHECK: vpmullw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd5,0xd1]
+; CHECK-LABEL: test_mask_mullo_epi16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmullw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd5,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmull.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_mullo_epi16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rrkz_256
- ;CHECK: vpmullw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd5,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmullw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd5,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmull.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_mullo_epi16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rm_256
- ;CHECK: vpmullw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd5,0x07]
+; CHECK-LABEL: test_mask_mullo_epi16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd5,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.pmull.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_mullo_epi16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rmk_256
- ;CHECK: vpmullw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd5,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmullw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd5,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.pmull.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_mullo_epi16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi16_rmkz_256
- ;CHECK: vpmullw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd5,0x07]
+; CHECK-LABEL: test_mask_mullo_epi16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmullw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd5,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.pmull.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
@@ -2054,53 +2698,73 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <8 x i16> @test_mask_packs_epi32_rr_128(<4 x i32> %a, <4 x i32> %b) {
- ;CHECK-LABEL: test_mask_packs_epi32_rr_128
- ;CHECK: vpackssdw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6b,0xc1]
+; CHECK-LABEL: test_mask_packs_epi32_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6b,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.packssdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packs_epi32_rrk_128(<4 x i32> %a, <4 x i32> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rrk_128
- ;CHECK: vpackssdw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x6b,0xd1]
+; CHECK-LABEL: test_mask_packs_epi32_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackssdw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x6b,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.packssdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packs_epi32_rrkz_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rrkz_128
- ;CHECK: vpackssdw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x6b,0xc1]
+; CHECK-LABEL: test_mask_packs_epi32_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackssdw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x6b,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.packssdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packs_epi32_rm_128(<4 x i32> %a, <4 x i32>* %ptr_b) {
- ;CHECK-LABEL: test_mask_packs_epi32_rm_128
- ;CHECK: vpackssdw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6b,0x07]
+; CHECK-LABEL: test_mask_packs_epi32_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackssdw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i32>, <4 x i32>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.packssdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packs_epi32_rmk_128(<4 x i32> %a, <4 x i32>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmk_128
- ;CHECK: vpackssdw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x6b,0x0f]
+; CHECK-LABEL: test_mask_packs_epi32_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackssdw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x6b,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i32>, <4 x i32>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.packssdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packs_epi32_rmkz_128(<4 x i32> %a, <4 x i32>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmkz_128
- ;CHECK: vpackssdw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x6b,0x07]
+; CHECK-LABEL: test_mask_packs_epi32_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackssdw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x6b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i32>, <4 x i32>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.packssdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packs_epi32_rmb_128(<4 x i32> %a, i32* %ptr_b) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmb_128
- ;CHECK: vpackssdw (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x18,0x6b,0x07]
+; CHECK-LABEL: test_mask_packs_epi32_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackssdw (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x18,0x6b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
%b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
@@ -2109,8 +2773,12 @@ define <8 x i16> @test_mask_packs_epi32_
}
define <8 x i16> @test_mask_packs_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmbk_128
- ;CHECK: vpackssdw (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x19,0x6b,0x0f]
+; CHECK-LABEL: test_mask_packs_epi32_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackssdw (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x19,0x6b,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
%b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
@@ -2119,8 +2787,11 @@ define <8 x i16> @test_mask_packs_epi32_
}
define <8 x i16> @test_mask_packs_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmbkz_128
- ;CHECK: vpackssdw (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0x6b,0x07]
+; CHECK-LABEL: test_mask_packs_epi32_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackssdw (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0x6b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
%b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
@@ -2131,53 +2802,73 @@ define <8 x i16> @test_mask_packs_epi32_
declare <8 x i16> @llvm.x86.avx512.mask.packssdw.128(<4 x i32>, <4 x i32>, <8 x i16>, i8)
define <16 x i16> @test_mask_packs_epi32_rr_256(<8 x i32> %a, <8 x i32> %b) {
- ;CHECK-LABEL: test_mask_packs_epi32_rr_256
- ;CHECK: vpackssdw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x6b,0xc1]
+; CHECK-LABEL: test_mask_packs_epi32_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackssdw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x6b,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.packssdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packs_epi32_rrk_256(<8 x i32> %a, <8 x i32> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rrk_256
- ;CHECK: vpackssdw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x6b,0xd1]
+; CHECK-LABEL: test_mask_packs_epi32_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackssdw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x6b,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.packssdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packs_epi32_rrkz_256(<8 x i32> %a, <8 x i32> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rrkz_256
- ;CHECK: vpackssdw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x6b,0xc1]
+; CHECK-LABEL: test_mask_packs_epi32_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackssdw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x6b,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.packssdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packs_epi32_rm_256(<8 x i32> %a, <8 x i32>* %ptr_b) {
- ;CHECK-LABEL: test_mask_packs_epi32_rm_256
- ;CHECK: vpackssdw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x6b,0x07]
+; CHECK-LABEL: test_mask_packs_epi32_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackssdw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x6b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i32>, <8 x i32>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.packssdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packs_epi32_rmk_256(<8 x i32> %a, <8 x i32>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmk_256
- ;CHECK: vpackssdw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x6b,0x0f]
+; CHECK-LABEL: test_mask_packs_epi32_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackssdw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x6b,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i32>, <8 x i32>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.packssdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packs_epi32_rmkz_256(<8 x i32> %a, <8 x i32>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmkz_256
- ;CHECK: vpackssdw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x6b,0x07]
+; CHECK-LABEL: test_mask_packs_epi32_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackssdw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x6b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i32>, <8 x i32>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.packssdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packs_epi32_rmb_256(<8 x i32> %a, i32* %ptr_b) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmb_256
- ;CHECK: vpackssdw (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x38,0x6b,0x07]
+; CHECK-LABEL: test_mask_packs_epi32_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackssdw (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x38,0x6b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
%b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
@@ -2186,8 +2877,12 @@ define <16 x i16> @test_mask_packs_epi32
}
define <16 x i16> @test_mask_packs_epi32_rmbk_256(<8 x i32> %a, i32* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmbk_256
- ;CHECK: vpackssdw (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x39,0x6b,0x0f]
+; CHECK-LABEL: test_mask_packs_epi32_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackssdw (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x39,0x6b,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
%b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
@@ -2196,8 +2891,11 @@ define <16 x i16> @test_mask_packs_epi32
}
define <16 x i16> @test_mask_packs_epi32_rmbkz_256(<8 x i32> %a, i32* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi32_rmbkz_256
- ;CHECK: vpackssdw (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xb9,0x6b,0x07]
+; CHECK-LABEL: test_mask_packs_epi32_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackssdw (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xb9,0x6b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
%b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
@@ -2208,45 +2906,63 @@ define <16 x i16> @test_mask_packs_epi32
declare <16 x i16> @llvm.x86.avx512.mask.packssdw.256(<8 x i32>, <8 x i32>, <16 x i16>, i16)
define <16 x i8> @test_mask_packs_epi16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_packs_epi16_rr_128
- ;CHECK: vpacksswb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x63,0xc1]
+; CHECK-LABEL: test_mask_packs_epi16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpacksswb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x63,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.packsswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packs_epi16_rrk_128(<8 x i16> %a, <8 x i16> %b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi16_rrk_128
- ;CHECK: vpacksswb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x63,0xd1]
+; CHECK-LABEL: test_mask_packs_epi16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpacksswb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x63,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.packsswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packs_epi16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi16_rrkz_128
- ;CHECK: vpacksswb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x63,0xc1]
+; CHECK-LABEL: test_mask_packs_epi16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpacksswb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x63,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.packsswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packs_epi16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_packs_epi16_rm_128
- ;CHECK: vpacksswb (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x63,0x07]
+; CHECK-LABEL: test_mask_packs_epi16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpacksswb (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x63,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.packsswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packs_epi16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi16_rmk_128
- ;CHECK: vpacksswb (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x63,0x0f]
+; CHECK-LABEL: test_mask_packs_epi16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpacksswb (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x63,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.packsswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packs_epi16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi16_rmkz_128
- ;CHECK: vpacksswb (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x63,0x07]
+; CHECK-LABEL: test_mask_packs_epi16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpacksswb (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x63,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.packsswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
@@ -2255,45 +2971,63 @@ define <16 x i8> @test_mask_packs_epi16_
declare <16 x i8> @llvm.x86.avx512.mask.packsswb.128(<8 x i16>, <8 x i16>, <16 x i8>, i16)
define <32 x i8> @test_mask_packs_epi16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_packs_epi16_rr_256
- ;CHECK: vpacksswb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x63,0xc1]
+; CHECK-LABEL: test_mask_packs_epi16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpacksswb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x63,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.packsswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packs_epi16_rrk_256(<16 x i16> %a, <16 x i16> %b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi16_rrk_256
- ;CHECK: vpacksswb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x63,0xd1]
+; CHECK-LABEL: test_mask_packs_epi16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpacksswb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x63,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.packsswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packs_epi16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi16_rrkz_256
- ;CHECK: vpacksswb %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x63,0xc1]
+; CHECK-LABEL: test_mask_packs_epi16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpacksswb %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x63,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.packsswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packs_epi16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_packs_epi16_rm_256
- ;CHECK: vpacksswb (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x63,0x07]
+; CHECK-LABEL: test_mask_packs_epi16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpacksswb (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x63,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.packsswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packs_epi16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi16_rmk_256
- ;CHECK: vpacksswb (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x63,0x0f]
+; CHECK-LABEL: test_mask_packs_epi16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpacksswb (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x63,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.packsswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packs_epi16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_packs_epi16_rmkz_256
- ;CHECK: vpacksswb (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x63,0x07]
+; CHECK-LABEL: test_mask_packs_epi16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpacksswb (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x63,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.packsswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
@@ -2303,53 +3037,73 @@ declare <32 x i8> @llvm.x86.avx512.mask.
define <8 x i16> @test_mask_packus_epi32_rr_128(<4 x i32> %a, <4 x i32> %b) {
- ;CHECK-LABEL: test_mask_packus_epi32_rr_128
- ;CHECK: vpackusdw %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_packus_epi32_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackusdw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x2b,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.packusdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packus_epi32_rrk_128(<4 x i32> %a, <4 x i32> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rrk_128
- ;CHECK: vpackusdw %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_packus_epi32_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackusdw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x2b,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.packusdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packus_epi32_rrkz_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rrkz_128
- ;CHECK: vpackusdw %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi32_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackusdw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x2b,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.packusdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packus_epi32_rm_128(<4 x i32> %a, <4 x i32>* %ptr_b) {
- ;CHECK-LABEL: test_mask_packus_epi32_rm_128
- ;CHECK: vpackusdw (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_packus_epi32_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackusdw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x2b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i32>, <4 x i32>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.packusdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packus_epi32_rmk_128(<4 x i32> %a, <4 x i32>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmk_128
- ;CHECK: vpackusdw (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_packus_epi32_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackusdw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x2b,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i32>, <4 x i32>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.packusdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packus_epi32_rmkz_128(<4 x i32> %a, <4 x i32>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmkz_128
- ;CHECK: vpackusdw (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi32_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackusdw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x2b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i32>, <4 x i32>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.packusdw.128(<4 x i32> %a, <4 x i32> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_packus_epi32_rmb_128(<4 x i32> %a, i32* %ptr_b) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmb_128
- ;CHECK: vpackusdw (%rdi){1to4}, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_packus_epi32_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackusdw (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x18,0x2b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
%b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
@@ -2358,8 +3112,12 @@ define <8 x i16> @test_mask_packus_epi32
}
define <8 x i16> @test_mask_packus_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmbk_128
- ;CHECK: vpackusdw (%rdi){1to4}, %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_packus_epi32_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackusdw (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x19,0x2b,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
%b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
@@ -2368,8 +3126,11 @@ define <8 x i16> @test_mask_packus_epi32
}
define <8 x i16> @test_mask_packus_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmbkz_128
- ;CHECK: vpackusdw (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi32_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackusdw (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x99,0x2b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
%b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
@@ -2380,53 +3141,73 @@ define <8 x i16> @test_mask_packus_epi32
declare <8 x i16> @llvm.x86.avx512.mask.packusdw.128(<4 x i32>, <4 x i32>, <8 x i16>, i8)
define <16 x i16> @test_mask_packus_epi32_rr_256(<8 x i32> %a, <8 x i32> %b) {
- ;CHECK-LABEL: test_mask_packus_epi32_rr_256
- ;CHECK: vpackusdw %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_packus_epi32_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackusdw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x2b,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.packusdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packus_epi32_rrk_256(<8 x i32> %a, <8 x i32> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rrk_256
- ;CHECK: vpackusdw %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_packus_epi32_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackusdw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x2b,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.packusdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packus_epi32_rrkz_256(<8 x i32> %a, <8 x i32> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rrkz_256
- ;CHECK: vpackusdw %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi32_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackusdw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x2b,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.packusdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packus_epi32_rm_256(<8 x i32> %a, <8 x i32>* %ptr_b) {
- ;CHECK-LABEL: test_mask_packus_epi32_rm_256
- ;CHECK: vpackusdw (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_packus_epi32_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackusdw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x2b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i32>, <8 x i32>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.packusdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packus_epi32_rmk_256(<8 x i32> %a, <8 x i32>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmk_256
- ;CHECK: vpackusdw (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_packus_epi32_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackusdw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x2b,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i32>, <8 x i32>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.packusdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packus_epi32_rmkz_256(<8 x i32> %a, <8 x i32>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmkz_256
- ;CHECK: vpackusdw (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi32_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackusdw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x2b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i32>, <8 x i32>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.packusdw.256(<8 x i32> %a, <8 x i32> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_packus_epi32_rmb_256(<8 x i32> %a, i32* %ptr_b) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmb_256
- ;CHECK: vpackusdw (%rdi){1to8}, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_packus_epi32_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackusdw (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x38,0x2b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
%b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
@@ -2435,8 +3216,12 @@ define <16 x i16> @test_mask_packus_epi3
}
define <16 x i16> @test_mask_packus_epi32_rmbk_256(<8 x i32> %a, i32* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmbk_256
- ;CHECK: vpackusdw (%rdi){1to8}, %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_packus_epi32_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackusdw (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x39,0x2b,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
%b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
@@ -2445,8 +3230,11 @@ define <16 x i16> @test_mask_packus_epi3
}
define <16 x i16> @test_mask_packus_epi32_rmbkz_256(<8 x i32> %a, i32* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi32_rmbkz_256
- ;CHECK: vpackusdw (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi32_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackusdw (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xb9,0x2b,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i32, i32* %ptr_b
%vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
%b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
@@ -2457,45 +3245,63 @@ define <16 x i16> @test_mask_packus_epi3
declare <16 x i16> @llvm.x86.avx512.mask.packusdw.256(<8 x i32>, <8 x i32>, <16 x i16>, i16)
define <16 x i8> @test_mask_packus_epi16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_packus_epi16_rr_128
- ;CHECK: vpackuswb %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_packus_epi16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackuswb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x67,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.packuswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packus_epi16_rrk_128(<8 x i16> %a, <8 x i16> %b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi16_rrk_128
- ;CHECK: vpackuswb %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_packus_epi16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackuswb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x67,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.packuswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packus_epi16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi16_rrkz_128
- ;CHECK: vpackuswb %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpackuswb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x67,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.packuswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packus_epi16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_packus_epi16_rm_128
- ;CHECK: vpackuswb (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_packus_epi16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackuswb (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x67,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.packuswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packus_epi16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi16_rmk_128
- ;CHECK: vpackuswb (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_packus_epi16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackuswb (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x67,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.packuswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_packus_epi16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi16_rmkz_128
- ;CHECK: vpackuswb (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpackuswb (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x67,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.packuswb.128(<8 x i16> %a, <8 x i16> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
@@ -2504,45 +3310,63 @@ define <16 x i8> @test_mask_packus_epi16
declare <16 x i8> @llvm.x86.avx512.mask.packuswb.128(<8 x i16>, <8 x i16>, <16 x i8>, i16)
define <32 x i8> @test_mask_packus_epi16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_packus_epi16_rr_256
- ;CHECK: vpackuswb %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_packus_epi16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackuswb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x67,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.packuswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packus_epi16_rrk_256(<16 x i16> %a, <16 x i16> %b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi16_rrk_256
- ;CHECK: vpackuswb %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_packus_epi16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpackuswb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x67,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.packuswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packus_epi16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi16_rrkz_256
- ;CHECK: vpackuswb %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpackuswb %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x67,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.packuswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packus_epi16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_packus_epi16_rm_256
- ;CHECK: vpackuswb (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_packus_epi16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpackuswb (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x67,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.packuswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packus_epi16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi16_rmk_256
- ;CHECK: vpackuswb (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_packus_epi16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpackuswb (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x67,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.packuswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_packus_epi16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_packus_epi16_rmkz_256
- ;CHECK: vpackuswb (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_packus_epi16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpackuswb (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x67,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.packuswb.256(<16 x i16> %a, <16 x i16> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
@@ -2551,45 +3375,63 @@ define <32 x i8> @test_mask_packus_epi16
declare <32 x i8> @llvm.x86.avx512.mask.packuswb.256(<16 x i16>, <16 x i16>, <32 x i8>, i32)
define <8 x i16> @test_mask_adds_epi16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_adds_epi16_rr_128
- ;CHECK: vpaddsw %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_adds_epi16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddsw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xed,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.padds.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epi16_rrk_128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi16_rrk_128
- ;CHECK: vpaddsw %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_adds_epi16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddsw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xed,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.padds.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epi16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi16_rrkz_128
- ;CHECK: vpaddsw %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epi16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddsw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xed,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.padds.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epi16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_adds_epi16_rm_128
- ;CHECK: vpaddsw (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_adds_epi16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddsw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xed,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.padds.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epi16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi16_rmk_128
- ;CHECK: vpaddsw (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_adds_epi16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddsw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xed,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.padds.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epi16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi16_rmkz_128
- ;CHECK: vpaddsw (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epi16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddsw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xed,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.padds.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
@@ -2598,45 +3440,63 @@ define <8 x i16> @test_mask_adds_epi16_r
declare <8 x i16> @llvm.x86.avx512.mask.padds.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <16 x i16> @test_mask_adds_epi16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_adds_epi16_rr_256
- ;CHECK: vpaddsw %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_adds_epi16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddsw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xed,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.padds.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epi16_rrk_256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi16_rrk_256
- ;CHECK: vpaddsw %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_adds_epi16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddsw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xed,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.padds.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epi16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi16_rrkz_256
- ;CHECK: vpaddsw %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epi16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddsw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xed,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.padds.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epi16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_adds_epi16_rm_256
- ;CHECK: vpaddsw (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_adds_epi16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddsw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xed,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.padds.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epi16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi16_rmk_256
- ;CHECK: vpaddsw (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_adds_epi16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddsw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xed,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.padds.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epi16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi16_rmkz_256
- ;CHECK: vpaddsw (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epi16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddsw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xed,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.padds.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
@@ -2645,45 +3505,63 @@ define <16 x i16> @test_mask_adds_epi16_
declare <16 x i16> @llvm.x86.avx512.mask.padds.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <8 x i16> @test_mask_subs_epi16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_subs_epi16_rr_128
- ;CHECK: vpsubsw %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_subs_epi16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubsw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psubs.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epi16_rrk_128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi16_rrk_128
- ;CHECK: vpsubsw %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_subs_epi16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubsw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe9,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psubs.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epi16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi16_rrkz_128
- ;CHECK: vpsubsw %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epi16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubsw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xe9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psubs.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epi16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_subs_epi16_rm_128
- ;CHECK: vpsubsw (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_subs_epi16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubsw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psubs.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epi16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi16_rmk_128
- ;CHECK: vpsubsw (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_subs_epi16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubsw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe9,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psubs.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epi16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi16_rmkz_128
- ;CHECK: vpsubsw (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epi16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubsw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xe9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psubs.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
@@ -2692,45 +3570,63 @@ define <8 x i16> @test_mask_subs_epi16_r
declare <8 x i16> @llvm.x86.avx512.mask.psubs.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <16 x i16> @test_mask_subs_epi16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_subs_epi16_rr_256
- ;CHECK: vpsubsw %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_subs_epi16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubsw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psubs.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epi16_rrk_256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi16_rrk_256
- ;CHECK: vpsubsw %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_subs_epi16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubsw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe9,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psubs.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epi16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi16_rrkz_256
- ;CHECK: vpsubsw %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epi16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubsw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xe9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psubs.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epi16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_subs_epi16_rm_256
- ;CHECK: vpsubsw (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_subs_epi16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubsw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psubs.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epi16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi16_rmk_256
- ;CHECK: vpsubsw (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_subs_epi16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubsw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe9,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psubs.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epi16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi16_rmkz_256
- ;CHECK: vpsubsw (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epi16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubsw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xe9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psubs.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
@@ -2739,45 +3635,63 @@ define <16 x i16> @test_mask_subs_epi16_
declare <16 x i16> @llvm.x86.avx512.mask.psubs.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <8 x i16> @test_mask_adds_epu16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_adds_epu16_rr_128
- ;CHECK: vpaddusw %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_adds_epu16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddusw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xdd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.paddus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epu16_rrk_128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu16_rrk_128
- ;CHECK: vpaddusw %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_adds_epu16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddusw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xdd,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.paddus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epu16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu16_rrkz_128
- ;CHECK: vpaddusw %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epu16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddusw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xdd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.paddus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epu16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_adds_epu16_rm_128
- ;CHECK: vpaddusw (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_adds_epu16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddusw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xdd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.paddus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epu16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu16_rmk_128
- ;CHECK: vpaddusw (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_adds_epu16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddusw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xdd,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.paddus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_adds_epu16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu16_rmkz_128
- ;CHECK: vpaddusw (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epu16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddusw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xdd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.paddus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
@@ -2786,45 +3700,63 @@ define <8 x i16> @test_mask_adds_epu16_r
declare <8 x i16> @llvm.x86.avx512.mask.paddus.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <16 x i16> @test_mask_adds_epu16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_adds_epu16_rr_256
- ;CHECK: vpaddusw %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_adds_epu16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddusw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xdd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.paddus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epu16_rrk_256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu16_rrk_256
- ;CHECK: vpaddusw %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_adds_epu16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddusw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xdd,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.paddus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epu16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu16_rrkz_256
- ;CHECK: vpaddusw %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epu16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddusw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xdd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.paddus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epu16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_adds_epu16_rm_256
- ;CHECK: vpaddusw (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_adds_epu16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddusw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xdd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.paddus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epu16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu16_rmk_256
- ;CHECK: vpaddusw (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_adds_epu16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddusw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xdd,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.paddus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_adds_epu16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu16_rmkz_256
- ;CHECK: vpaddusw (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epu16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddusw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xdd,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.paddus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
@@ -2833,45 +3765,63 @@ define <16 x i16> @test_mask_adds_epu16_
declare <16 x i16> @llvm.x86.avx512.mask.paddus.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <8 x i16> @test_mask_subs_epu16_rr_128(<8 x i16> %a, <8 x i16> %b) {
- ;CHECK-LABEL: test_mask_subs_epu16_rr_128
- ;CHECK: vpsubusw %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_subs_epu16_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubusw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psubus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epu16_rrk_128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu16_rrk_128
- ;CHECK: vpsubusw %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_subs_epu16_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubusw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd9,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psubus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epu16_rrkz_128(<8 x i16> %a, <8 x i16> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu16_rrkz_128
- ;CHECK: vpsubusw %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epu16_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubusw %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psubus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epu16_rm_128(<8 x i16> %a, <8 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_subs_epu16_rm_128
- ;CHECK: vpsubusw (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_subs_epu16_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubusw (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psubus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 -1)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epu16_rmk_128(<8 x i16> %a, <8 x i16>* %ptr_b, <8 x i16> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu16_rmk_128
- ;CHECK: vpsubusw (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_subs_epu16_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubusw (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd9,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psubus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> %passThru, i8 %mask)
ret <8 x i16> %res
}
define <8 x i16> @test_mask_subs_epu16_rmkz_128(<8 x i16> %a, <8 x i16>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu16_rmkz_128
- ;CHECK: vpsubusw (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epu16_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubusw (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i16>, <8 x i16>* %ptr_b
%res = call <8 x i16> @llvm.x86.avx512.mask.psubus.w.128(<8 x i16> %a, <8 x i16> %b, <8 x i16> zeroinitializer, i8 %mask)
ret <8 x i16> %res
@@ -2880,45 +3830,63 @@ define <8 x i16> @test_mask_subs_epu16_r
declare <8 x i16> @llvm.x86.avx512.mask.psubus.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <16 x i16> @test_mask_subs_epu16_rr_256(<16 x i16> %a, <16 x i16> %b) {
- ;CHECK-LABEL: test_mask_subs_epu16_rr_256
- ;CHECK: vpsubusw %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_subs_epu16_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubusw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psubus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epu16_rrk_256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu16_rrk_256
- ;CHECK: vpsubusw %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_subs_epu16_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubusw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd9,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psubus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epu16_rrkz_256(<16 x i16> %a, <16 x i16> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu16_rrkz_256
- ;CHECK: vpsubusw %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epu16_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubusw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd9,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psubus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epu16_rm_256(<16 x i16> %a, <16 x i16>* %ptr_b) {
- ;CHECK-LABEL: test_mask_subs_epu16_rm_256
- ;CHECK: vpsubusw (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_subs_epu16_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubusw (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psubus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 -1)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epu16_rmk_256(<16 x i16> %a, <16 x i16>* %ptr_b, <16 x i16> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu16_rmk_256
- ;CHECK: vpsubusw (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_subs_epu16_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubusw (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd9,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psubus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> %passThru, i16 %mask)
ret <16 x i16> %res
}
define <16 x i16> @test_mask_subs_epu16_rmkz_256(<16 x i16> %a, <16 x i16>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu16_rmkz_256
- ;CHECK: vpsubusw (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epu16_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubusw (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd9,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i16>, <16 x i16>* %ptr_b
%res = call <16 x i16> @llvm.x86.avx512.mask.psubus.w.256(<16 x i16> %a, <16 x i16> %b, <16 x i16> zeroinitializer, i16 %mask)
ret <16 x i16> %res
@@ -2927,45 +3895,63 @@ define <16 x i16> @test_mask_subs_epu16_
declare <16 x i16> @llvm.x86.avx512.mask.psubus.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i8> @test_mask_adds_epi8_rr_128(<16 x i8> %a, <16 x i8> %b) {
- ;CHECK-LABEL: test_mask_adds_epi8_rr_128
- ;CHECK: vpaddsb %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_adds_epi8_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddsb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xec,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.padds.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epi8_rrk_128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi8_rrk_128
- ;CHECK: vpaddsb %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_adds_epi8_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddsb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xec,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.padds.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epi8_rrkz_128(<16 x i8> %a, <16 x i8> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi8_rrkz_128
- ;CHECK: vpaddsb %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epi8_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddsb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xec,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.padds.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epi8_rm_128(<16 x i8> %a, <16 x i8>* %ptr_b) {
- ;CHECK-LABEL: test_mask_adds_epi8_rm_128
- ;CHECK: vpaddsb (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_adds_epi8_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddsb (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xec,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.padds.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epi8_rmk_128(<16 x i8> %a, <16 x i8>* %ptr_b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi8_rmk_128
- ;CHECK: vpaddsb (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_adds_epi8_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddsb (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xec,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.padds.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epi8_rmkz_128(<16 x i8> %a, <16 x i8>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi8_rmkz_128
- ;CHECK: vpaddsb (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epi8_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddsb (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xec,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.padds.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
@@ -2974,45 +3960,63 @@ define <16 x i8> @test_mask_adds_epi8_rm
declare <16 x i8> @llvm.x86.avx512.mask.padds.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <32 x i8> @test_mask_adds_epi8_rr_256(<32 x i8> %a, <32 x i8> %b) {
- ;CHECK-LABEL: test_mask_adds_epi8_rr_256
- ;CHECK: vpaddsb %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_adds_epi8_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddsb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xec,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.padds.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epi8_rrk_256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi8_rrk_256
- ;CHECK: vpaddsb %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_adds_epi8_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpaddsb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xec,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.padds.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epi8_rrkz_256(<32 x i8> %a, <32 x i8> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi8_rrkz_256
- ;CHECK: vpaddsb %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epi8_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpaddsb %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xec,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.padds.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epi8_rm_256(<32 x i8> %a, <32 x i8>* %ptr_b) {
- ;CHECK-LABEL: test_mask_adds_epi8_rm_256
- ;CHECK: vpaddsb (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_adds_epi8_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddsb (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xec,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.padds.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epi8_rmk_256(<32 x i8> %a, <32 x i8>* %ptr_b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi8_rmk_256
- ;CHECK: vpaddsb (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_adds_epi8_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpaddsb (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xec,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.padds.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epi8_rmkz_256(<32 x i8> %a, <32 x i8>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_adds_epi8_rmkz_256
- ;CHECK: vpaddsb (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epi8_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpaddsb (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xec,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.padds.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
@@ -3021,45 +4025,63 @@ define <32 x i8> @test_mask_adds_epi8_rm
declare <32 x i8> @llvm.x86.avx512.mask.padds.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <16 x i8> @test_mask_subs_epi8_rr_128(<16 x i8> %a, <16 x i8> %b) {
- ;CHECK-LABEL: test_mask_subs_epi8_rr_128
- ;CHECK: vpsubsb %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_subs_epi8_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubsb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe8,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.psubs.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epi8_rrk_128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi8_rrk_128
- ;CHECK: vpsubsb %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_subs_epi8_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubsb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe8,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.psubs.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epi8_rrkz_128(<16 x i8> %a, <16 x i8> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi8_rrkz_128
- ;CHECK: vpsubsb %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epi8_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubsb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xe8,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.psubs.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epi8_rm_128(<16 x i8> %a, <16 x i8>* %ptr_b) {
- ;CHECK-LABEL: test_mask_subs_epi8_rm_128
- ;CHECK: vpsubsb (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_subs_epi8_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubsb (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.psubs.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epi8_rmk_128(<16 x i8> %a, <16 x i8>* %ptr_b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi8_rmk_128
- ;CHECK: vpsubsb (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_subs_epi8_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubsb (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe8,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.psubs.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epi8_rmkz_128(<16 x i8> %a, <16 x i8>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi8_rmkz_128
- ;CHECK: vpsubsb (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epi8_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubsb (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xe8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.psubs.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
@@ -3068,45 +4090,63 @@ define <16 x i8> @test_mask_subs_epi8_rm
declare <16 x i8> @llvm.x86.avx512.mask.psubs.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <32 x i8> @test_mask_subs_epi8_rr_256(<32 x i8> %a, <32 x i8> %b) {
- ;CHECK-LABEL: test_mask_subs_epi8_rr_256
- ;CHECK: vpsubsb %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_subs_epi8_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubsb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe8,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.psubs.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epi8_rrk_256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi8_rrk_256
- ;CHECK: vpsubsb %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_subs_epi8_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpsubsb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe8,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.psubs.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epi8_rrkz_256(<32 x i8> %a, <32 x i8> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi8_rrkz_256
- ;CHECK: vpsubsb %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epi8_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpsubsb %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xe8,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.psubs.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epi8_rm_256(<32 x i8> %a, <32 x i8>* %ptr_b) {
- ;CHECK-LABEL: test_mask_subs_epi8_rm_256
- ;CHECK: vpsubsb (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_subs_epi8_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubsb (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.psubs.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epi8_rmk_256(<32 x i8> %a, <32 x i8>* %ptr_b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi8_rmk_256
- ;CHECK: vpsubsb (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_subs_epi8_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpsubsb (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe8,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.psubs.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epi8_rmkz_256(<32 x i8> %a, <32 x i8>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_subs_epi8_rmkz_256
- ;CHECK: vpsubsb (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epi8_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpsubsb (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xe8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.psubs.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
@@ -3115,45 +4155,63 @@ define <32 x i8> @test_mask_subs_epi8_rm
declare <32 x i8> @llvm.x86.avx512.mask.psubs.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <16 x i8> @test_mask_adds_epu8_rr_128(<16 x i8> %a, <16 x i8> %b) {
- ;CHECK-LABEL: test_mask_adds_epu8_rr_128
- ;CHECK: vpaddusb %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_adds_epu8_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddusb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xdc,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.paddus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epu8_rrk_128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu8_rrk_128
- ;CHECK: vpaddusb %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_adds_epu8_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddusb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xdc,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.paddus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epu8_rrkz_128(<16 x i8> %a, <16 x i8> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu8_rrkz_128
- ;CHECK: vpaddusb %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epu8_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpaddusb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xdc,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.paddus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epu8_rm_128(<16 x i8> %a, <16 x i8>* %ptr_b) {
- ;CHECK-LABEL: test_mask_adds_epu8_rm_128
- ;CHECK: vpaddusb (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_adds_epu8_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddusb (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xdc,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.paddus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epu8_rmk_128(<16 x i8> %a, <16 x i8>* %ptr_b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu8_rmk_128
- ;CHECK: vpaddusb (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_adds_epu8_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddusb (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xdc,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.paddus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_adds_epu8_rmkz_128(<16 x i8> %a, <16 x i8>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu8_rmkz_128
- ;CHECK: vpaddusb (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epu8_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpaddusb (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xdc,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.paddus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
@@ -3162,45 +4220,63 @@ define <16 x i8> @test_mask_adds_epu8_rm
declare <16 x i8> @llvm.x86.avx512.mask.paddus.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <32 x i8> @test_mask_adds_epu8_rr_256(<32 x i8> %a, <32 x i8> %b) {
- ;CHECK-LABEL: test_mask_adds_epu8_rr_256
- ;CHECK: vpaddusb %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_adds_epu8_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddusb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xdc,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.paddus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epu8_rrk_256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu8_rrk_256
- ;CHECK: vpaddusb %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_adds_epu8_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpaddusb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xdc,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.paddus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epu8_rrkz_256(<32 x i8> %a, <32 x i8> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu8_rrkz_256
- ;CHECK: vpaddusb %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epu8_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpaddusb %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xdc,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.paddus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epu8_rm_256(<32 x i8> %a, <32 x i8>* %ptr_b) {
- ;CHECK-LABEL: test_mask_adds_epu8_rm_256
- ;CHECK: vpaddusb (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_adds_epu8_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpaddusb (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xdc,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.paddus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epu8_rmk_256(<32 x i8> %a, <32 x i8>* %ptr_b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu8_rmk_256
- ;CHECK: vpaddusb (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_adds_epu8_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpaddusb (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xdc,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.paddus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_adds_epu8_rmkz_256(<32 x i8> %a, <32 x i8>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_adds_epu8_rmkz_256
- ;CHECK: vpaddusb (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_adds_epu8_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpaddusb (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xdc,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.paddus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
@@ -3209,45 +4285,63 @@ define <32 x i8> @test_mask_adds_epu8_rm
declare <32 x i8> @llvm.x86.avx512.mask.paddus.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <16 x i8> @test_mask_subs_epu8_rr_128(<16 x i8> %a, <16 x i8> %b) {
- ;CHECK-LABEL: test_mask_subs_epu8_rr_128
- ;CHECK: vpsubusb %xmm1, %xmm0, %xmm0
+; CHECK-LABEL: test_mask_subs_epu8_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubusb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd8,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.psubus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epu8_rrk_128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu8_rrk_128
- ;CHECK: vpsubusb %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK-LABEL: test_mask_subs_epu8_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubusb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd8,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.psubus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epu8_rrkz_128(<16 x i8> %a, <16 x i8> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu8_rrkz_128
- ;CHECK: vpsubusb %xmm1, %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epu8_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsubusb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd8,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.psubus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epu8_rm_128(<16 x i8> %a, <16 x i8>* %ptr_b) {
- ;CHECK-LABEL: test_mask_subs_epu8_rm_128
- ;CHECK: vpsubusb (%rdi), %xmm0, %xmm0
+; CHECK-LABEL: test_mask_subs_epu8_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubusb (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.psubus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 -1)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epu8_rmk_128(<16 x i8> %a, <16 x i8>* %ptr_b, <16 x i8> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu8_rmk_128
- ;CHECK: vpsubusb (%rdi), %xmm0, %xmm1 {%k1}
+; CHECK-LABEL: test_mask_subs_epu8_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubusb (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd8,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.psubus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> %passThru, i16 %mask)
ret <16 x i8> %res
}
define <16 x i8> @test_mask_subs_epu8_rmkz_128(<16 x i8> %a, <16 x i8>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu8_rmkz_128
- ;CHECK: vpsubusb (%rdi), %xmm0, %xmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epu8_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsubusb (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x i8>, <16 x i8>* %ptr_b
%res = call <16 x i8> @llvm.x86.avx512.mask.psubus.b.128(<16 x i8> %a, <16 x i8> %b, <16 x i8> zeroinitializer, i16 %mask)
ret <16 x i8> %res
@@ -3256,45 +4350,63 @@ define <16 x i8> @test_mask_subs_epu8_rm
declare <16 x i8> @llvm.x86.avx512.mask.psubus.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <32 x i8> @test_mask_subs_epu8_rr_256(<32 x i8> %a, <32 x i8> %b) {
- ;CHECK-LABEL: test_mask_subs_epu8_rr_256
- ;CHECK: vpsubusb %ymm1, %ymm0, %ymm0
+; CHECK-LABEL: test_mask_subs_epu8_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubusb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd8,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.psubus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epu8_rrk_256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu8_rrk_256
- ;CHECK: vpsubusb %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK-LABEL: test_mask_subs_epu8_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpsubusb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd8,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.psubus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epu8_rrkz_256(<32 x i8> %a, <32 x i8> %b, i32 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu8_rrkz_256
- ;CHECK: vpsubusb %ymm1, %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epu8_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpsubusb %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd8,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.psubus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epu8_rm_256(<32 x i8> %a, <32 x i8>* %ptr_b) {
- ;CHECK-LABEL: test_mask_subs_epu8_rm_256
- ;CHECK: vpsubusb (%rdi), %ymm0, %ymm0
+; CHECK-LABEL: test_mask_subs_epu8_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpsubusb (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.psubus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 -1)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epu8_rmk_256(<32 x i8> %a, <32 x i8>* %ptr_b, <32 x i8> %passThru, i32 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu8_rmk_256
- ;CHECK: vpsubusb (%rdi), %ymm0, %ymm1 {%k1}
+; CHECK-LABEL: test_mask_subs_epu8_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpsubusb (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd8,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.psubus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> %passThru, i32 %mask)
ret <32 x i8> %res
}
define <32 x i8> @test_mask_subs_epu8_rmkz_256(<32 x i8> %a, <32 x i8>* %ptr_b, i32 %mask) {
- ;CHECK-LABEL: test_mask_subs_epu8_rmkz_256
- ;CHECK: vpsubusb (%rdi), %ymm0, %ymm0 {%k1} {z}
+; CHECK-LABEL: test_mask_subs_epu8_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpsubusb (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd8,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <32 x i8>, <32 x i8>* %ptr_b
%res = call <32 x i8> @llvm.x86.avx512.mask.psubus.b.256(<32 x i8> %a, <32 x i8> %b, <32 x i8> zeroinitializer, i32 %mask)
ret <32 x i8> %res
@@ -3305,10 +4417,13 @@ declare <32 x i8> @llvm.x86.avx512.mask.
declare <16 x i8> @llvm.x86.avx512.mask.pmaxs.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <16 x i8>@test_int_x86_avx512_mask_pmaxs_b_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %mask) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxs_b_128
-; CHECK-NOT: call
-; CHECK: vpmaxsb %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmaxs_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaxsb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x3c,0xd1]
+; CHECK-NEXT: vpmaxsb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x3c,0xc1]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.pmaxs.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2 ,i16 %mask)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmaxs.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> zeroinitializer, i16 %mask)
%res2 = add <16 x i8> %res, %res1
@@ -3318,10 +4433,13 @@ define <16 x i8>@test_int_x86_avx512_mas
declare <32 x i8> @llvm.x86.avx512.mask.pmaxs.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <32 x i8>@test_int_x86_avx512_mask_pmaxs_b_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxs_b_256
-; CHECK-NOT: call
-; CHECK: vpmaxsb %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmaxs_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpmaxsb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x3c,0xd1]
+; CHECK-NEXT: vpmaxsb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x3c,0xc1]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.pmaxs.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.pmaxs.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -3331,10 +4449,13 @@ define <32 x i8>@test_int_x86_avx512_mas
declare <8 x i16> @llvm.x86.avx512.mask.pmaxs.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pmaxs_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxs_w_128
-; CHECK-NOT: call
-; CHECK: vpmaxsw %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmaxs_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaxsw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xee,0xd1]
+; CHECK-NEXT: vpmaxsw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xee,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmaxs.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmaxs.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3344,10 +4465,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pmaxs.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pmaxs_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %mask) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxs_w_256
-; CHECK-NOT: call
-; CHECK: vpmaxsw %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmaxs_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaxsw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xee,0xd1]
+; CHECK-NEXT: vpmaxsw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xee,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmaxs.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %mask)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmaxs.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %mask)
%res2 = add <16 x i16> %res, %res1
@@ -3357,10 +4481,13 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <16 x i8> @llvm.x86.avx512.mask.pmaxu.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <16 x i8>@test_int_x86_avx512_mask_pmaxu_b_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2,i16 %mask) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxu_b_128
-; CHECK-NOT: call
-; CHECK: vpmaxub %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmaxu_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaxub %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xde,0xd1]
+; CHECK-NEXT: vpmaxub %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xde,0xc1]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.pmaxu.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %mask)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmaxu.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> zeroinitializer, i16 %mask)
%res2 = add <16 x i8> %res, %res1
@@ -3370,10 +4497,13 @@ define <16 x i8>@test_int_x86_avx512_mas
declare <32 x i8> @llvm.x86.avx512.mask.pmaxu.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <32 x i8>@test_int_x86_avx512_mask_pmaxu_b_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxu_b_256
-; CHECK-NOT: call
-; CHECK: vpmaxub %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmaxu_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpmaxub %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xde,0xd1]
+; CHECK-NEXT: vpmaxub %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xde,0xc1]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.pmaxu.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.pmaxu.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -3383,10 +4513,13 @@ define <32 x i8>@test_int_x86_avx512_mas
declare <8 x i16> @llvm.x86.avx512.mask.pmaxu.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pmaxu_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxu_w_128
-; CHECK-NOT: call
-; CHECK: vpmaxuw %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmaxu_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaxuw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x3e,0xd1]
+; CHECK-NEXT: vpmaxuw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x3e,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmaxu.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmaxu.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3396,10 +4529,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pmaxu.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pmaxu_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %mask) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmaxu_w_256
-; CHECK-NOT: call
-; CHECK: vpmaxuw %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmaxu_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaxuw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x3e,0xd1]
+; CHECK-NEXT: vpmaxuw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x3e,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmaxu.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %mask)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmaxu.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %mask)
%res2 = add <16 x i16> %res, %res1
@@ -3409,10 +4545,13 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <16 x i8> @llvm.x86.avx512.mask.pmins.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <16 x i8>@test_int_x86_avx512_mask_pmins_b_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %mask) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmins_b_128
-; CHECK-NOT: call
-; CHECK: vpminsb %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmins_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpminsb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x38,0xd1]
+; CHECK-NEXT: vpminsb %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x38,0xc1]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.pmins.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %mask)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmins.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> zeroinitializer, i16 %mask)
%res2 = add <16 x i8> %res, %res1
@@ -3422,10 +4561,13 @@ define <16 x i8>@test_int_x86_avx512_mas
declare <32 x i8> @llvm.x86.avx512.mask.pmins.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <32 x i8>@test_int_x86_avx512_mask_pmins_b_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmins_b_256
-; CHECK-NOT: call
-; CHECK: vpminsb %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmins_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpminsb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x38,0xd1]
+; CHECK-NEXT: vpminsb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x38,0xc1]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.pmins.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.pmins.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -3435,10 +4577,13 @@ define <32 x i8>@test_int_x86_avx512_mas
declare <8 x i16> @llvm.x86.avx512.mask.pmins.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pmins_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmins_w_128
-; CHECK-NOT: call
-; CHECK: vpminsw %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmins_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpminsw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xea,0xd1]
+; CHECK-NEXT: vpminsw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xea,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmins.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmins.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3448,10 +4593,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pmins.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pmins_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %mask) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmins_w_256
-; CHECK-NOT: call
-; CHECK: vpminsw %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pmins_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpminsw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xea,0xd1]
+; CHECK-NEXT: vpminsw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xea,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmins.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %mask)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmins.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %mask)
%res2 = add <16 x i16> %res, %res1
@@ -3461,10 +4609,13 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <16 x i8> @llvm.x86.avx512.mask.pminu.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <16 x i8>@test_int_x86_avx512_mask_pminu_b_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %mask) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pminu_b_128
-; CHECK-NOT: call
-; CHECK: vpminub %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pminu_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpminub %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xda,0xd1]
+; CHECK-NEXT: vpminub %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xda,0xc1]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.pminu.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %mask)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pminu.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> zeroinitializer, i16 %mask)
%res2 = add <16 x i8> %res, %res1
@@ -3474,10 +4625,13 @@ define <16 x i8>@test_int_x86_avx512_mas
declare <32 x i8> @llvm.x86.avx512.mask.pminu.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <32 x i8>@test_int_x86_avx512_mask_pminu_b_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pminu_b_256
-; CHECK-NOT: call
-; CHECK: vpminub %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pminu_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpminub %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xda,0xd1]
+; CHECK-NEXT: vpminub %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xda,0xc1]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.pminu.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.pminu.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -3487,10 +4641,13 @@ define <32 x i8>@test_int_x86_avx512_mas
declare <8 x i16> @llvm.x86.avx512.mask.pminu.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pminu_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pminu_w_128
-; CHECK-NOT: call
-; CHECK: vpminuw %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pminu_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpminuw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x3a,0xd1]
+; CHECK-NEXT: vpminuw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x3a,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pminu.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pminu.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3500,10 +4657,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pminu.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pminu_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %mask) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pminu_w_256
-; CHECK-NOT: call
-; CHECK: vpminuw %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pminu_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpminuw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x3a,0xd1]
+; CHECK-NEXT: vpminuw %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x3a,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pminu.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %mask)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pminu.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %mask)
%res2 = add <16 x i16> %res, %res1
@@ -3513,11 +4673,14 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <8 x i16> @llvm.x86.avx512.mask.vpermt2var.hi.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_vpermt2var_hi_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_vpermt2var_hi_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpermt2w %xmm{{.*}}{%k1}
-; CHECK-NOT: {z}
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermt2var_hi_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm1, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; CHECK-NEXT: vpermt2w %xmm2, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x7d,0xda]
+; CHECK-NEXT: vpermt2w %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0x7d,0xca]
+; CHECK-NEXT: vpaddw %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x65,0x08,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.vpermt2var.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.vpermt2var.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3527,10 +4690,14 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <8 x i16> @llvm.x86.avx512.maskz.vpermt2var.hi.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_maskz_vpermt2var_hi_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_maskz_vpermt2var_hi_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpermt2w %xmm{{.*}}{%k1} {z}
+; CHECK-LABEL: test_int_x86_avx512_maskz_vpermt2var_hi_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm1, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; CHECK-NEXT: vpermt2w %xmm2, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x7d,0xda]
+; CHECK-NEXT: vpermt2w %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0x7d,0xca]
+; CHECK-NEXT: vpaddw %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x65,0x08,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.maskz.vpermt2var.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.maskz.vpermt2var.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3540,10 +4707,14 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.vpermt2var.hi.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_vpermt2var_hi_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_vpermt2var_hi_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpermt2w %ymm{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermt2var_hi_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm1, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; CHECK-NEXT: vpermt2w %ymm2, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x7d,0xda]
+; CHECK-NEXT: vpermt2w %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0x7d,0xca]
+; CHECK-NEXT: vpaddw %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.vpermt2var.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.vpermt2var.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -3553,10 +4724,14 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <16 x i16> @llvm.x86.avx512.maskz.vpermt2var.hi.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_maskz_vpermt2var_hi_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_maskz_vpermt2var_hi_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpermt2w %ymm{{.*}}{%k1} {z}
+; CHECK-LABEL: test_int_x86_avx512_maskz_vpermt2var_hi_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm1, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; CHECK-NEXT: vpermt2w %ymm2, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x7d,0xda]
+; CHECK-NEXT: vpermt2w %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0x7d,0xca]
+; CHECK-NEXT: vpaddw %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.maskz.vpermt2var.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.maskz.vpermt2var.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -3566,10 +4741,14 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <8 x i16> @llvm.x86.avx512.mask.vpermi2var.hi.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_vpermi2var_hi_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_vpermi2var_hi_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpermi2w %xmm{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermi2var_hi_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm1, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; CHECK-NEXT: vpermi2w %xmm2, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x75,0xda]
+; CHECK-NEXT: vpermi2w %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0x75,0xca]
+; CHECK-NEXT: vpaddw %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x65,0x08,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.vpermi2var.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.vpermi2var.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3579,10 +4758,14 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.vpermi2var.hi.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_vpermi2var_hi_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_vpermi2var_hi_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpermi2w %ymm{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_vpermi2var_hi_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovaps %zmm1, %zmm3 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xd9]
+; CHECK-NEXT: vpermi2w %ymm2, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x75,0xda]
+; CHECK-NEXT: vpermi2w %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0x75,0xca]
+; CHECK-NEXT: vpaddw %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.vpermi2var.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.vpermi2var.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -3592,10 +4775,13 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <16 x i8> @llvm.x86.avx512.mask.pavg.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <16 x i8>@test_int_x86_avx512_mask_pavg_b_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pavg_b_128
-; CHECK-NOT: call
-; CHECK: vpavgb %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pavg_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpavgb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe0,0xd1]
+; CHECK-NEXT: vpavgb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe0,0xc1]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.pavg.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pavg.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 -1)
%res2 = add <16 x i8> %res, %res1
@@ -3605,10 +4791,13 @@ define <16 x i8>@test_int_x86_avx512_mas
declare <32 x i8> @llvm.x86.avx512.mask.pavg.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <32 x i8>@test_int_x86_avx512_mask_pavg_b_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pavg_b_256
-; CHECK-NOT: call
-; CHECK: vpavgb %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pavg_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpavgb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe0,0xd1]
+; CHECK-NEXT: vpavgb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe0,0xc1]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.pavg.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.pavg.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -3618,10 +4807,13 @@ define <32 x i8>@test_int_x86_avx512_mas
declare <8 x i16> @llvm.x86.avx512.mask.pavg.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pavg_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pavg_w_128
-; CHECK-NOT: call
-; CHECK: vpavgw %xmm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pavg_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpavgw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe3,0xd1]
+; CHECK-NEXT: vpavgw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe3,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pavg.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pavg.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3631,10 +4823,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pavg.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pavg_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pavg_w_256
-; CHECK-NOT: call
-; CHECK: vpavgw %ymm
-; CHECK: {%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pavg_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpavgw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe3,0xd1]
+; CHECK-NEXT: vpavgw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe3,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pavg.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pavg.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -3644,10 +4839,13 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <16 x i8> @llvm.x86.avx512.mask.pshuf.b.128(<16 x i8>, <16 x i8>, <16 x i8>, i16)
define <16 x i8>@test_int_x86_avx512_mask_pshuf_b_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pshuf_b_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpshufb %xmm{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pshuf_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpshufb %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x00,0xd1]
+; CHECK-NEXT: vpshufb %xmm1, %xmm0, %xmm0 ## encoding: [0xc4,0xe2,0x79,0x00,0xc1]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.pshuf.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pshuf.b.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 -1)
%res2 = add <16 x i8> %res, %res1
@@ -3657,10 +4855,13 @@ define <16 x i8>@test_int_x86_avx512_mas
declare <32 x i8> @llvm.x86.avx512.mask.pshuf.b.256(<32 x i8>, <32 x i8>, <32 x i8>, i32)
define <32 x i8>@test_int_x86_avx512_mask_pshuf_b_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pshuf_b_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpshufb %ymm{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pshuf_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpshufb %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x00,0xd1]
+; CHECK-NEXT: vpshufb %ymm1, %ymm0, %ymm0 ## encoding: [0xc4,0xe2,0x7d,0x00,0xc1]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.pshuf.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.pshuf.b.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -3670,10 +4871,13 @@ define <32 x i8>@test_int_x86_avx512_mas
declare <16 x i8> @llvm.x86.avx512.mask.pabs.b.128(<16 x i8>, <16 x i8>, i16)
define <16 x i8>@test_int_x86_avx512_mask_pabs_b_128(<16 x i8> %x0, <16 x i8> %x1, i16 %x2) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pabs_b_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpabsb{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pabs_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpabsb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x1c,0xc8]
+; CHECK-NEXT: vpabsb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x1c,0xc0]
+; CHECK-NEXT: vpaddb %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.pabs.b.128(<16 x i8> %x0, <16 x i8> %x1, i16 %x2)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pabs.b.128(<16 x i8> %x0, <16 x i8> %x1, i16 -1)
%res2 = add <16 x i8> %res, %res1
@@ -3683,10 +4887,13 @@ define <16 x i8>@test_int_x86_avx512_mas
declare <32 x i8> @llvm.x86.avx512.mask.pabs.b.256(<32 x i8>, <32 x i8>, i32)
define <32 x i8>@test_int_x86_avx512_mask_pabs_b_256(<32 x i8> %x0, <32 x i8> %x1, i32 %x2) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pabs_b_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpabsb{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pabs_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpabsb %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x1c,0xc8]
+; CHECK-NEXT: vpabsb %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x1c,0xc0]
+; CHECK-NEXT: vpaddb %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.pabs.b.256(<32 x i8> %x0, <32 x i8> %x1, i32 %x2)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.pabs.b.256(<32 x i8> %x0, <32 x i8> %x1, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -3696,10 +4903,13 @@ define <32 x i8>@test_int_x86_avx512_mas
declare <8 x i16> @llvm.x86.avx512.mask.pabs.w.128(<8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pabs_w_128(<8 x i16> %x0, <8 x i16> %x1, i8 %x2) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pabs_w_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpabsw{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pabs_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpabsw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x1d,0xc8]
+; CHECK-NEXT: vpabsw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x1d,0xc0]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pabs.w.128(<8 x i16> %x0, <8 x i16> %x1, i8 %x2)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pabs.w.128(<8 x i16> %x0, <8 x i16> %x1, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3709,10 +4919,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pabs.w.256(<16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pabs_w_256(<16 x i16> %x0, <16 x i16> %x1, i16 %x2) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pabs_w_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vpabsw{{.*}}{%k1}
+; CHECK-LABEL: test_int_x86_avx512_mask_pabs_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpabsw %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x1d,0xc8]
+; CHECK-NEXT: vpabsw %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x1d,0xc0]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pabs.w.256(<16 x i16> %x0, <16 x i16> %x1, i16 %x2)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pabs.w.256(<16 x i16> %x0, <16 x i16> %x1, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -3720,32 +4933,44 @@ define <16 x i16>@test_int_x86_avx512_ma
}
define <32 x i8> @test_x86_mask_blend_b_256(i32 %a0, <32 x i8> %a1, <32 x i8> %a2) {
-; CHECK-LABEL: test_x86_mask_blend_b_256
-; CHECK: vpblendmb
+; CHECK-LABEL: test_x86_mask_blend_b_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpblendmb %ymm1, %ymm0, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x66,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.blend.b.256(<32 x i8> %a1, <32 x i8> %a2, i32 %a0) ; <<32 x i8>> [#uses=1]
ret <32 x i8> %res
}
declare <32 x i8> @llvm.x86.avx512.mask.blend.b.256(<32 x i8>, <32 x i8>, i32) nounwind readonly
define <16 x i16> @test_x86_mask_blend_w_256(i16 %mask, <16 x i16> %a1, <16 x i16> %a2) {
-; CHECK-LABEL: test_x86_mask_blend_w_256
-; CHECK: vpblendmw
+; CHECK-LABEL: test_x86_mask_blend_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpblendmw %ymm1, %ymm0, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x66,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.blend.w.256(<16 x i16> %a1, <16 x i16> %a2, i16 %mask) ; <<16 x i16>> [#uses=1]
ret <16 x i16> %res
}
declare <16 x i16> @llvm.x86.avx512.mask.blend.w.256(<16 x i16>, <16 x i16>, i16) nounwind readonly
define <16 x i8> @test_x86_mask_blend_b_128(i16 %a0, <16 x i8> %a1, <16 x i8> %a2) {
-; CHECK-LABEL: test_x86_mask_blend_b_128
-; CHECK: vpblendmb
+; CHECK-LABEL: test_x86_mask_blend_b_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpblendmb %xmm1, %xmm0, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x66,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.blend.b.128(<16 x i8> %a1, <16 x i8> %a2, i16 %a0) ; <<16 x i8>> [#uses=1]
ret <16 x i8> %res
}
declare <16 x i8> @llvm.x86.avx512.mask.blend.b.128(<16 x i8>, <16 x i8>, i16) nounwind readonly
define <8 x i16> @test_x86_mask_blend_w_128(i8 %mask, <8 x i16> %a1, <8 x i16> %a2) {
-; CHECK-LABEL: test_x86_mask_blend_w_128
-; CHECK: vpblendmw
+; CHECK-LABEL: test_x86_mask_blend_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpblendmw %xmm1, %xmm0, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x66,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.blend.w.128(<8 x i16> %a1, <8 x i16> %a2, i8 %mask) ; <<8 x i16>> [#uses=1]
ret <8 x i16> %res
}
@@ -3754,11 +4979,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
declare <8 x i16> @llvm.x86.avx512.mask.pmulhu.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pmulhu_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmulhu_w_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: {%k1}
-; CHECK: vpmulhuw {{.*}}encoding: [0x62
+; CHECK-LABEL: test_int_x86_avx512_mask_pmulhu_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmulhuw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe4,0xd1]
+; CHECK-NEXT: vpmulhuw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe4,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmulhu.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmulhu.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3768,11 +4995,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pmulhu.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pmulhu_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmulhu_w_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: {%k1}
-; CHECK: vpmulhuw {{.*}}encoding: [0x62
+; CHECK-LABEL: test_int_x86_avx512_mask_pmulhu_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmulhuw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe4,0xd1]
+; CHECK-NEXT: vpmulhuw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe4,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmulhu.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmulhu.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -3782,11 +5011,13 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <8 x i16> @llvm.x86.avx512.mask.pmulh.w.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pmulh_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmulh_w_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: {%k1}
-; CHECK: vpmulhw {{.*}}encoding: [0x62
+; CHECK-LABEL: test_int_x86_avx512_mask_pmulh_w_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmulhw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe5,0xd1]
+; CHECK-NEXT: vpmulhw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe5,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmulh.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmulh.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3796,11 +5027,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pmulh.w.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pmulh_w_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmulh_w_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: {%k1}
-; CHECK: vpmulhw {{.*}}encoding: [0x62
+; CHECK-LABEL: test_int_x86_avx512_mask_pmulh_w_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmulhw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe5,0xd1]
+; CHECK-NEXT: vpmulhw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe5,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmulh.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmulh.w.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -3810,11 +5043,13 @@ define <16 x i16>@test_int_x86_avx512_ma
declare <8 x i16> @llvm.x86.avx512.mask.pmul.hr.sw.128(<8 x i16>, <8 x i16>, <8 x i16>, i8)
define <8 x i16>@test_int_x86_avx512_mask_pmulhr_sw_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmulhr_sw_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: {%k1}
-; CHECK: vpmulhrsw {{.*}}encoding: [0x62
+; CHECK-LABEL: test_int_x86_avx512_mask_pmulhr_sw_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmulhrsw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x0b,0xd1]
+; CHECK-NEXT: vpmulhrsw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x0b,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmul.hr.sw.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmul.hr.sw.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -3824,11 +5059,13 @@ define <8 x i16>@test_int_x86_avx512_mas
declare <16 x i16> @llvm.x86.avx512.mask.pmul.hr.sw.256(<16 x i16>, <16 x i16>, <16 x i16>, i16)
define <16 x i16>@test_int_x86_avx512_mask_pmulhr_sw_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
-; CHECK-LABEL: @test_int_x86_avx512_mask_pmulhr_sw_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: {%k1}
-; CHECK: vpmulhrsw {{.*}}encoding: [0x62
+; CHECK-LABEL: test_int_x86_avx512_mask_pmulhr_sw_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmulhrsw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x0b,0xd1]
+; CHECK-NEXT: vpmulhrsw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x0b,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmul.hr.sw.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmul.hr.sw.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -3839,9 +5076,14 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_pmov_wb_128(<8 x i16> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_wb_128:
-; CHECK: vpmovwb %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpmovwb %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpmovwb %xmm0, %xmm0
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovwb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x30,0xc1]
+; CHECK-NEXT: vpmovwb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x30,0xc2]
+; CHECK-NEXT: vpmovwb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x30,0xc0]
+; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc1]
+; CHECK-NEXT: vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <16 x i8> @llvm.x86.avx512.mask.pmov.wb.128(<8 x i16> %x0, <16 x i8> %x1, i8 -1)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmov.wb.128(<8 x i16> %x0, <16 x i8> %x1, i8 %x2)
%res2 = call <16 x i8> @llvm.x86.avx512.mask.pmov.wb.128(<8 x i16> %x0, <16 x i8> zeroinitializer, i8 %x2)
@@ -3854,8 +5096,11 @@ declare void @llvm.x86.avx512.mask.pmov.
define void @test_int_x86_avx512_mask_pmov_wb_mem_128(i8* %ptr, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_wb_mem_128:
-; CHECK: vpmovwb %xmm0, (%rdi)
-; CHECK: vpmovwb %xmm0, (%rdi) {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmovwb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x30,0x07]
+; CHECK-NEXT: vpmovwb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x30,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.pmov.wb.mem.128(i8* %ptr, <8 x i16> %x1, i8 -1)
call void @llvm.x86.avx512.mask.pmov.wb.mem.128(i8* %ptr, <8 x i16> %x1, i8 %x2)
ret void
@@ -3865,9 +5110,14 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_pmovs_wb_128(<8 x i16> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_wb_128:
-; CHECK: vpmovswb %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpmovswb %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpmovswb %xmm0, %xmm0
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovswb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x20,0xc1]
+; CHECK-NEXT: vpmovswb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x20,0xc2]
+; CHECK-NEXT: vpmovswb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x20,0xc0]
+; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc1]
+; CHECK-NEXT: vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.wb.128(<8 x i16> %x0, <16 x i8> %x1, i8 -1)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.wb.128(<8 x i16> %x0, <16 x i8> %x1, i8 %x2)
%res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.wb.128(<8 x i16> %x0, <16 x i8> zeroinitializer, i8 %x2)
@@ -3880,8 +5130,11 @@ declare void @llvm.x86.avx512.mask.pmovs
define void @test_int_x86_avx512_mask_pmovs_wb_mem_128(i8* %ptr, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_wb_mem_128:
-; CHECK: vpmovswb %xmm0, (%rdi)
-; CHECK: vpmovswb %xmm0, (%rdi) {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmovswb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x20,0x07]
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmovswb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x20,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.pmovs.wb.mem.128(i8* %ptr, <8 x i16> %x1, i8 -1)
call void @llvm.x86.avx512.mask.pmovs.wb.mem.128(i8* %ptr, <8 x i16> %x1, i8 %x2)
ret void
@@ -3891,9 +5144,14 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_pmovus_wb_128(<8 x i16> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_wb_128:
-; CHECK: vpmovuswb %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpmovuswb %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpmovuswb %xmm0, %xmm0
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovuswb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x10,0xc1]
+; CHECK-NEXT: vpmovuswb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x10,0xc2]
+; CHECK-NEXT: vpmovuswb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x10,0xc0]
+; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc1]
+; CHECK-NEXT: vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.wb.128(<8 x i16> %x0, <16 x i8> %x1, i8 -1)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.wb.128(<8 x i16> %x0, <16 x i8> %x1, i8 %x2)
%res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.wb.128(<8 x i16> %x0, <16 x i8> zeroinitializer, i8 %x2)
@@ -3906,8 +5164,11 @@ declare void @llvm.x86.avx512.mask.pmovu
define void @test_int_x86_avx512_mask_pmovus_wb_mem_128(i8* %ptr, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_wb_mem_128:
-; CHECK: vpmovuswb %xmm0, (%rdi)
-; CHECK: vpmovuswb %xmm0, (%rdi) {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmovuswb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x10,0x07]
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmovuswb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x10,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.pmovus.wb.mem.128(i8* %ptr, <8 x i16> %x1, i8 -1)
call void @llvm.x86.avx512.mask.pmovus.wb.mem.128(i8* %ptr, <8 x i16> %x1, i8 %x2)
ret void
@@ -3917,9 +5178,14 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_pmov_wb_256(<16 x i16> %x0, <16 x i8> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_wb_256:
-; CHECK: vpmovwb %ymm0, %xmm1 {%k1}
-; CHECK-NEXT: vpmovwb %ymm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpmovwb %ymm0, %xmm0
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovwb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x30,0xc1]
+; CHECK-NEXT: vpmovwb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x30,0xc2]
+; CHECK-NEXT: vpmovwb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x30,0xc0]
+; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc1]
+; CHECK-NEXT: vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <16 x i8> @llvm.x86.avx512.mask.pmov.wb.256(<16 x i16> %x0, <16 x i8> %x1, i16 -1)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmov.wb.256(<16 x i16> %x0, <16 x i8> %x1, i16 %x2)
%res2 = call <16 x i8> @llvm.x86.avx512.mask.pmov.wb.256(<16 x i16> %x0, <16 x i8> zeroinitializer, i16 %x2)
@@ -3932,8 +5198,11 @@ declare void @llvm.x86.avx512.mask.pmov.
define void @test_int_x86_avx512_mask_pmov_wb_mem_256(i8* %ptr, <16 x i16> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_wb_mem_256:
-; CHECK: vpmovwb %ymm0, (%rdi)
-; CHECK: vpmovwb %ymm0, (%rdi) {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmovwb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x30,0x07]
+; CHECK-NEXT: vpmovwb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x30,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.pmov.wb.mem.256(i8* %ptr, <16 x i16> %x1, i16 -1)
call void @llvm.x86.avx512.mask.pmov.wb.mem.256(i8* %ptr, <16 x i16> %x1, i16 %x2)
ret void
@@ -3943,9 +5212,14 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_pmovs_wb_256(<16 x i16> %x0, <16 x i8> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_wb_256:
-; CHECK: vpmovswb %ymm0, %xmm1 {%k1}
-; CHECK-NEXT: vpmovswb %ymm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpmovswb %ymm0, %xmm0
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovswb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x20,0xc1]
+; CHECK-NEXT: vpmovswb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x20,0xc2]
+; CHECK-NEXT: vpmovswb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x20,0xc0]
+; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc1]
+; CHECK-NEXT: vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.wb.256(<16 x i16> %x0, <16 x i8> %x1, i16 -1)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.wb.256(<16 x i16> %x0, <16 x i8> %x1, i16 %x2)
%res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.wb.256(<16 x i16> %x0, <16 x i8> zeroinitializer, i16 %x2)
@@ -3958,8 +5232,11 @@ declare void @llvm.x86.avx512.mask.pmovs
define void @test_int_x86_avx512_mask_pmovs_wb_mem_256(i8* %ptr, <16 x i16> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_wb_mem_256:
-; CHECK: vpmovswb %ymm0, (%rdi)
-; CHECK: vpmovswb %ymm0, (%rdi) {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmovswb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x20,0x07]
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmovswb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x20,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.pmovs.wb.mem.256(i8* %ptr, <16 x i16> %x1, i16 -1)
call void @llvm.x86.avx512.mask.pmovs.wb.mem.256(i8* %ptr, <16 x i16> %x1, i16 %x2)
ret void
@@ -3969,9 +5246,14 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_pmovus_wb_256(<16 x i16> %x0, <16 x i8> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_wb_256:
-; CHECK: vpmovuswb %ymm0, %xmm1 {%k1}
-; CHECK-NEXT: vpmovuswb %ymm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpmovuswb %ymm0, %xmm0
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovuswb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x10,0xc1]
+; CHECK-NEXT: vpmovuswb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x10,0xc2]
+; CHECK-NEXT: vpmovuswb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x10,0xc0]
+; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc1]
+; CHECK-NEXT: vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.wb.256(<16 x i16> %x0, <16 x i8> %x1, i16 -1)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.wb.256(<16 x i16> %x0, <16 x i8> %x1, i16 %x2)
%res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.wb.256(<16 x i16> %x0, <16 x i8> zeroinitializer, i16 %x2)
@@ -3984,8 +5266,11 @@ declare void @llvm.x86.avx512.mask.pmovu
define void @test_int_x86_avx512_mask_pmovus_wb_mem_256(i8* %ptr, <16 x i16> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_wb_mem_256:
-; CHECK: vpmovuswb %ymm0, (%rdi)
-; CHECK: vpmovuswb %ymm0, (%rdi) {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmovuswb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x10,0x07]
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpmovuswb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x10,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.pmovus.wb.mem.256(i8* %ptr, <16 x i16> %x1, i16 -1)
call void @llvm.x86.avx512.mask.pmovus.wb.mem.256(i8* %ptr, <16 x i16> %x1, i16 %x2)
ret void
@@ -3996,11 +5281,11 @@ declare <4 x i32> @llvm.x86.avx512.mask.
define <4 x i32>@test_int_x86_avx512_mask_pmaddw_d_128(<8 x i16> %x0, <8 x i16> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaddw_d_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmaddwd %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpmaddwd %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddd %xmm0, %xmm2, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaddwd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xf5,0xd1]
+; CHECK-NEXT: vpmaddwd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xf5,0xc1]
+; CHECK-NEXT: vpaddd %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i32> @llvm.x86.avx512.mask.pmaddw.d.128(<8 x i16> %x0, <8 x i16> %x1, <4 x i32> %x2, i8 %x3)
%res1 = call <4 x i32> @llvm.x86.avx512.mask.pmaddw.d.128(<8 x i16> %x0, <8 x i16> %x1, <4 x i32> %x2, i8 -1)
%res2 = add <4 x i32> %res, %res1
@@ -4012,11 +5297,11 @@ declare <8 x i32> @llvm.x86.avx512.mask.
define <8 x i32>@test_int_x86_avx512_mask_pmaddw_d_256(<16 x i16> %x0, <16 x i16> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaddw_d_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmaddwd %ymm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpmaddwd %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddd %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaddwd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xf5,0xd1]
+; CHECK-NEXT: vpmaddwd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xf5,0xc1]
+; CHECK-NEXT: vpaddd %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i32> @llvm.x86.avx512.mask.pmaddw.d.256(<16 x i16> %x0, <16 x i16> %x1, <8 x i32> %x2, i8 %x3)
%res1 = call <8 x i32> @llvm.x86.avx512.mask.pmaddw.d.256(<16 x i16> %x0, <16 x i16> %x1, <8 x i32> %x2, i8 -1)
%res2 = add <8 x i32> %res, %res1
@@ -4028,11 +5313,11 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_pmaddubs_w_128(<16 x i8> %x0, <16 x i8> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaddubs_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmaddubsw %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpmaddubsw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaddubsw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x04,0xd1]
+; CHECK-NEXT: vpmaddubsw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x04,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmaddubs.w.128(<16 x i8> %x0, <16 x i8> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmaddubs.w.128(<16 x i8> %x0, <16 x i8> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -4044,11 +5329,11 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_pmaddubs_w_256(<32 x i8> %x0, <32 x i8> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaddubs_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmaddubsw %ymm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpmaddubsw %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmaddubsw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x04,0xd1]
+; CHECK-NEXT: vpmaddubsw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x04,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmaddubs.w.256(<32 x i8> %x0, <32 x i8> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmaddubs.w.256(<32 x i8> %x0, <32 x i8> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -4059,10 +5344,14 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_punpckhb_w_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckhb_w_128:
-; CHECK: vpunpckhbw %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpunpckhbw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x68,0xd1]
; CHECK-NEXT: ## xmm2 = xmm2[8],k1[8],xmm2[9],k1[9],xmm2[10],k1[10],xmm2[11],k1[11],xmm2[12],k1[12],xmm2[13],k1[13],xmm2[14],k1[14],xmm2[15],k1[15]
; CHECK-NEXT: vpunpckhbw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x68,0xc1]
; CHECK-NEXT: ## xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.punpckhb.w.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.punpckhb.w.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 -1)
%res2 = add <16 x i8> %res, %res1
@@ -4073,10 +5362,14 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_punpcklb_w_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpcklb_w_128:
-; CHECK: vpunpcklbw %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpunpcklbw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x60,0xd1]
; CHECK-NEXT: ## xmm2 = xmm2[0],k1[0],xmm2[1],k1[1],xmm2[2],k1[2],xmm2[3],k1[3],xmm2[4],k1[4],xmm2[5],k1[5],xmm2[6],k1[6],xmm2[7],k1[7]
; CHECK-NEXT: vpunpcklbw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x60,0xc1]
; CHECK-NEXT: ## xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.punpcklb.w.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 %x3)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.punpcklb.w.128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x2, i16 -1)
%res2 = add <16 x i8> %res, %res1
@@ -4087,10 +5380,14 @@ declare <32 x i8> @llvm.x86.avx512.mask.
define <32 x i8>@test_int_x86_avx512_mask_punpckhb_w_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckhb_w_256:
-; CHECK: vpunpckhbw %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpunpckhbw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x68,0xd1]
; CHECK-NEXT: ## ymm2 = ymm2[8],k1[8],ymm2[9],k1[9],ymm2[10],k1[10],ymm2[11],k1[11],ymm2[12],k1[12],ymm2[13],k1[13],ymm2[14],k1[14],ymm2[15],k1[15],ymm2[24],k1[24],ymm2[25],k1[25],ymm2[26],k1[26],ymm2[27],k1[27],ymm2[28],k1[28],ymm2[29],k1[29],ymm2[30],k1[30],ymm2[31],k1[31]
; CHECK-NEXT: vpunpckhbw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x68,0xc1]
; CHECK-NEXT: ## ymm0 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.punpckhb.w.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.punpckhb.w.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -4101,10 +5398,14 @@ declare <32 x i8> @llvm.x86.avx512.mask.
define <32 x i8>@test_int_x86_avx512_mask_punpcklb_w_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpcklb_w_256:
-; CHECK: vpunpcklbw %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpunpcklbw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x60,0xd1]
; CHECK-NEXT: ## ymm2 = ymm2[0],k1[0],ymm2[1],k1[1],ymm2[2],k1[2],ymm2[3],k1[3],ymm2[4],k1[4],ymm2[5],k1[5],ymm2[6],k1[6],ymm2[7],k1[7],ymm2[16],k1[16],ymm2[17],k1[17],ymm2[18],k1[18],ymm2[19],k1[19],ymm2[20],k1[20],ymm2[21],k1[21],ymm2[22],k1[22],ymm2[23],k1[23]
; CHECK-NEXT: vpunpcklbw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x60,0xc1]
; CHECK-NEXT: ## ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[16],ymm1[16],ymm0[17],ymm1[17],ymm0[18],ymm1[18],ymm0[19],ymm1[19],ymm0[20],ymm1[20],ymm0[21],ymm1[21],ymm0[22],ymm1[22],ymm0[23],ymm1[23]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.punpcklb.w.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 %x3)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.punpcklb.w.256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x2, i32 -1)
%res2 = add <32 x i8> %res, %res1
@@ -4115,10 +5416,14 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_punpcklw_d_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpcklw_d_128:
-; CHECK: vpunpcklwd %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpunpcklwd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x61,0xd1]
; CHECK-NEXT: ## xmm2 = xmm2[0],k1[0],xmm2[1],k1[1],xmm2[2],k1[2],xmm2[3],k1[3]
; CHECK-NEXT: vpunpcklwd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x61,0xc1]
; CHECK-NEXT: ## xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.punpcklw.d.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.punpcklw.d.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -4129,10 +5434,14 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_punpckhw_d_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckhw_d_128:
-; CHECK: vpunpckhwd %xmm1, %xmm0, %xmm2 {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpunpckhwd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x69,0xd1]
; CHECK-NEXT: ## xmm2 = xmm2[4],k1[4],xmm2[5],k1[5],xmm2[6],k1[6],xmm2[7],k1[7]
; CHECK-NEXT: vpunpckhwd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x69,0xc1]
; CHECK-NEXT: ## xmm0 = xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.punpckhw.d.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.punpckhw.d.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = add <8 x i16> %res, %res1
@@ -4143,10 +5452,14 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_punpcklw_d_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpcklw_d_256:
-; CHECK: vpunpcklwd %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpunpcklwd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x61,0xd1]
; CHECK-NEXT: ## ymm2 = ymm2[0],k1[0],ymm2[1],k1[1],ymm2[2],k1[2],ymm2[3],k1[3],ymm2[8],k1[8],ymm2[9],k1[9],ymm2[10],k1[10],ymm2[11],k1[11]
; CHECK-NEXT: vpunpcklwd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x61,0xc1]
; CHECK-NEXT: ## ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.punpcklw.d.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.punpcklw.d.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -4157,10 +5470,14 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_punpckhw_d_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckhw_d_256:
-; CHECK: vpunpckhwd %ymm1, %ymm0, %ymm2 {%k1}
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpunpckhwd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x69,0xd1]
; CHECK-NEXT: ## ymm2 = ymm2[4],k1[4],ymm2[5],k1[5],ymm2[6],k1[6],ymm2[7],k1[7],ymm2[12],k1[12],ymm2[13],k1[13],ymm2[14],k1[14],ymm2[15],k1[15]
; CHECK-NEXT: vpunpckhwd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x69,0xc1]
; CHECK-NEXT: ## ymm0 = ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.punpckhw.d.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.punpckhw.d.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = add <16 x i16> %res, %res1
@@ -4172,13 +5489,13 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_palignr_128(<16 x i8> %x0, <16 x i8> %x1, <16 x i8> %x3, i16 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_palignr_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpalignr $2, %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpalignr $2, %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vpalignr $2, %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddb %xmm3, %xmm2, %xmm1
-; CHECK-NEXT: vpaddb %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpalignr $2, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x0f,0xd1,0x02]
+; CHECK-NEXT: vpalignr $2, %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0x89,0x0f,0xd9,0x02]
+; CHECK-NEXT: vpalignr $2, %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x0f,0xc1,0x02]
+; CHECK-NEXT: vpaddb %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xcb]
+; CHECK-NEXT: vpaddb %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.palignr.128(<16 x i8> %x0, <16 x i8> %x1, i32 2, <16 x i8> %x3, i16 %x4)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.palignr.128(<16 x i8> %x0, <16 x i8> %x1, i32 2, <16 x i8> zeroinitializer, i16 %x4)
%res2 = call <16 x i8> @llvm.x86.avx512.mask.palignr.128(<16 x i8> %x0, <16 x i8> %x1, i32 2, <16 x i8> %x3, i16 -1)
@@ -4192,13 +5509,13 @@ declare <32 x i8> @llvm.x86.avx512.mask.
define <32 x i8>@test_int_x86_avx512_mask_palignr_256(<32 x i8> %x0, <32 x i8> %x1, <32 x i8> %x3, i32 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_palignr_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovd %edi, %k1
-; CHECK-NEXT: vpalignr $2, %ymm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpalignr $2, %ymm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vpalignr $2, %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddb %ymm3, %ymm2, %ymm1
-; CHECK-NEXT: vpaddb %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpalignr $2, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x0f,0xd1,0x02]
+; CHECK-NEXT: vpalignr $2, %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x0f,0xd9,0x02]
+; CHECK-NEXT: vpalignr $2, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x0f,0xc1,0x02]
+; CHECK-NEXT: vpaddb %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xcb]
+; CHECK-NEXT: vpaddb %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.palignr.256(<32 x i8> %x0, <32 x i8> %x1, i32 2, <32 x i8> %x3, i32 %x4)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.palignr.256(<32 x i8> %x0, <32 x i8> %x1, i32 2, <32 x i8> zeroinitializer, i32 %x4)
%res2 = call <32 x i8> @llvm.x86.avx512.mask.palignr.256(<32 x i8> %x0, <32 x i8> %x1, i32 2, <32 x i8> %x3, i32 -1)
@@ -4212,13 +5529,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_dbpsadbw_128(<16 x i8> %x0, <16 x i8> %x1, <8 x i16> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_dbpsadbw_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1
-; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x42,0xd1,0x02]
+; CHECK-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0x89,0x42,0xd9,0x02]
+; CHECK-NEXT: vdbpsadbw $2, %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x42,0xc1,0x02]
+; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.dbpsadbw.128(<16 x i8> %x0, <16 x i8> %x1, i32 2, <8 x i16> %x3, i8 %x4)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.dbpsadbw.128(<16 x i8> %x0, <16 x i8> %x1, i32 2, <8 x i16> zeroinitializer, i8 %x4)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.dbpsadbw.128(<16 x i8> %x0, <16 x i8> %x1, i32 2, <8 x i16> %x3, i8 -1)
@@ -4232,13 +5549,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_dbpsadbw_256(<32 x i8> %x0, <32 x i8> %x1, <16 x i16> %x3, i16 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_dbpsadbw_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x42,0xd1,0x02]
+; CHECK-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x42,0xd9,0x02]
+; CHECK-NEXT: vdbpsadbw $2, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x42,0xc1,0x02]
+; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.dbpsadbw.256(<32 x i8> %x0, <32 x i8> %x1, i32 2, <16 x i16> %x3, i16 %x4)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.dbpsadbw.256(<32 x i8> %x0, <32 x i8> %x1, i32 2, <16 x i16> zeroinitializer, i16 %x4)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.dbpsadbw.256(<32 x i8> %x0, <32 x i8> %x1, i32 2, <16 x i16> %x3, i16 -1)
@@ -4252,13 +5569,13 @@ declare <32 x i8> @llvm.x86.avx512.pbroa
define <32 x i8>@test_int_x86_avx512_pbroadcastb_256(<16 x i8> %x0, <32 x i8> %x1, i32 %mask) {
; CHECK-LABEL: test_int_x86_avx512_pbroadcastb_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovd %edi, %k1
-; CHECK-NEXT: vpbroadcastb %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vpbroadcastb %xmm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpbroadcastb %xmm0, %ymm0
-; CHECK-NEXT: vpaddb %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vpbroadcastb %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x78,0xc8]
+; CHECK-NEXT: vpbroadcastb %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x78,0xd0]
+; CHECK-NEXT: vpbroadcastb %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x78,0xc0]
+; CHECK-NEXT: vpaddb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfc,0xc1]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.pbroadcastb.256(<16 x i8> %x0, <32 x i8> %x1, i32 -1)
%res1 = call <32 x i8> @llvm.x86.avx512.pbroadcastb.256(<16 x i8> %x0, <32 x i8> %x1, i32 %mask)
%res2 = call <32 x i8> @llvm.x86.avx512.pbroadcastb.256(<16 x i8> %x0, <32 x i8> zeroinitializer, i32 %mask)
@@ -4272,13 +5589,13 @@ declare <16 x i8> @llvm.x86.avx512.pbroa
define <16 x i8>@test_int_x86_avx512_pbroadcastb_128(<16 x i8> %x0, <16 x i8> %x1, i16 %mask) {
; CHECK-LABEL: test_int_x86_avx512_pbroadcastb_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpbroadcastb %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpbroadcastb %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpbroadcastb %xmm0, %xmm0
-; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpbroadcastb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x78,0xc8]
+; CHECK-NEXT: vpbroadcastb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x78,0xd0]
+; CHECK-NEXT: vpbroadcastb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x78,0xc0]
+; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc1]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.pbroadcastb.128(<16 x i8> %x0, <16 x i8> %x1, i16 -1)
%res1 = call <16 x i8> @llvm.x86.avx512.pbroadcastb.128(<16 x i8> %x0, <16 x i8> %x1, i16 %mask)
%res2 = call <16 x i8> @llvm.x86.avx512.pbroadcastb.128(<16 x i8> %x0, <16 x i8> zeroinitializer, i16 %mask)
@@ -4292,13 +5609,13 @@ declare <16 x i16> @llvm.x86.avx512.pbro
define <16 x i16>@test_int_x86_avx512_pbroadcastw_256(<8 x i16> %x0, <16 x i16> %x1, i16 %mask) {
; CHECK-LABEL: test_int_x86_avx512_pbroadcastw_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpbroadcastw %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vpbroadcastw %xmm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpbroadcastw %xmm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpbroadcastw %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x79,0xc8]
+; CHECK-NEXT: vpbroadcastw %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x79,0xd0]
+; CHECK-NEXT: vpbroadcastw %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x79,0xc0]
+; CHECK-NEXT: vpaddw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfd,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.pbroadcastw.256(<8 x i16> %x0, <16 x i16> %x1, i16 -1)
%res1 = call <16 x i16> @llvm.x86.avx512.pbroadcastw.256(<8 x i16> %x0, <16 x i16> %x1, i16 %mask)
%res2 = call <16 x i16> @llvm.x86.avx512.pbroadcastw.256(<8 x i16> %x0, <16 x i16> zeroinitializer, i16 %mask)
@@ -4312,13 +5629,13 @@ declare <8 x i16> @llvm.x86.avx512.pbroa
define <8 x i16>@test_int_x86_avx512_pbroadcastw_128(<8 x i16> %x0, <8 x i16> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_pbroadcastw_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpbroadcastw %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpbroadcastw %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpbroadcastw %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpbroadcastw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x79,0xc8]
+; CHECK-NEXT: vpbroadcastw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x79,0xd0]
+; CHECK-NEXT: vpbroadcastw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x79,0xc0]
+; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfd,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.pbroadcastw.128(<8 x i16> %x0, <8 x i16> %x1, i8 -1)
%res1 = call <8 x i16> @llvm.x86.avx512.pbroadcastw.128(<8 x i16> %x0, <8 x i16> %x1, i8 %mask)
%res2 = call <8 x i16> @llvm.x86.avx512.pbroadcastw.128(<8 x i16> %x0, <8 x i16> zeroinitializer, i8 %mask)
@@ -4372,9 +5689,9 @@ declare i16 @llvm.x86.avx512.cvtb2mask.1
define i16 at test_int_x86_avx512_cvtb2mask_128(<16 x i8> %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtb2mask_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpmovb2m %xmm0, %k0
-; CHECK-NEXT: kmovw %k0, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpmovb2m %xmm0, %k0 ## encoding: [0x62,0xf2,0x7e,0x08,0x29,0xc0]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.cvtb2mask.128(<16 x i8> %x0)
ret i16 %res
}
@@ -4384,9 +5701,9 @@ declare i32 @llvm.x86.avx512.cvtb2mask.2
define i32 at test_int_x86_avx512_cvtb2mask_256(<32 x i8> %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtb2mask_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpmovb2m %ymm0, %k0
-; CHECK-NEXT: kmovd %k0, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpmovb2m %ymm0, %k0 ## encoding: [0x62,0xf2,0x7e,0x28,0x29,0xc0]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i32 @llvm.x86.avx512.cvtb2mask.256(<32 x i8> %x0)
ret i32 %res
}
@@ -4396,9 +5713,9 @@ declare i8 @llvm.x86.avx512.cvtw2mask.12
define i8 at test_int_x86_avx512_cvtw2mask_128(<8 x i16> %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtw2mask_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpmovw2m %xmm0, %k0
-; CHECK-NEXT: kmovw %k0, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpmovw2m %xmm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x08,0x29,0xc0]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.cvtw2mask.128(<8 x i16> %x0)
ret i8 %res
}
@@ -4408,9 +5725,9 @@ declare i16 @llvm.x86.avx512.cvtw2mask.2
define i16 at test_int_x86_avx512_cvtw2mask_256(<16 x i16> %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtw2mask_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpmovw2m %ymm0, %k0
-; CHECK-NEXT: kmovw %k0, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpmovw2m %ymm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x28,0x29,0xc0]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.cvtw2mask.256(<16 x i16> %x0)
ret i16 %res
}
@@ -4420,9 +5737,9 @@ declare <16 x i8> @llvm.x86.avx512.cvtma
define <16 x i8>@test_int_x86_avx512_cvtmask2b_128(i16 %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtmask2b_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k0
-; CHECK-NEXT: vpmovm2b %k0, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k0 ## encoding: [0xc5,0xf8,0x92,0xc7]
+; CHECK-NEXT: vpmovm2b %k0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x28,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.cvtmask2b.128(i16 %x0)
ret <16 x i8> %res
}
@@ -4432,9 +5749,9 @@ declare <32 x i8> @llvm.x86.avx512.cvtma
define <32 x i8>@test_int_x86_avx512_cvtmask2b_256(i32 %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtmask2b_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovd %edi, %k0
-; CHECK-NEXT: vpmovm2b %k0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovd %edi, %k0 ## encoding: [0xc5,0xfb,0x92,0xc7]
+; CHECK-NEXT: vpmovm2b %k0, %ymm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x28,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.cvtmask2b.256(i32 %x0)
ret <32 x i8> %res
}
@@ -4444,9 +5761,9 @@ declare <8 x i16> @llvm.x86.avx512.cvtma
define <8 x i16>@test_int_x86_avx512_cvtmask2w_128(i8 %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtmask2w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k0
-; CHECK-NEXT: vpmovm2w %k0, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k0 ## encoding: [0xc5,0xf8,0x92,0xc7]
+; CHECK-NEXT: vpmovm2w %k0, %xmm0 ## encoding: [0x62,0xf2,0xfe,0x08,0x28,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.cvtmask2w.128(i8 %x0)
ret <8 x i16> %res
}
@@ -4456,9 +5773,9 @@ declare <16 x i16> @llvm.x86.avx512.cvtm
define <16 x i16>@test_int_x86_avx512_cvtmask2w_256(i16 %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtmask2w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k0
-; CHECK-NEXT: vpmovm2w %k0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k0 ## encoding: [0xc5,0xf8,0x92,0xc7]
+; CHECK-NEXT: vpmovm2w %k0, %ymm0 ## encoding: [0x62,0xf2,0xfe,0x28,0x28,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.cvtmask2w.256(i16 %x0)
ret <16 x i16> %res
}
@@ -4468,13 +5785,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psrl_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsrlw %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpsrlw %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vpsrlw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0
-; CHECK-NEXT: vpaddw %xmm0, %xmm3, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsrlw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd1,0xd1]
+; CHECK-NEXT: vpsrlw %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd1,0xd9]
+; CHECK-NEXT: vpsrlw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd1,0xc1]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: vpaddw %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x65,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psrl.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psrl.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psrl.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
@@ -4488,13 +5805,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psrl_w_256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsrlw %xmm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpsrlw %xmm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vpsrlw %xmm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: vpaddw %ymm3, %ymm0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsrlw %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd1,0xd1]
+; CHECK-NEXT: vpsrlw %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd1,0xd9]
+; CHECK-NEXT: vpsrlw %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd1,0xc1]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: vpaddw %ymm3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfd,0xc3]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psrl.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psrl.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 -1)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psrl.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
@@ -4508,13 +5825,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psrl_wi_128(<8 x i16> %x0, i32 %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_wi_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpsrlw $3, %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpsrlw $3, %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpsrlw $3, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsrlw $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x09,0x71,0xd0,0x03]
+; CHECK-NEXT: vpsrlw $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0x89,0x71,0xd0,0x03]
+; CHECK-NEXT: vpsrlw $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x71,0xd0,0x03]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 -1)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psrl.wi.128(<8 x i16> %x0, i32 3, <8 x i16> zeroinitializer, i8 %x3)
@@ -4528,13 +5845,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psrl_wi_256(<16 x i16> %x0, i32 %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_wi_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpsrlw $3, %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vpsrlw $3, %ymm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpsrlw $3, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: vpaddw %ymm2, %ymm0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsrlw $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x29,0x71,0xd0,0x03]
+; CHECK-NEXT: vpsrlw $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xa9,0x71,0xd0,0x03]
+; CHECK-NEXT: vpsrlw $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x71,0xd0,0x03]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: vpaddw %ymm2, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfd,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 -1)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psrl.wi.256(<16 x i16> %x0, i32 3, <16 x i16> zeroinitializer, i16 %x3)
@@ -4548,13 +5865,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psrlv16_hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrlv16_hi:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsrlvw %ymm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpsrlvw %ymm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vpsrlvw %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsrlvw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x10,0xd1]
+; CHECK-NEXT: vpsrlvw %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x10,0xd9]
+; CHECK-NEXT: vpsrlvw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x10,0xc1]
+; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psrlv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
@@ -4568,13 +5885,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psrlv8_hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrlv8_hi:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsrlvw %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpsrlvw %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vpsrlvw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsrlvw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x10,0xd1]
+; CHECK-NEXT: vpsrlvw %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x10,0xd9]
+; CHECK-NEXT: vpsrlvw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x10,0xc1]
+; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psrlv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
@@ -4588,13 +5905,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psra_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsraw %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpsraw %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vpsraw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsraw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe1,0xd1]
+; CHECK-NEXT: vpsraw %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xe1,0xd9]
+; CHECK-NEXT: vpsraw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe1,0xc1]
+; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psra.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psra.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psra.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
@@ -4608,13 +5925,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psra_wi_128(<8 x i16> %x0, i32 %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_wi_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpsraw $3, %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpsraw $3, %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpsraw $3, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsraw $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x09,0x71,0xe0,0x03]
+; CHECK-NEXT: vpsraw $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0x89,0x71,0xe0,0x03]
+; CHECK-NEXT: vpsraw $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x71,0xe0,0x03]
+; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xca]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psra.wi.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psra.wi.128(<8 x i16> %x0, i32 3, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psra.wi.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 -1)
@@ -4628,13 +5945,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psra_w_256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsraw %xmm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpsraw %xmm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vpsraw %xmm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsraw %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe1,0xd1]
+; CHECK-NEXT: vpsraw %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xe1,0xd9]
+; CHECK-NEXT: vpsraw %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe1,0xc1]
+; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psra.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psra.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psra.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 -1)
@@ -4648,13 +5965,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psra_wi_256(<16 x i16> %x0, i32 %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_wi_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpsraw $3, %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vpsraw $3, %ymm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpsraw $3, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsraw $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x29,0x71,0xe0,0x03]
+; CHECK-NEXT: vpsraw $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xa9,0x71,0xe0,0x03]
+; CHECK-NEXT: vpsraw $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x71,0xe0,0x03]
+; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xca]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psra.wi.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psra.wi.256(<16 x i16> %x0, i32 3, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psra.wi.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 -1)
@@ -4668,14 +5985,14 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_pshufh_w_128(<8 x i16> %x0, i32 %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pshufh_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpshufhw $3, %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpshufhw $3, %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpshufhw $3, %xmm0, %xmm0
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpshufhw $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x09,0x70,0xc8,0x03]
+; CHECK-NEXT: vpshufhw $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x7e,0x89,0x70,0xd0,0x03]
+; CHECK-NEXT: vpshufhw $3, %xmm0, %xmm0 ## encoding: [0xc5,0xfa,0x70,0xc0,0x03]
; CHECK-NEXT: ## xmm0 = xmm0[0,1,2,3,7,4,4,4]
-; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xca]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pshufh.w.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pshufh.w.128(<8 x i16> %x0, i32 3, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.pshufh.w.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 -1)
@@ -4689,14 +6006,14 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_pshufh_w_256(<16 x i16> %x0, i32 %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pshufh_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpshufhw $3, %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vpshufhw $3, %ymm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpshufhw $3, %ymm0, %ymm0
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpshufhw $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x29,0x70,0xc8,0x03]
+; CHECK-NEXT: vpshufhw $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x7e,0xa9,0x70,0xd0,0x03]
+; CHECK-NEXT: vpshufhw $3, %ymm0, %ymm0 ## encoding: [0xc5,0xfe,0x70,0xc0,0x03]
; CHECK-NEXT: ## ymm0 = ymm0[0,1,2,3,7,4,4,4,8,9,10,11,15,12,12,12]
-; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xca]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pshufh.w.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pshufh.w.256(<16 x i16> %x0, i32 3, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.pshufh.w.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 -1)
@@ -4710,14 +6027,14 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_pshufl_w_128(<8 x i16> %x0, i32 %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pshufl_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpshuflw $3, %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpshuflw $3, %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpshuflw $3, %xmm0, %xmm0
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpshuflw $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7f,0x09,0x70,0xc8,0x03]
+; CHECK-NEXT: vpshuflw $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x7f,0x89,0x70,0xd0,0x03]
+; CHECK-NEXT: vpshuflw $3, %xmm0, %xmm0 ## encoding: [0xc5,0xfb,0x70,0xc0,0x03]
; CHECK-NEXT: ## xmm0 = xmm0[3,0,0,0,4,5,6,7]
-; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xca]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pshufl.w.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pshufl.w.128(<8 x i16> %x0, i32 3, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.pshufl.w.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 -1)
@@ -4731,14 +6048,14 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_pshufl_w_256(<16 x i16> %x0, i32 %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pshufl_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpshuflw $3, %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vpshuflw $3, %ymm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpshuflw $3, %ymm0, %ymm0
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpshuflw $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7f,0x29,0x70,0xc8,0x03]
+; CHECK-NEXT: vpshuflw $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x7f,0xa9,0x70,0xd0,0x03]
+; CHECK-NEXT: vpshuflw $3, %ymm0, %ymm0 ## encoding: [0xc5,0xff,0x70,0xc0,0x03]
; CHECK-NEXT: ## ymm0 = ymm0[3,0,0,0,4,5,6,7,11,8,8,8,12,13,14,15]
-; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xca]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pshufl.w.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pshufl.w.256(<16 x i16> %x0, i32 3, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.pshufl.w.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 -1)
@@ -4752,13 +6069,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psrav16_hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrav16_hi:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsravw %ymm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpsravw %ymm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vpsravw %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsravw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x11,0xd1]
+; CHECK-NEXT: vpsravw %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x11,0xd9]
+; CHECK-NEXT: vpsravw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x11,0xc1]
+; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psrav16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psrav16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psrav16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
@@ -4772,13 +6089,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psrav8_hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrav8_hi:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsravw %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpsravw %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vpsravw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsravw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x11,0xd1]
+; CHECK-NEXT: vpsravw %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x11,0xd9]
+; CHECK-NEXT: vpsravw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x11,0xc1]
+; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psrav8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psrav8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psrav8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
@@ -4793,13 +6110,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psll_w_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsllw %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpsllw %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vpsllw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsllw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xf1,0xd1]
+; CHECK-NEXT: vpsllw %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xf1,0xd9]
+; CHECK-NEXT: vpsllw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xf1,0xc1]
+; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psll.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psll.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psll.w.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
@@ -4813,13 +6130,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psll_w_256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsllw %xmm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpsllw %xmm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vpsllw %xmm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsllw %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xf1,0xd1]
+; CHECK-NEXT: vpsllw %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xf1,0xd9]
+; CHECK-NEXT: vpsllw %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xf1,0xc1]
+; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psll.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psll.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psll.w.256(<16 x i16> %x0, <8 x i16> %x1, <16 x i16> %x2, i16 -1)
@@ -4833,13 +6150,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psll_wi_128(<8 x i16> %x0, i32 %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_wi_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpsllw $3, %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpsllw $3, %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpsllw $3, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsllw $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x09,0x71,0xf0,0x03]
+; CHECK-NEXT: vpsllw $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0x89,0x71,0xf0,0x03]
+; CHECK-NEXT: vpsllw $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x71,0xf0,0x03]
+; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xca]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psll.wi.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psll.wi.128(<8 x i16> %x0, i32 3, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psll.wi.128(<8 x i16> %x0, i32 3, <8 x i16> %x2, i8 -1)
@@ -4853,13 +6170,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psll_wi_256(<16 x i16> %x0, i32 %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_wi_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpsllw $3, %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vpsllw $3, %ymm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpsllw $3, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpsllw $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x29,0x71,0xf0,0x03]
+; CHECK-NEXT: vpsllw $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xa9,0x71,0xf0,0x03]
+; CHECK-NEXT: vpsllw $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x71,0xf0,0x03]
+; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xca]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psll.wi.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psll.wi.256(<16 x i16> %x0, i32 3, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psll.wi.256(<16 x i16> %x0, i32 3, <16 x i16> %x2, i16 -1)
@@ -4873,13 +6190,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_psllv16_hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psllv16_hi:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsllvw %ymm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpsllvw %ymm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vpsllvw %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsllvw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x12,0xd1]
+; CHECK-NEXT: vpsllvw %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x12,0xd9]
+; CHECK-NEXT: vpsllvw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x12,0xc1]
+; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.psllv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.psllv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.psllv16.hi(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
@@ -4893,13 +6210,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_psllv8_hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psllv8_hi:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpsllvw %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpsllvw %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vpsllvw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpsllvw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x12,0xd1]
+; CHECK-NEXT: vpsllvw %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x12,0xd9]
+; CHECK-NEXT: vpsllvw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x12,0xc1]
+; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.psllv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.psllv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.psllv8.hi(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
@@ -4913,12 +6230,12 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_loadu_w_128(i8* %ptr, i8* %ptr2, <8 x i16> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_loadu_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edx, %k1
-; CHECK-NEXT: vmovdqu16 (%rdi), %xmm0
-; CHECK-NEXT: vmovdqu16 (%rsi), %xmm0 {%k1}
-; CHECK-NEXT: vmovdqu16 (%rdi), %xmm1 {%k1} {z}
-; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
+; CHECK-NEXT: vmovdqu16 (%rdi), %xmm0 ## encoding: [0x62,0xf1,0xff,0x08,0x6f,0x07]
+; CHECK-NEXT: vmovdqu16 (%rsi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0xff,0x09,0x6f,0x06]
+; CHECK-NEXT: vmovdqu16 (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0xff,0x89,0x6f,0x0f]
+; CHECK-NEXT: vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <8 x i16> @llvm.x86.avx512.mask.loadu.w.128(i8* %ptr, <8 x i16> %x1, i8 -1)
%res = call <8 x i16> @llvm.x86.avx512.mask.loadu.w.128(i8* %ptr2, <8 x i16> %res0, i8 %mask)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.loadu.w.128(i8* %ptr, <8 x i16> zeroinitializer, i8 %mask)
@@ -4931,12 +6248,12 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_loadu_w_256(i8* %ptr, i8* %ptr2, <16 x i16> %x1, i16 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_loadu_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edx, %k1
-; CHECK-NEXT: vmovdqu16 (%rdi), %ymm0
-; CHECK-NEXT: vmovdqu16 (%rsi), %ymm0 {%k1}
-; CHECK-NEXT: vmovdqu16 (%rdi), %ymm1 {%k1} {z}
-; CHECK-NEXT: vpaddw %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
+; CHECK-NEXT: vmovdqu16 (%rdi), %ymm0 ## encoding: [0x62,0xf1,0xff,0x28,0x6f,0x07]
+; CHECK-NEXT: vmovdqu16 (%rsi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0xff,0x29,0x6f,0x06]
+; CHECK-NEXT: vmovdqu16 (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0xff,0xa9,0x6f,0x0f]
+; CHECK-NEXT: vpaddw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfd,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <16 x i16> @llvm.x86.avx512.mask.loadu.w.256(i8* %ptr, <16 x i16> %x1, i16 -1)
%res = call <16 x i16> @llvm.x86.avx512.mask.loadu.w.256(i8* %ptr2, <16 x i16> %res0, i16 %mask)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.loadu.w.256(i8* %ptr, <16 x i16> zeroinitializer, i16 %mask)
@@ -4949,12 +6266,12 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_loadu_b_128(i8* %ptr, i8* %ptr2, <16 x i8> %x1, i16 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_loadu_b_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edx, %k1
-; CHECK-NEXT: vmovdqu8 (%rdi), %xmm0
-; CHECK-NEXT: vmovdqu8 (%rsi), %xmm0 {%k1}
-; CHECK-NEXT: vmovdqu8 (%rdi), %xmm1 {%k1} {z}
-; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
+; CHECK-NEXT: vmovdqu8 (%rdi), %xmm0 ## encoding: [0x62,0xf1,0x7f,0x08,0x6f,0x07]
+; CHECK-NEXT: vmovdqu8 (%rsi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0x7f,0x09,0x6f,0x06]
+; CHECK-NEXT: vmovdqu8 (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7f,0x89,0x6f,0x0f]
+; CHECK-NEXT: vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfc,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <16 x i8> @llvm.x86.avx512.mask.loadu.b.128(i8* %ptr, <16 x i8> %x1, i16 -1)
%res = call <16 x i8> @llvm.x86.avx512.mask.loadu.b.128(i8* %ptr2, <16 x i8> %res0, i16 %mask)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.loadu.b.128(i8* %ptr, <16 x i8> zeroinitializer, i16 %mask)
@@ -4967,12 +6284,12 @@ declare <32 x i8> @llvm.x86.avx512.mask.
define <32 x i8>@test_int_x86_avx512_mask_loadu_b_256(i8* %ptr, i8* %ptr2, <32 x i8> %x1, i32 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_loadu_b_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovd %edx, %k1
-; CHECK-NEXT: vmovdqu8 (%rdi), %ymm0
-; CHECK-NEXT: vmovdqu8 (%rsi), %ymm0 {%k1}
-; CHECK-NEXT: vmovdqu8 (%rdi), %ymm1 {%k1} {z}
-; CHECK-NEXT: vpaddb %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovd %edx, %k1 ## encoding: [0xc5,0xfb,0x92,0xca]
+; CHECK-NEXT: vmovdqu8 (%rdi), %ymm0 ## encoding: [0x62,0xf1,0x7f,0x28,0x6f,0x07]
+; CHECK-NEXT: vmovdqu8 (%rsi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0x7f,0x29,0x6f,0x06]
+; CHECK-NEXT: vmovdqu8 (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7f,0xa9,0x6f,0x0f]
+; CHECK-NEXT: vpaddb %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfc,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res0 = call <32 x i8> @llvm.x86.avx512.mask.loadu.b.256(i8* %ptr, <32 x i8> %x1, i32 -1)
%res = call <32 x i8> @llvm.x86.avx512.mask.loadu.b.256(i8* %ptr2, <32 x i8> %res0, i32 %mask)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.loadu.b.256(i8* %ptr, <32 x i8> zeroinitializer, i32 %mask)
@@ -4985,16 +6302,16 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_pmovzxb_w_128(<16 x i8> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxb_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmovzxbw %xmm0, %xmm1 {%k1}
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovzxbw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x30,0xc8]
; CHECK-NEXT: ## xmm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; CHECK-NEXT: vpmovzxbw %xmm0, %xmm2 {%k1} {z}
+; CHECK-NEXT: vpmovzxbw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x30,0xd0]
; CHECK-NEXT: ## xmm2 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; CHECK-NEXT: vpmovzxbw %xmm0, %xmm0
+; CHECK-NEXT: vpmovzxbw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x30,0xc0]
; CHECK-NEXT: ## xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
-; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xca]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmovzxb.w.128(<16 x i8> %x0, <8 x i16> %x1, i8 %x2)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovzxb.w.128(<16 x i8> %x0, <8 x i16> zeroinitializer, i8 %x2)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovzxb.w.128(<16 x i8> %x0, <8 x i16> %x1, i8 -1)
@@ -5008,16 +6325,16 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_pmovzxb_w_256(<16 x i8> %x0, <16 x i16> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxb_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmovzxbw %xmm0, %ymm1 {%k1}
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovzxbw %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x30,0xc8]
; CHECK-NEXT: ## ymm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
-; CHECK-NEXT: vpmovzxbw %xmm0, %ymm2 {%k1} {z}
+; CHECK-NEXT: vpmovzxbw %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x30,0xd0]
; CHECK-NEXT: ## ymm2 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
-; CHECK-NEXT: vpmovzxbw %xmm0, %ymm0
+; CHECK-NEXT: vpmovzxbw %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x30,0xc0]
; CHECK-NEXT: ## ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
-; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xca]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmovzxb.w.256(<16 x i8> %x0, <16 x i16> %x1, i16 %x2)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmovzxb.w.256(<16 x i8> %x0, <16 x i16> zeroinitializer, i16 %x2)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.pmovzxb.w.256(<16 x i8> %x0, <16 x i16> %x1, i16 -1)
@@ -5032,13 +6349,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_pmovsxb_w_128(<16 x i8> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxb_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmovsxbw %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpmovsxbw %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpmovsxbw %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovsxbw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x20,0xc8]
+; CHECK-NEXT: vpmovsxbw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x20,0xd0]
+; CHECK-NEXT: vpmovsxbw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x20,0xc0]
+; CHECK-NEXT: vpaddw %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xca]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pmovsxb.w.128(<16 x i8> %x0, <8 x i16> %x1, i8 %x2)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovsxb.w.128(<16 x i8> %x0, <8 x i16> zeroinitializer, i8 %x2)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovsxb.w.128(<16 x i8> %x0, <8 x i16> %x1, i8 -1)
@@ -5052,13 +6369,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_pmovsxb_w_256(<16 x i8> %x0, <16 x i16> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxb_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmovsxbw %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vpmovsxbw %xmm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpmovsxbw %xmm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovsxbw %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x20,0xc8]
+; CHECK-NEXT: vpmovsxbw %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x20,0xd0]
+; CHECK-NEXT: vpmovsxbw %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x20,0xc0]
+; CHECK-NEXT: vpaddw %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xca]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pmovsxb.w.256(<16 x i8> %x0, <16 x i16> %x1, i16 %x2)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pmovsxb.w.256(<16 x i8> %x0, <16 x i16> zeroinitializer, i16 %x2)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.pmovsxb.w.256(<16 x i8> %x0, <16 x i16> %x1, i16 -1)
@@ -5072,13 +6389,13 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_pmovsxd_q_128(<4 x i32> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxd_q_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmovsxdq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vpmovsxdq %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vpmovsxdq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm2, %xmm1, %xmm1
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovsxdq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x25,0xc8]
+; CHECK-NEXT: vpmovsxdq %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x25,0xd0]
+; CHECK-NEXT: vpmovsxdq %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x25,0xc0]
+; CHECK-NEXT: vpaddq %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xca]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.pmovsxd.q.128(<4 x i32> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.pmovsxd.q.128(<4 x i32> %x0, <2 x i64> zeroinitializer, i8 %x2)
%res2 = call <2 x i64> @llvm.x86.avx512.mask.pmovsxd.q.128(<4 x i32> %x0, <2 x i64> %x1, i8 -1)
@@ -5092,13 +6409,13 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_pmovsxd_q_256(<4 x i32> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxd_q_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpmovsxdq %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vpmovsxdq %xmm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vpmovsxdq %xmm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpmovsxdq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x25,0xc8]
+; CHECK-NEXT: vpmovsxdq %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x25,0xd0]
+; CHECK-NEXT: vpmovsxdq %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x25,0xc0]
+; CHECK-NEXT: vpaddq %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xca]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.pmovsxd.q.256(<4 x i32> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.pmovsxd.q.256(<4 x i32> %x0, <4 x i64> zeroinitializer, i8 %x2)
%res2 = call <4 x i64> @llvm.x86.avx512.mask.pmovsxd.q.256(<4 x i32> %x0, <4 x i64> %x1, i8 -1)
@@ -5112,13 +6429,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_permvar_hi_128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_permvar_hi_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpermw %xmm1, %xmm0, %xmm2 {%k1}
-; CHECK-NEXT: vpermw %xmm1, %xmm0, %xmm3 {%k1} {z}
-; CHECK-NEXT: vpermw %xmm1, %xmm0, %xmm0
-; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpermw %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x8d,0xd1]
+; CHECK-NEXT: vpermw %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x8d,0xd9]
+; CHECK-NEXT: vpermw %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x8d,0xc1]
+; CHECK-NEXT: vpaddw %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.permvar.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 %x3)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.permvar.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> zeroinitializer, i8 %x3)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.permvar.hi.128(<8 x i16> %x0, <8 x i16> %x1, <8 x i16> %x2, i8 -1)
@@ -5132,13 +6449,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_permvar_hi_256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_permvar_hi_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vpermw %ymm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vpermw %ymm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vpermw %ymm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vpermw %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x8d,0xd1]
+; CHECK-NEXT: vpermw %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x8d,0xd9]
+; CHECK-NEXT: vpermw %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x8d,0xc1]
+; CHECK-NEXT: vpaddw %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xcb]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.permvar.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 %x3)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.permvar.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> zeroinitializer, i16 %x3)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.permvar.hi.256(<16 x i16> %x0, <16 x i16> %x1, <16 x i16> %x2, i16 -1)
@@ -5152,10 +6469,10 @@ declare void @llvm.x86.avx512.mask.store
define void at test_int_x86_avx512_mask_storeu_b_128(i8* %ptr1, i8* %ptr2, <16 x i8> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_b_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edx, %k1
-; CHECK-NEXT: vmovdqu8 %xmm0, (%rdi) {%k1}
-; CHECK-NEXT: vmovdqu8 %xmm0, (%rsi)
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
+; CHECK-NEXT: vmovdqu8 %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7f,0x09,0x7f,0x07]
+; CHECK-NEXT: vmovdqu8 %xmm0, (%rsi) ## encoding: [0x62,0xf1,0x7f,0x08,0x7f,0x06]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.storeu.b.128(i8* %ptr1, <16 x i8> %x1, i16 %x2)
call void @llvm.x86.avx512.mask.storeu.b.128(i8* %ptr2, <16 x i8> %x1, i16 -1)
ret void
@@ -5166,10 +6483,10 @@ declare void @llvm.x86.avx512.mask.store
define void at test_int_x86_avx512_mask_storeu_b_256(i8* %ptr1, i8* %ptr2, <32 x i8> %x1, i32 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_b_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovd %edx, %k1
-; CHECK-NEXT: vmovdqu8 %ymm0, (%rdi) {%k1}
-; CHECK-NEXT: vmovdqu8 %ymm0, (%rsi)
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovd %edx, %k1 ## encoding: [0xc5,0xfb,0x92,0xca]
+; CHECK-NEXT: vmovdqu8 %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7f,0x29,0x7f,0x07]
+; CHECK-NEXT: vmovdqu8 %ymm0, (%rsi) ## encoding: [0x62,0xf1,0x7f,0x28,0x7f,0x06]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.storeu.b.256(i8* %ptr1, <32 x i8> %x1, i32 %x2)
call void @llvm.x86.avx512.mask.storeu.b.256(i8* %ptr2, <32 x i8> %x1, i32 -1)
ret void
@@ -5180,10 +6497,10 @@ declare void @llvm.x86.avx512.mask.store
define void at test_int_x86_avx512_mask_storeu_w_128(i8* %ptr1, i8* %ptr2, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edx, %k1
-; CHECK-NEXT: vmovdqu16 %xmm0, (%rdi) {%k1}
-; CHECK-NEXT: vmovdqu16 %xmm0, (%rsi)
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
+; CHECK-NEXT: vmovdqu16 %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xff,0x09,0x7f,0x07]
+; CHECK-NEXT: vmovdqu16 %xmm0, (%rsi) ## encoding: [0x62,0xf1,0xff,0x08,0x7f,0x06]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.storeu.w.128(i8* %ptr1, <8 x i16> %x1, i8 %x2)
call void @llvm.x86.avx512.mask.storeu.w.128(i8* %ptr2, <8 x i16> %x1, i8 -1)
ret void
@@ -5194,10 +6511,10 @@ declare void @llvm.x86.avx512.mask.store
define void at test_int_x86_avx512_mask_storeu_w_256(i8* %ptr1, i8* %ptr2, <16 x i16> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edx, %k1
-; CHECK-NEXT: vmovdqu16 %ymm0, (%rdi) {%k1}
-; CHECK-NEXT: vmovdqu16 %ymm0, (%rsi)
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
+; CHECK-NEXT: vmovdqu16 %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xff,0x29,0x7f,0x07]
+; CHECK-NEXT: vmovdqu16 %ymm0, (%rsi) ## encoding: [0x62,0xf1,0xff,0x28,0x7f,0x06]
+; CHECK-NEXT: retq ## encoding: [0xc3]
call void @llvm.x86.avx512.mask.storeu.w.256(i8* %ptr1, <16 x i16> %x1, i16 %x2)
call void @llvm.x86.avx512.mask.storeu.w.256(i8* %ptr2, <16 x i16> %x1, i16 -1)
ret void
@@ -5208,11 +6525,11 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_movu_w_128(<8 x i16> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movu_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovdqu16 %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vmovdqu16 %xmm0, %xmm0 {%k1} {z}
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovdqu16 %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xff,0x09,0x6f,0xc8]
+; CHECK-NEXT: vmovdqu16 %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0xff,0x89,0x6f,0xc0]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.movu.w.128(<8 x i16> %x0, <8 x i16> %x1, i8 %x2)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.movu.w.128(<8 x i16> %x0, <8 x i16> zeroinitializer, i8 %x2)
%res2 = add <8 x i16> %res, %res1
@@ -5224,11 +6541,11 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_movu_w_256(<16 x i16> %x0, <16 x i16> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movu_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovdqu16 %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vmovdqu16 %ymm0, %ymm0 {%k1} {z}
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovdqu16 %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xff,0x29,0x6f,0xc8]
+; CHECK-NEXT: vmovdqu16 %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xff,0xa9,0x6f,0xc0]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.movu.w.256(<16 x i16> %x0, <16 x i16> %x1, i16 %x2)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.movu.w.256(<16 x i16> %x0, <16 x i16> zeroinitializer, i16 %x2)
%res2 = add <16 x i16> %res, %res1
@@ -5240,11 +6557,11 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_movu_b_128(<16 x i8> %x0, <16 x i8> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movu_b_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vmovdqu8 %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vmovdqu8 %xmm0, %xmm0 {%k1} {z}
-; CHECK-NEXT: vpaddb %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vmovdqu8 %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7f,0x09,0x6f,0xc8]
+; CHECK-NEXT: vmovdqu8 %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7f,0x89,0x6f,0xc0]
+; CHECK-NEXT: vpaddb %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.movu.b.128(<16 x i8> %x0, <16 x i8> %x1, i16 %x2)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.movu.b.128(<16 x i8> %x0, <16 x i8> zeroinitializer, i16 %x2)
%res2 = add <16 x i8> %res, %res1
@@ -5256,11 +6573,11 @@ declare <32 x i8> @llvm.x86.avx512.mask.
define <32 x i8>@test_int_x86_avx512_mask_movu_b_256(<32 x i8> %x0, <32 x i8> %x1, i32 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movu_b_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovd %edi, %k1
-; CHECK-NEXT: vmovdqu8 %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vmovdqu8 %ymm0, %ymm0 {%k1} {z}
-; CHECK-NEXT: vpaddb %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vmovdqu8 %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7f,0x29,0x6f,0xc8]
+; CHECK-NEXT: vmovdqu8 %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7f,0xa9,0x6f,0xc0]
+; CHECK-NEXT: vpaddb %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.movu.b.256(<32 x i8> %x0, <32 x i8> %x1, i32 %x2)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.movu.b.256(<32 x i8> %x0, <32 x i8> zeroinitializer, i32 %x2)
%res2 = add <32 x i8> %res, %res1
@@ -5344,13 +6661,13 @@ declare i16 @llvm.x86.avx512.ptestnm.b.1
define i16 at test_int_x86_avx512_ptestnm_b_128(<16 x i8> %x0, <16 x i8> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestnm_b_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vptestnmb %xmm1, %xmm0, %k0 {%k1}
-; CHECK-NEXT: kmovw %k0, %ecx
-; CHECK-NEXT: vptestnmb %xmm1, %xmm0, %k0
-; CHECK-NEXT: kmovw %k0, %eax
-; CHECK-NEXT: addl %ecx, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vptestnmb %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x26,0xc1]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vptestnmb %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf2,0x7e,0x08,0x26,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: addl %ecx, %eax ## encoding: [0x01,0xc8]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.ptestnm.b.128(<16 x i8> %x0, <16 x i8> %x1, i16 %x2)
%res1 = call i16 @llvm.x86.avx512.ptestnm.b.128(<16 x i8> %x0, <16 x i8> %x1, i16-1)
%res2 = add i16 %res, %res1
@@ -5362,13 +6679,13 @@ declare i32 @llvm.x86.avx512.ptestnm.b.2
define i32 at test_int_x86_avx512_ptestnm_b_256(<32 x i8> %x0, <32 x i8> %x1, i32 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestnm_b_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovd %edi, %k1
-; CHECK-NEXT: vptestnmb %ymm1, %ymm0, %k0 {%k1}
-; CHECK-NEXT: kmovd %k0, %ecx
-; CHECK-NEXT: vptestnmb %ymm1, %ymm0, %k0
-; CHECK-NEXT: kmovd %k0, %eax
-; CHECK-NEXT: addl %ecx, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovd %edi, %k1 ## encoding: [0xc5,0xfb,0x92,0xcf]
+; CHECK-NEXT: vptestnmb %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x26,0xc1]
+; CHECK-NEXT: kmovd %k0, %ecx ## encoding: [0xc5,0xfb,0x93,0xc8]
+; CHECK-NEXT: vptestnmb %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf2,0x7e,0x28,0x26,0xc1]
+; CHECK-NEXT: kmovd %k0, %eax ## encoding: [0xc5,0xfb,0x93,0xc0]
+; CHECK-NEXT: addl %ecx, %eax ## encoding: [0x01,0xc8]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i32 @llvm.x86.avx512.ptestnm.b.256(<32 x i8> %x0, <32 x i8> %x1, i32 %x2)
%res1 = call i32 @llvm.x86.avx512.ptestnm.b.256(<32 x i8> %x0, <32 x i8> %x1, i32-1)
%res2 = add i32 %res, %res1
@@ -5380,13 +6697,13 @@ declare i8 @llvm.x86.avx512.ptestnm.w.12
define i8 at test_int_x86_avx512_ptestnm_w_128(<8 x i16> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestnm_w_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vptestnmw %xmm1, %xmm0, %k0 {%k1}
-; CHECK-NEXT: kmovw %k0, %ecx
-; CHECK-NEXT: vptestnmw %xmm1, %xmm0, %k0
-; CHECK-NEXT: kmovw %k0, %eax
-; CHECK-NEXT: addb %cl, %al
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vptestnmw %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfe,0x09,0x26,0xc1]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vptestnmw %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x08,0x26,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: addb %cl, %al ## encoding: [0x00,0xc8]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.ptestnm.w.128(<8 x i16> %x0, <8 x i16> %x1, i8 %x2)
%res1 = call i8 @llvm.x86.avx512.ptestnm.w.128(<8 x i16> %x0, <8 x i16> %x1, i8-1)
%res2 = add i8 %res, %res1
@@ -5398,13 +6715,13 @@ declare i16 @llvm.x86.avx512.ptestnm.w.2
define i16 at test_int_x86_avx512_ptestnm_w_256(<16 x i16> %x0, <16 x i16> %x1, i16 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestnm_w_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %edi, %k1
-; CHECK-NEXT: vptestnmw %ymm1, %ymm0, %k0 {%k1}
-; CHECK-NEXT: kmovw %k0, %ecx
-; CHECK-NEXT: vptestnmw %ymm1, %ymm0, %k0
-; CHECK-NEXT: kmovw %k0, %eax
-; CHECK-NEXT: addl %ecx, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vptestnmw %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfe,0x29,0x26,0xc1]
+; CHECK-NEXT: kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
+; CHECK-NEXT: vptestnmw %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x28,0x26,0xc1]
+; CHECK-NEXT: kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
+; CHECK-NEXT: addl %ecx, %eax ## encoding: [0x01,0xc8]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i16 @llvm.x86.avx512.ptestnm.w.256(<16 x i16> %x0, <16 x i16> %x1, i16 %x2)
%res1 = call i16 @llvm.x86.avx512.ptestnm.w.256(<16 x i16> %x0, <16 x i16> %x1, i16-1)
%res2 = add i16 %res, %res1
@@ -5416,13 +6733,13 @@ declare <32 x i8> @llvm.x86.avx512.mask.
define <32 x i8>@test_int_x86_avx512_mask_pbroadcast_b_gpr_256(i8 %x0, <32 x i8> %x1, i32 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pbroadcast_b_gpr_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovd %esi, %k1
-; CHECK-NEXT: vpbroadcastb %dil, %ymm0 {%k1}
-; CHECK-NEXT: vpbroadcastb %dil, %ymm1 {%k1} {z}
-; CHECK-NEXT: vpbroadcastb %dil, %ymm2
-; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: vpaddb %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovd %esi, %k1 ## encoding: [0xc5,0xfb,0x92,0xce]
+; CHECK-NEXT: vpbroadcastb %dil, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x7a,0xc7]
+; CHECK-NEXT: vpbroadcastb %dil, %ymm1 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x7a,0xcf]
+; CHECK-NEXT: vpbroadcastb %dil, %ymm2 ## encoding: [0x62,0xf2,0x7d,0x28,0x7a,0xd7]
+; CHECK-NEXT: vpaddb %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfc,0xc0]
+; CHECK-NEXT: vpaddb %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <32 x i8> @llvm.x86.avx512.mask.pbroadcast.b.gpr.256(i8 %x0, <32 x i8> %x1, i32 -1)
%res1 = call <32 x i8> @llvm.x86.avx512.mask.pbroadcast.b.gpr.256(i8 %x0, <32 x i8> %x1, i32 %mask)
%res2 = call <32 x i8> @llvm.x86.avx512.mask.pbroadcast.b.gpr.256(i8 %x0, <32 x i8> zeroinitializer, i32 %mask)
@@ -5436,13 +6753,13 @@ declare <16 x i8> @llvm.x86.avx512.mask.
define <16 x i8>@test_int_x86_avx512_mask_pbroadcast_b_gpr_128(i8 %x0, <16 x i8> %x1, i16 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pbroadcast_b_gpr_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpbroadcastb %dil, %xmm0 {%k1}
-; CHECK-NEXT: vpbroadcastb %dil, %xmm1 {%k1} {z}
-; CHECK-NEXT: vpbroadcastb %dil, %xmm2
-; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0
-; CHECK-NEXT: vpaddb %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpbroadcastb %dil, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x7a,0xc7]
+; CHECK-NEXT: vpbroadcastb %dil, %xmm1 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x7a,0xcf]
+; CHECK-NEXT: vpbroadcastb %dil, %xmm2 ## encoding: [0x62,0xf2,0x7d,0x08,0x7a,0xd7]
+; CHECK-NEXT: vpaddb %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfc,0xc0]
+; CHECK-NEXT: vpaddb %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfc,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i8> @llvm.x86.avx512.mask.pbroadcast.b.gpr.128(i8 %x0, <16 x i8> %x1, i16 -1)
%res1 = call <16 x i8> @llvm.x86.avx512.mask.pbroadcast.b.gpr.128(i8 %x0, <16 x i8> %x1, i16 %mask)
%res2 = call <16 x i8> @llvm.x86.avx512.mask.pbroadcast.b.gpr.128(i8 %x0, <16 x i8> zeroinitializer, i16 %mask)
@@ -5456,13 +6773,13 @@ declare <16 x i16> @llvm.x86.avx512.mask
define <16 x i16>@test_int_x86_avx512_mask_pbroadcast_w_gpr_256(i16 %x0, <16 x i16> %x1, i16 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pbroadcast_w_gpr_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpbroadcastw %di, %ymm0 {%k1}
-; CHECK-NEXT: vpbroadcastw %di, %ymm1 {%k1} {z}
-; CHECK-NEXT: vpbroadcastw %di, %ymm2
-; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpbroadcastw %di, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x7b,0xc7]
+; CHECK-NEXT: vpbroadcastw %di, %ymm1 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x7b,0xcf]
+; CHECK-NEXT: vpbroadcastw %di, %ymm2 ## encoding: [0x62,0xf2,0x7d,0x28,0x7b,0xd7]
+; CHECK-NEXT: vpaddw %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfd,0xc0]
+; CHECK-NEXT: vpaddw %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x i16> @llvm.x86.avx512.mask.pbroadcast.w.gpr.256(i16 %x0, <16 x i16> %x1, i16 -1)
%res1 = call <16 x i16> @llvm.x86.avx512.mask.pbroadcast.w.gpr.256(i16 %x0, <16 x i16> %x1, i16 %mask)
%res2 = call <16 x i16> @llvm.x86.avx512.mask.pbroadcast.w.gpr.256(i16 %x0, <16 x i16> zeroinitializer, i16 %mask)
@@ -5476,13 +6793,13 @@ declare <8 x i16> @llvm.x86.avx512.mask.
define <8 x i16>@test_int_x86_avx512_mask_pbroadcast_w_gpr_128(i16 %x0, <8 x i16> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pbroadcast_w_gpr_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovw %esi, %k1
-; CHECK-NEXT: vpbroadcastw %di, %xmm0 {%k1}
-; CHECK-NEXT: vpbroadcastw %di, %xmm1 {%k1} {z}
-; CHECK-NEXT: vpbroadcastw %di, %xmm2
-; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0
-; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vpbroadcastw %di, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x7b,0xc7]
+; CHECK-NEXT: vpbroadcastw %di, %xmm1 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x7b,0xcf]
+; CHECK-NEXT: vpbroadcastw %di, %xmm2 ## encoding: [0x62,0xf2,0x7d,0x08,0x7b,0xd7]
+; CHECK-NEXT: vpaddw %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfd,0xc0]
+; CHECK-NEXT: vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfd,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i16> @llvm.x86.avx512.mask.pbroadcast.w.gpr.128(i16 %x0, <8 x i16> %x1, i8 -1)
%res1 = call <8 x i16> @llvm.x86.avx512.mask.pbroadcast.w.gpr.128(i16 %x0, <8 x i16> %x1, i8 %mask)
%res2 = call <8 x i16> @llvm.x86.avx512.mask.pbroadcast.w.gpr.128(i16 %x0, <8 x i16> zeroinitializer, i8 %mask)
Modified: llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics.ll?rev=270315&r1=270314&r2=270315&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics.ll (original)
+++ llvm/trunk/test/CodeGen/X86/avx512dqvl-intrinsics.ll Sat May 21 00:46:58 2016
@@ -1,53 +1,74 @@
+; NOTE: Assertions have been autogenerated by update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl -mattr=+avx512dq -mattr=+avx512vl --show-mc-encoding| FileCheck %s
define <8 x i64> @test_mask_mullo_epi64_rr_512(<8 x i64> %a, <8 x i64> %b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rr_512
- ;CHECK: vpmullq %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf2,0xfd,0x48,0x40,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi64_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf2,0xfd,0x48,0x40,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.pmull.q.512(<8 x i64> %a, <8 x i64> %b, <8 x i64> zeroinitializer, i8 -1)
ret <8 x i64> %res
}
define <8 x i64> @test_mask_mullo_epi64_rrk_512(<8 x i64> %a, <8 x i64> %b, <8 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rrk_512
- ;CHECK: vpmullq %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0x40,0xd1]
+; CHECK-LABEL: test_mask_mullo_epi64_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vpmullq %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0x40,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.pmull.q.512(<8 x i64> %a, <8 x i64> %b, <8 x i64> %passThru, i8 %mask)
ret <8 x i64> %res
}
define <8 x i64> @test_mask_mullo_epi64_rrkz_512(<8 x i64> %a, <8 x i64> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rrkz_512
- ;CHECK: vpmullq %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xc9,0x40,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi64_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vpmullq %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xc9,0x40,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i64> @llvm.x86.avx512.mask.pmull.q.512(<8 x i64> %a, <8 x i64> %b, <8 x i64> zeroinitializer, i8 %mask)
ret <8 x i64> %res
}
define <8 x i64> @test_mask_mullo_epi64_rm_512(<8 x i64> %a, <8 x i64>* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rm_512
- ;CHECK: vpmullq (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf2,0xfd,0x48,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf2,0xfd,0x48,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i64>, <8 x i64>* %ptr_b
%res = call <8 x i64> @llvm.x86.avx512.mask.pmull.q.512(<8 x i64> %a, <8 x i64> %b, <8 x i64> zeroinitializer, i8 -1)
ret <8 x i64> %res
}
define <8 x i64> @test_mask_mullo_epi64_rmk_512(<8 x i64> %a, <8 x i64>* %ptr_b, <8 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmk_512
- ;CHECK: vpmullq (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0x40,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi64_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0x40,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i64>, <8 x i64>* %ptr_b
%res = call <8 x i64> @llvm.x86.avx512.mask.pmull.q.512(<8 x i64> %a, <8 x i64> %b, <8 x i64> %passThru, i8 %mask)
ret <8 x i64> %res
}
define <8 x i64> @test_mask_mullo_epi64_rmkz_512(<8 x i64> %a, <8 x i64>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmkz_512
- ;CHECK: vpmullq (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xc9,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xc9,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x i64>, <8 x i64>* %ptr_b
%res = call <8 x i64> @llvm.x86.avx512.mask.pmull.q.512(<8 x i64> %a, <8 x i64> %b, <8 x i64> zeroinitializer, i8 %mask)
ret <8 x i64> %res
}
define <8 x i64> @test_mask_mullo_epi64_rmb_512(<8 x i64> %a, i64* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmb_512
- ;CHECK: vpmullq (%rdi){1to8}, %zmm0, %zmm0 ## encoding: [0x62,0xf2,0xfd,0x58,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq (%rdi){1to8}, %zmm0, %zmm0 ## encoding: [0x62,0xf2,0xfd,0x58,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <8 x i64> undef, i64 %q, i32 0
%b = shufflevector <8 x i64> %vecinit.i, <8 x i64> undef, <8 x i32> zeroinitializer
@@ -56,8 +77,12 @@ define <8 x i64> @test_mask_mullo_epi64_
}
define <8 x i64> @test_mask_mullo_epi64_rmbk_512(<8 x i64> %a, i64* %ptr_b, <8 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmbk_512
- ;CHECK: vpmullq (%rdi){1to8}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x59,0x40,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi64_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi){1to8}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x59,0x40,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <8 x i64> undef, i64 %q, i32 0
%b = shufflevector <8 x i64> %vecinit.i, <8 x i64> undef, <8 x i32> zeroinitializer
@@ -66,8 +91,11 @@ define <8 x i64> @test_mask_mullo_epi64_
}
define <8 x i64> @test_mask_mullo_epi64_rmbkz_512(<8 x i64> %a, i64* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmbkz_512
- ;CHECK: vpmullq (%rdi){1to8}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xd9,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi){1to8}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xd9,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <8 x i64> undef, i64 %q, i32 0
%b = shufflevector <8 x i64> %vecinit.i, <8 x i64> undef, <8 x i32> zeroinitializer
@@ -77,53 +105,73 @@ define <8 x i64> @test_mask_mullo_epi64_
declare <8 x i64> @llvm.x86.avx512.mask.pmull.q.512(<8 x i64>, <8 x i64>, <8 x i64>, i8)
define <4 x i64> @test_mask_mullo_epi64_rr_256(<4 x i64> %a, <4 x i64> %b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rr_256
- ;CHECK: vpmullq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x40,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi64_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x40,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.pmull.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 -1)
ret <4 x i64> %res
}
define <4 x i64> @test_mask_mullo_epi64_rrk_256(<4 x i64> %a, <4 x i64> %b, <4 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rrk_256
- ;CHECK: vpmullq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x40,0xd1]
+; CHECK-LABEL: test_mask_mullo_epi64_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vpmullq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x40,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.pmull.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> %passThru, i8 %mask)
ret <4 x i64> %res
}
define <4 x i64> @test_mask_mullo_epi64_rrkz_256(<4 x i64> %a, <4 x i64> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rrkz_256
- ;CHECK: vpmullq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x40,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi64_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vpmullq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x40,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.pmull.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 %mask)
ret <4 x i64> %res
}
define <4 x i64> @test_mask_mullo_epi64_rm_256(<4 x i64> %a, <4 x i64>* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rm_256
- ;CHECK: vpmullq (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i64>, <4 x i64>* %ptr_b
%res = call <4 x i64> @llvm.x86.avx512.mask.pmull.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 -1)
ret <4 x i64> %res
}
define <4 x i64> @test_mask_mullo_epi64_rmk_256(<4 x i64> %a, <4 x i64>* %ptr_b, <4 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmk_256
- ;CHECK: vpmullq (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x40,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi64_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x40,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i64>, <4 x i64>* %ptr_b
%res = call <4 x i64> @llvm.x86.avx512.mask.pmull.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> %passThru, i8 %mask)
ret <4 x i64> %res
}
define <4 x i64> @test_mask_mullo_epi64_rmkz_256(<4 x i64> %a, <4 x i64>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmkz_256
- ;CHECK: vpmullq (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x i64>, <4 x i64>* %ptr_b
%res = call <4 x i64> @llvm.x86.avx512.mask.pmull.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 %mask)
ret <4 x i64> %res
}
define <4 x i64> @test_mask_mullo_epi64_rmb_256(<4 x i64> %a, i64* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmb_256
- ;CHECK: vpmullq (%rdi){1to4}, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x38,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq (%rdi){1to4}, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x38,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <4 x i64> undef, i64 %q, i32 0
%b = shufflevector <4 x i64> %vecinit.i, <4 x i64> undef, <4 x i32> zeroinitializer
@@ -132,8 +180,12 @@ define <4 x i64> @test_mask_mullo_epi64_
}
define <4 x i64> @test_mask_mullo_epi64_rmbk_256(<4 x i64> %a, i64* %ptr_b, <4 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmbk_256
- ;CHECK: vpmullq (%rdi){1to4}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x39,0x40,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi64_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi){1to4}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x39,0x40,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <4 x i64> undef, i64 %q, i32 0
%b = shufflevector <4 x i64> %vecinit.i, <4 x i64> undef, <4 x i32> zeroinitializer
@@ -142,8 +194,11 @@ define <4 x i64> @test_mask_mullo_epi64_
}
define <4 x i64> @test_mask_mullo_epi64_rmbkz_256(<4 x i64> %a, i64* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmbkz_256
- ;CHECK: vpmullq (%rdi){1to4}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xb9,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi){1to4}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xb9,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <4 x i64> undef, i64 %q, i32 0
%b = shufflevector <4 x i64> %vecinit.i, <4 x i64> undef, <4 x i32> zeroinitializer
@@ -154,53 +209,73 @@ define <4 x i64> @test_mask_mullo_epi64_
declare <4 x i64> @llvm.x86.avx512.mask.pmull.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)
define <2 x i64> @test_mask_mullo_epi64_rr_128(<2 x i64> %a, <2 x i64> %b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rr_128
- ;CHECK: vpmullq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x40,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi64_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x40,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.pmull.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 -1)
ret <2 x i64> %res
}
define <2 x i64> @test_mask_mullo_epi64_rrk_128(<2 x i64> %a, <2 x i64> %b, <2 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rrk_128
- ;CHECK: vpmullq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x40,0xd1]
+; CHECK-LABEL: test_mask_mullo_epi64_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vpmullq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x40,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.pmull.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> %passThru, i8 %mask)
ret <2 x i64> %res
}
define <2 x i64> @test_mask_mullo_epi64_rrkz_128(<2 x i64> %a, <2 x i64> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rrkz_128
- ;CHECK: vpmullq %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x40,0xc1]
+; CHECK-LABEL: test_mask_mullo_epi64_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vpmullq %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x40,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.pmull.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 %mask)
ret <2 x i64> %res
}
define <2 x i64> @test_mask_mullo_epi64_rm_128(<2 x i64> %a, <2 x i64>* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rm_128
- ;CHECK: vpmullq (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <2 x i64>, <2 x i64>* %ptr_b
%res = call <2 x i64> @llvm.x86.avx512.mask.pmull.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 -1)
ret <2 x i64> %res
}
define <2 x i64> @test_mask_mullo_epi64_rmk_128(<2 x i64> %a, <2 x i64>* %ptr_b, <2 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmk_128
- ;CHECK: vpmullq (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x40,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi64_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x40,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <2 x i64>, <2 x i64>* %ptr_b
%res = call <2 x i64> @llvm.x86.avx512.mask.pmull.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> %passThru, i8 %mask)
ret <2 x i64> %res
}
define <2 x i64> @test_mask_mullo_epi64_rmkz_128(<2 x i64> %a, <2 x i64>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmkz_128
- ;CHECK: vpmullq (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <2 x i64>, <2 x i64>* %ptr_b
%res = call <2 x i64> @llvm.x86.avx512.mask.pmull.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 %mask)
ret <2 x i64> %res
}
define <2 x i64> @test_mask_mullo_epi64_rmb_128(<2 x i64> %a, i64* %ptr_b) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmb_128
- ;CHECK: vpmullq (%rdi){1to2}, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x18,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vpmullq (%rdi){1to2}, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x18,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <2 x i64> undef, i64 %q, i32 0
%b = shufflevector <2 x i64> %vecinit.i, <2 x i64> undef, <2 x i32> zeroinitializer
@@ -209,8 +284,12 @@ define <2 x i64> @test_mask_mullo_epi64_
}
define <2 x i64> @test_mask_mullo_epi64_rmbk_128(<2 x i64> %a, i64* %ptr_b, <2 x i64> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmbk_128
- ;CHECK: vpmullq (%rdi){1to2}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0x40,0x0f]
+; CHECK-LABEL: test_mask_mullo_epi64_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi){1to2}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0x40,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <2 x i64> undef, i64 %q, i32 0
%b = shufflevector <2 x i64> %vecinit.i, <2 x i64> undef, <2 x i32> zeroinitializer
@@ -219,8 +298,11 @@ define <2 x i64> @test_mask_mullo_epi64_
}
define <2 x i64> @test_mask_mullo_epi64_rmbkz_128(<2 x i64> %a, i64* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_mullo_epi64_rmbkz_128
- ;CHECK: vpmullq (%rdi){1to2}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x99,0x40,0x07]
+; CHECK-LABEL: test_mask_mullo_epi64_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vpmullq (%rdi){1to2}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x99,0x40,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load i64, i64* %ptr_b
%vecinit.i = insertelement <2 x i64> undef, i64 %q, i32 0
%b = shufflevector <2 x i64> %vecinit.i, <2 x i64> undef, <2 x i32> zeroinitializer
@@ -231,53 +313,73 @@ define <2 x i64> @test_mask_mullo_epi64_
declare <2 x i64> @llvm.x86.avx512.mask.pmull.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)
define <4 x float> @test_mask_andnot_ps_rr_128(<4 x float> %a, <4 x float> %b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rr_128
- ;CHECK: vandnps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x55,0xc1]
+; CHECK-LABEL: test_mask_andnot_ps_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
ret <4 x float> %res
}
define <4 x float> @test_mask_andnot_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rrk_128
- ;CHECK: vandnps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x55,0xd1]
+; CHECK-LABEL: test_mask_andnot_ps_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x55,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_andnot_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rrkz_128
- ;CHECK: vandnps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x55,0xc1]
+; CHECK-LABEL: test_mask_andnot_ps_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandnps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_andnot_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rm_128
- ;CHECK: vandnps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
ret <4 x float> %res
}
define <4 x float> @test_mask_andnot_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmk_128
- ;CHECK: vandnps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x55,0x0f]
+; CHECK-LABEL: test_mask_andnot_ps_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x55,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_andnot_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmkz_128
- ;CHECK: vandnps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_andnot_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmb_128
- ;CHECK: vandnps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -286,8 +388,12 @@ define <4 x float> @test_mask_andnot_ps_
}
define <4 x float> @test_mask_andnot_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmbk_128
- ;CHECK: vandnps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x55,0x0f]
+; CHECK-LABEL: test_mask_andnot_ps_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x55,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -296,8 +402,11 @@ define <4 x float> @test_mask_andnot_ps_
}
define <4 x float> @test_mask_andnot_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmbkz_128
- ;CHECK: vandnps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -308,53 +417,73 @@ define <4 x float> @test_mask_andnot_ps_
declare <4 x float> @llvm.x86.avx512.mask.andn.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
define <8 x float> @test_mask_andnot_ps_rr_256(<8 x float> %a, <8 x float> %b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rr_256
- ;CHECK: vandnps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x55,0xc1]
+; CHECK-LABEL: test_mask_andnot_ps_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
ret <8 x float> %res
}
define <8 x float> @test_mask_andnot_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rrk_256
- ;CHECK: vandnps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x55,0xd1]
+; CHECK-LABEL: test_mask_andnot_ps_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x55,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_andnot_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rrkz_256
- ;CHECK: vandnps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x55,0xc1]
+; CHECK-LABEL: test_mask_andnot_ps_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandnps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_andnot_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rm_256
- ;CHECK: vandnps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
ret <8 x float> %res
}
define <8 x float> @test_mask_andnot_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmk_256
- ;CHECK: vandnps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x55,0x0f]
+; CHECK-LABEL: test_mask_andnot_ps_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x55,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_andnot_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmkz_256
- ;CHECK: vandnps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_andnot_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmb_256
- ;CHECK: vandnps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -363,8 +492,12 @@ define <8 x float> @test_mask_andnot_ps_
}
define <8 x float> @test_mask_andnot_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmbk_256
- ;CHECK: vandnps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x55,0x0f]
+; CHECK-LABEL: test_mask_andnot_ps_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x55,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -373,8 +506,11 @@ define <8 x float> @test_mask_andnot_ps_
}
define <8 x float> @test_mask_andnot_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmbkz_256
- ;CHECK: vandnps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -385,53 +521,73 @@ define <8 x float> @test_mask_andnot_ps_
declare <8 x float> @llvm.x86.avx512.mask.andn.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
define <16 x float> @test_mask_andnot_ps_rr_512(<16 x float> %a, <16 x float> %b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rr_512
- ;CHECK: vandnps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x55,0xc1]
+; CHECK-LABEL: test_mask_andnot_ps_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
ret <16 x float> %res
}
define <16 x float> @test_mask_andnot_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rrk_512
- ;CHECK: vandnps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x55,0xd1]
+; CHECK-LABEL: test_mask_andnot_ps_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x55,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_andnot_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rrkz_512
- ;CHECK: vandnps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x55,0xc1]
+; CHECK-LABEL: test_mask_andnot_ps_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vandnps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x55,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_andnot_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rm_512
- ;CHECK: vandnps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
ret <16 x float> %res
}
define <16 x float> @test_mask_andnot_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmk_512
- ;CHECK: vandnps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x55,0x0f]
+; CHECK-LABEL: test_mask_andnot_ps_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x55,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_andnot_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmkz_512
- ;CHECK: vandnps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_andnot_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmb_512
- ;CHECK: vandnps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -440,8 +596,12 @@ define <16 x float> @test_mask_andnot_ps
}
define <16 x float> @test_mask_andnot_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmbk_512
- ;CHECK: vandnps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x55,0x0f]
+; CHECK-LABEL: test_mask_andnot_ps_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x55,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -450,8 +610,11 @@ define <16 x float> @test_mask_andnot_ps
}
define <16 x float> @test_mask_andnot_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_andnot_ps_rmbkz_512
- ;CHECK: vandnps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x55,0x07]
+; CHECK-LABEL: test_mask_andnot_ps_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandnps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x55,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -462,53 +625,73 @@ define <16 x float> @test_mask_andnot_ps
declare <16 x float> @llvm.x86.avx512.mask.andn.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
define <4 x float> @test_mask_and_ps_rr_128(<4 x float> %a, <4 x float> %b) {
- ;CHECK-LABEL: test_mask_and_ps_rr_128
- ;CHECK: vandps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x54,0xc1]
+; CHECK-LABEL: test_mask_and_ps_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
ret <4 x float> %res
}
define <4 x float> @test_mask_and_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rrk_128
- ;CHECK: vandps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x54,0xd1]
+; CHECK-LABEL: test_mask_and_ps_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x54,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_and_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rrkz_128
- ;CHECK: vandps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x54,0xc1]
+; CHECK-LABEL: test_mask_and_ps_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_and_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_and_ps_rm_128
- ;CHECK: vandps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
ret <4 x float> %res
}
define <4 x float> @test_mask_and_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmk_128
- ;CHECK: vandps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x54,0x0f]
+; CHECK-LABEL: test_mask_and_ps_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x54,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_and_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmkz_128
- ;CHECK: vandps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_and_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_and_ps_rmb_128
- ;CHECK: vandps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -517,8 +700,12 @@ define <4 x float> @test_mask_and_ps_rmb
}
define <4 x float> @test_mask_and_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmbk_128
- ;CHECK: vandps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x54,0x0f]
+; CHECK-LABEL: test_mask_and_ps_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x54,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -527,8 +714,11 @@ define <4 x float> @test_mask_and_ps_rmb
}
define <4 x float> @test_mask_and_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmbkz_128
- ;CHECK: vandps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -539,53 +729,73 @@ define <4 x float> @test_mask_and_ps_rmb
declare <4 x float> @llvm.x86.avx512.mask.and.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
define <8 x float> @test_mask_and_ps_rr_256(<8 x float> %a, <8 x float> %b) {
- ;CHECK-LABEL: test_mask_and_ps_rr_256
- ;CHECK: vandps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x54,0xc1]
+; CHECK-LABEL: test_mask_and_ps_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
ret <8 x float> %res
}
define <8 x float> @test_mask_and_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rrk_256
- ;CHECK: vandps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x54,0xd1]
+; CHECK-LABEL: test_mask_and_ps_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x54,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_and_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rrkz_256
- ;CHECK: vandps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x54,0xc1]
+; CHECK-LABEL: test_mask_and_ps_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vandps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_and_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_and_ps_rm_256
- ;CHECK: vandps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
ret <8 x float> %res
}
define <8 x float> @test_mask_and_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmk_256
- ;CHECK: vandps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x54,0x0f]
+; CHECK-LABEL: test_mask_and_ps_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x54,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_and_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmkz_256
- ;CHECK: vandps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_and_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_and_ps_rmb_256
- ;CHECK: vandps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -594,8 +804,12 @@ define <8 x float> @test_mask_and_ps_rmb
}
define <8 x float> @test_mask_and_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmbk_256
- ;CHECK: vandps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x54,0x0f]
+; CHECK-LABEL: test_mask_and_ps_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x54,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -604,8 +818,11 @@ define <8 x float> @test_mask_and_ps_rmb
}
define <8 x float> @test_mask_and_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmbkz_256
- ;CHECK: vandps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -616,53 +833,73 @@ define <8 x float> @test_mask_and_ps_rmb
declare <8 x float> @llvm.x86.avx512.mask.and.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
define <16 x float> @test_mask_and_ps_rr_512(<16 x float> %a, <16 x float> %b) {
- ;CHECK-LABEL: test_mask_and_ps_rr_512
- ;CHECK: vandps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x54,0xc1]
+; CHECK-LABEL: test_mask_and_ps_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
ret <16 x float> %res
}
define <16 x float> @test_mask_and_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rrk_512
- ;CHECK: vandps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x54,0xd1]
+; CHECK-LABEL: test_mask_and_ps_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x54,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_and_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rrkz_512
- ;CHECK: vandps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x54,0xc1]
+; CHECK-LABEL: test_mask_and_ps_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vandps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x54,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_and_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_and_ps_rm_512
- ;CHECK: vandps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
ret <16 x float> %res
}
define <16 x float> @test_mask_and_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmk_512
- ;CHECK: vandps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x54,0x0f]
+; CHECK-LABEL: test_mask_and_ps_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x54,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_and_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmkz_512
- ;CHECK: vandps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_and_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_and_ps_rmb_512
- ;CHECK: vandps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -671,8 +908,12 @@ define <16 x float> @test_mask_and_ps_rm
}
define <16 x float> @test_mask_and_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmbk_512
- ;CHECK: vandps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x54,0x0f]
+; CHECK-LABEL: test_mask_and_ps_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x54,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -681,8 +922,11 @@ define <16 x float> @test_mask_and_ps_rm
}
define <16 x float> @test_mask_and_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_and_ps_rmbkz_512
- ;CHECK: vandps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x54,0x07]
+; CHECK-LABEL: test_mask_and_ps_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vandps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x54,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -693,53 +937,73 @@ define <16 x float> @test_mask_and_ps_rm
declare <16 x float> @llvm.x86.avx512.mask.and.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
define <4 x float> @test_mask_or_ps_rr_128(<4 x float> %a, <4 x float> %b) {
- ;CHECK-LABEL: test_mask_or_ps_rr_128
- ;CHECK: vorps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x56,0xc1]
+; CHECK-LABEL: test_mask_or_ps_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
ret <4 x float> %res
}
define <4 x float> @test_mask_or_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rrk_128
- ;CHECK: vorps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x56,0xd1]
+; CHECK-LABEL: test_mask_or_ps_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x56,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_or_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rrkz_128
- ;CHECK: vorps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x56,0xc1]
+; CHECK-LABEL: test_mask_or_ps_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vorps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_or_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_or_ps_rm_128
- ;CHECK: vorps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
ret <4 x float> %res
}
define <4 x float> @test_mask_or_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmk_128
- ;CHECK: vorps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x56,0x0f]
+; CHECK-LABEL: test_mask_or_ps_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x56,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_or_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmkz_128
- ;CHECK: vorps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_or_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_or_ps_rmb_128
- ;CHECK: vorps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -748,8 +1012,12 @@ define <4 x float> @test_mask_or_ps_rmb_
}
define <4 x float> @test_mask_or_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmbk_128
- ;CHECK: vorps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x56,0x0f]
+; CHECK-LABEL: test_mask_or_ps_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x56,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -758,8 +1026,11 @@ define <4 x float> @test_mask_or_ps_rmbk
}
define <4 x float> @test_mask_or_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmbkz_128
- ;CHECK: vorps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -770,53 +1041,73 @@ define <4 x float> @test_mask_or_ps_rmbk
declare <4 x float> @llvm.x86.avx512.mask.or.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
define <8 x float> @test_mask_or_ps_rr_256(<8 x float> %a, <8 x float> %b) {
- ;CHECK-LABEL: test_mask_or_ps_rr_256
- ;CHECK: vorps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x56,0xc1]
+; CHECK-LABEL: test_mask_or_ps_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
ret <8 x float> %res
}
define <8 x float> @test_mask_or_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rrk_256
- ;CHECK: vorps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x56,0xd1]
+; CHECK-LABEL: test_mask_or_ps_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x56,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_or_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rrkz_256
- ;CHECK: vorps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x56,0xc1]
+; CHECK-LABEL: test_mask_or_ps_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vorps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_or_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_or_ps_rm_256
- ;CHECK: vorps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
ret <8 x float> %res
}
define <8 x float> @test_mask_or_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmk_256
- ;CHECK: vorps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x56,0x0f]
+; CHECK-LABEL: test_mask_or_ps_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x56,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_or_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmkz_256
- ;CHECK: vorps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_or_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_or_ps_rmb_256
- ;CHECK: vorps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -825,8 +1116,12 @@ define <8 x float> @test_mask_or_ps_rmb_
}
define <8 x float> @test_mask_or_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmbk_256
- ;CHECK: vorps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x56,0x0f]
+; CHECK-LABEL: test_mask_or_ps_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x56,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -835,8 +1130,11 @@ define <8 x float> @test_mask_or_ps_rmbk
}
define <8 x float> @test_mask_or_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmbkz_256
- ;CHECK: vorps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -847,53 +1145,73 @@ define <8 x float> @test_mask_or_ps_rmbk
declare <8 x float> @llvm.x86.avx512.mask.or.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
define <16 x float> @test_mask_or_ps_rr_512(<16 x float> %a, <16 x float> %b) {
- ;CHECK-LABEL: test_mask_or_ps_rr_512
- ;CHECK: vorps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x56,0xc1]
+; CHECK-LABEL: test_mask_or_ps_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
ret <16 x float> %res
}
define <16 x float> @test_mask_or_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rrk_512
- ;CHECK: vorps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x56,0xd1]
+; CHECK-LABEL: test_mask_or_ps_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x56,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_or_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rrkz_512
- ;CHECK: vorps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x56,0xc1]
+; CHECK-LABEL: test_mask_or_ps_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vorps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x56,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_or_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_or_ps_rm_512
- ;CHECK: vorps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
ret <16 x float> %res
}
define <16 x float> @test_mask_or_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmk_512
- ;CHECK: vorps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x56,0x0f]
+; CHECK-LABEL: test_mask_or_ps_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x56,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_or_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmkz_512
- ;CHECK: vorps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_or_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_or_ps_rmb_512
- ;CHECK: vorps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -902,8 +1220,12 @@ define <16 x float> @test_mask_or_ps_rmb
}
define <16 x float> @test_mask_or_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmbk_512
- ;CHECK: vorps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x56,0x0f]
+; CHECK-LABEL: test_mask_or_ps_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x56,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -912,8 +1234,11 @@ define <16 x float> @test_mask_or_ps_rmb
}
define <16 x float> @test_mask_or_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_or_ps_rmbkz_512
- ;CHECK: vorps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x56,0x07]
+; CHECK-LABEL: test_mask_or_ps_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vorps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x56,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -924,53 +1249,73 @@ define <16 x float> @test_mask_or_ps_rmb
declare <16 x float> @llvm.x86.avx512.mask.or.ps.512(<16 x float>, <16 x float>, <16 x float>, i16)
define <4 x float> @test_mask_xor_ps_rr_128(<4 x float> %a, <4 x float> %b) {
- ;CHECK-LABEL: test_mask_xor_ps_rr_128
- ;CHECK: vxorps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x57,0xc1]
+; CHECK-LABEL: test_mask_xor_ps_rr_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
ret <4 x float> %res
}
define <4 x float> @test_mask_xor_ps_rrk_128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rrk_128
- ;CHECK: vxorps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x57,0xd1]
+; CHECK-LABEL: test_mask_xor_ps_rrk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x57,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_xor_ps_rrkz_128(<4 x float> %a, <4 x float> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rrkz_128
- ;CHECK: vxorps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x57,0xc1]
+; CHECK-LABEL: test_mask_xor_ps_rrkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vxorps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_xor_ps_rm_128(<4 x float> %a, <4 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_xor_ps_rm_128
- ;CHECK: vxorps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rm_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 -1)
ret <4 x float> %res
}
define <4 x float> @test_mask_xor_ps_rmk_128(<4 x float> %a, <4 x float>* %ptr_b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmk_128
- ;CHECK: vxorps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x57,0x0f]
+; CHECK-LABEL: test_mask_xor_ps_rmk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x57,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> %passThru, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_xor_ps_rmkz_128(<4 x float> %a, <4 x float>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmkz_128
- ;CHECK: vxorps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rmkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <4 x float>, <4 x float>* %ptr_b
%res = call <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float> %a, <4 x float> %b, <4 x float> zeroinitializer, i8 %mask)
ret <4 x float> %res
}
define <4 x float> @test_mask_xor_ps_rmb_128(<4 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_xor_ps_rmb_128
- ;CHECK: vxorps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rmb_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x18,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -979,8 +1324,12 @@ define <4 x float> @test_mask_xor_ps_rmb
}
define <4 x float> @test_mask_xor_ps_rmbk_128(<4 x float> %a, float* %ptr_b, <4 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmbk_128
- ;CHECK: vxorps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x57,0x0f]
+; CHECK-LABEL: test_mask_xor_ps_rmbk_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x19,0x57,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -989,8 +1338,11 @@ define <4 x float> @test_mask_xor_ps_rmb
}
define <4 x float> @test_mask_xor_ps_rmbkz_128(<4 x float> %a, float* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmbkz_128
- ;CHECK: vxorps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rmbkz_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x99,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <4 x float> undef, float %q, i32 0
%b = shufflevector <4 x float> %vecinit.i, <4 x float> undef, <4 x i32> zeroinitializer
@@ -1001,53 +1353,73 @@ define <4 x float> @test_mask_xor_ps_rmb
declare <4 x float> @llvm.x86.avx512.mask.xor.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)
define <8 x float> @test_mask_xor_ps_rr_256(<8 x float> %a, <8 x float> %b) {
- ;CHECK-LABEL: test_mask_xor_ps_rr_256
- ;CHECK: vxorps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x57,0xc1]
+; CHECK-LABEL: test_mask_xor_ps_rr_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
ret <8 x float> %res
}
define <8 x float> @test_mask_xor_ps_rrk_256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rrk_256
- ;CHECK: vxorps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x57,0xd1]
+; CHECK-LABEL: test_mask_xor_ps_rrk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x57,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_xor_ps_rrkz_256(<8 x float> %a, <8 x float> %b, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rrkz_256
- ;CHECK: vxorps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x57,0xc1]
+; CHECK-LABEL: test_mask_xor_ps_rrkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vxorps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_xor_ps_rm_256(<8 x float> %a, <8 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_xor_ps_rm_256
- ;CHECK: vxorps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rm_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 -1)
ret <8 x float> %res
}
define <8 x float> @test_mask_xor_ps_rmk_256(<8 x float> %a, <8 x float>* %ptr_b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmk_256
- ;CHECK: vxorps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x57,0x0f]
+; CHECK-LABEL: test_mask_xor_ps_rmk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x57,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %passThru, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_xor_ps_rmkz_256(<8 x float> %a, <8 x float>* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmkz_256
- ;CHECK: vxorps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rmkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <8 x float>, <8 x float>* %ptr_b
%res = call <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> zeroinitializer, i8 %mask)
ret <8 x float> %res
}
define <8 x float> @test_mask_xor_ps_rmb_256(<8 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_xor_ps_rmb_256
- ;CHECK: vxorps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rmb_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x38,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -1056,8 +1428,12 @@ define <8 x float> @test_mask_xor_ps_rmb
}
define <8 x float> @test_mask_xor_ps_rmbk_256(<8 x float> %a, float* %ptr_b, <8 x float> %passThru, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmbk_256
- ;CHECK: vxorps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x57,0x0f]
+; CHECK-LABEL: test_mask_xor_ps_rmbk_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x39,0x57,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -1066,8 +1442,11 @@ define <8 x float> @test_mask_xor_ps_rmb
}
define <8 x float> @test_mask_xor_ps_rmbkz_256(<8 x float> %a, float* %ptr_b, i8 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmbkz_256
- ;CHECK: vxorps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rmbkz_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %esi, %k1 ## encoding: [0xc5,0xf9,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xb9,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <8 x float> undef, float %q, i32 0
%b = shufflevector <8 x float> %vecinit.i, <8 x float> undef, <8 x i32> zeroinitializer
@@ -1078,53 +1457,73 @@ define <8 x float> @test_mask_xor_ps_rmb
declare <8 x float> @llvm.x86.avx512.mask.xor.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)
define <16 x float> @test_mask_xor_ps_rr_512(<16 x float> %a, <16 x float> %b) {
- ;CHECK-LABEL: test_mask_xor_ps_rr_512
- ;CHECK: vxorps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x57,0xc1]
+; CHECK-LABEL: test_mask_xor_ps_rr_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
ret <16 x float> %res
}
define <16 x float> @test_mask_xor_ps_rrk_512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rrk_512
- ;CHECK: vxorps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x57,0xd1]
+; CHECK-LABEL: test_mask_xor_ps_rrk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x57,0xd1]
+; CHECK-NEXT: vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_xor_ps_rrkz_512(<16 x float> %a, <16 x float> %b, i16 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rrkz_512
- ;CHECK: vxorps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x57,0xc1]
+; CHECK-LABEL: test_mask_xor_ps_rrkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
+; CHECK-NEXT: vxorps %zmm1, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x57,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_xor_ps_rm_512(<16 x float> %a, <16 x float>* %ptr_b) {
- ;CHECK-LABEL: test_mask_xor_ps_rm_512
- ;CHECK: vxorps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rm_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 -1)
ret <16 x float> %res
}
define <16 x float> @test_mask_xor_ps_rmk_512(<16 x float> %a, <16 x float>* %ptr_b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmk_512
- ;CHECK: vxorps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x57,0x0f]
+; CHECK-LABEL: test_mask_xor_ps_rmk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x49,0x57,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> %passThru, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_xor_ps_rmkz_512(<16 x float> %a, <16 x float>* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmkz_512
- ;CHECK: vxorps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rmkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi), %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xc9,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%b = load <16 x float>, <16 x float>* %ptr_b
%res = call <16 x float> @llvm.x86.avx512.mask.xor.ps.512(<16 x float> %a, <16 x float> %b, <16 x float> zeroinitializer, i16 %mask)
ret <16 x float> %res
}
define <16 x float> @test_mask_xor_ps_rmb_512(<16 x float> %a, float* %ptr_b) {
- ;CHECK-LABEL: test_mask_xor_ps_rmb_512
- ;CHECK: vxorps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x57,0x07]
+; CHECK-LABEL: test_mask_xor_ps_rmb_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x58,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -1133,8 +1532,12 @@ define <16 x float> @test_mask_xor_ps_rm
}
define <16 x float> @test_mask_xor_ps_rmbk_512(<16 x float> %a, float* %ptr_b, <16 x float> %passThru, i16 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmbk_512
- ;CHECK: vxorps (%rdi){1to16}, %zmm0, %zmm1 {%k1}
+; CHECK-LABEL: test_mask_xor_ps_rmbk_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x59,0x57,0x0f]
+; CHECK-NEXT: vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -1143,8 +1546,11 @@ define <16 x float> @test_mask_xor_ps_rm
}
define <16 x float> @test_mask_xor_ps_rmbkz_512(<16 x float> %a, float* %ptr_b, i16 %mask) {
- ;CHECK-LABEL: test_mask_xor_ps_rmbkz_512
- ;CHECK: vxorps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z}
+; CHECK-LABEL: test_mask_xor_ps_rmbkz_512:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
+; CHECK-NEXT: vxorps (%rdi){1to16}, %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xd9,0x57,0x07]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%q = load float, float* %ptr_b
%vecinit.i = insertelement <16 x float> undef, float %q, i32 0
%b = shufflevector <16 x float> %vecinit.i, <16 x float> undef, <16 x i32> zeroinitializer
@@ -1159,11 +1565,11 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_cvt_pd2qq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2qq_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtpd2qq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtpd2qq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtpd2qq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x7b,0xc8]
+; CHECK-NEXT: vcvtpd2qq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x7b,0xc0]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.cvtpd2qq.128(<2 x double> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.cvtpd2qq.128(<2 x double> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
@@ -1175,11 +1581,11 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_cvt_pd2qq_256(<4 x double> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2qq_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtpd2qq %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvtpd2qq %ymm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtpd2qq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x7b,0xc8]
+; CHECK-NEXT: vcvtpd2qq %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x7b,0xc0]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.cvtpd2qq.256(<4 x double> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.cvtpd2qq.256(<4 x double> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
@@ -1191,11 +1597,11 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_cvt_pd2uqq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2uqq_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtpd2uqq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtpd2uqq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtpd2uqq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x79,0xc8]
+; CHECK-NEXT: vcvtpd2uqq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x79,0xc0]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.128(<2 x double> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.128(<2 x double> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
@@ -1207,11 +1613,11 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_cvt_pd2uqq_256(<4 x double> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2uqq_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtpd2uqq %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvtpd2uqq %ymm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtpd2uqq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x79,0xc8]
+; CHECK-NEXT: vcvtpd2uqq %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x79,0xc0]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.256(<4 x double> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.cvtpd2uqq.256(<4 x double> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
@@ -1223,11 +1629,11 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_cvt_ps2qq_128(<4 x float> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtps2qq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtps2qq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtps2qq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x7b,0xc8]
+; CHECK-NEXT: vcvtps2qq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x7b,0xc0]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.cvtps2qq.128(<4 x float> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
@@ -1239,11 +1645,11 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_cvt_ps2qq_256(<4 x float> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2qq_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtps2qq %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvtps2qq %xmm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtps2qq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x7b,0xc8]
+; CHECK-NEXT: vcvtps2qq %xmm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x7b,0xc0]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.cvtps2qq.256(<4 x float> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.cvtps2qq.256(<4 x float> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
@@ -1255,11 +1661,11 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_cvt_ps2uqq_128(<4 x float> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtps2uqq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtps2uqq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtps2uqq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x79,0xc8]
+; CHECK-NEXT: vcvtps2uqq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x79,0xc0]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.cvtps2uqq.128(<4 x float> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
@@ -1271,11 +1677,11 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_cvt_ps2uqq_256(<4 x float> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2uqq_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtps2uqq %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvtps2uqq %xmm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtps2uqq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x79,0xc8]
+; CHECK-NEXT: vcvtps2uqq %xmm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x79,0xc0]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.cvtps2uqq.256(<4 x float> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.cvtps2uqq.256(<4 x float> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
@@ -1287,11 +1693,11 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_mask_cvt_qq2pd_128(<2 x i64> %x0, <2 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_qq2pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtqq2pd %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtqq2pd %xmm0, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtqq2pd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfe,0x09,0xe6,0xc8]
+; CHECK-NEXT: vcvtqq2pd %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfe,0x08,0xe6,0xc0]
+; CHECK-NEXT: vaddpd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.cvtqq2pd.128(<2 x i64> %x0, <2 x double> %x1, i8 %x2)
%res1 = call <2 x double> @llvm.x86.avx512.mask.cvtqq2pd.128(<2 x i64> %x0, <2 x double> %x1, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1303,11 +1709,11 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask_cvt_qq2pd_256(<4 x i64> %x0, <4 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_qq2pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtqq2pd %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvtqq2pd %ymm0, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtqq2pd %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfe,0x29,0xe6,0xc8]
+; CHECK-NEXT: vcvtqq2pd %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfe,0x28,0xe6,0xc0]
+; CHECK-NEXT: vaddpd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.cvtqq2pd.256(<4 x i64> %x0, <4 x double> %x1, i8 %x2)
%res1 = call <4 x double> @llvm.x86.avx512.mask.cvtqq2pd.256(<4 x i64> %x0, <4 x double> %x1, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1319,11 +1725,11 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask_cvt_qq2ps_128(<2 x i64> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_qq2ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtqq2ps %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtqq2ps %xmm0, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtqq2ps %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfc,0x09,0x5b,0xc8]
+; CHECK-NEXT: vcvtqq2ps %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfc,0x08,0x5b,0xc0]
+; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x74,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 %x2)
%res1 = call <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1335,11 +1741,11 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask_cvt_qq2ps_256(<4 x i64> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_qq2ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtqq2ps %ymm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtqq2ps %ymm0, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtqq2ps %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfc,0x29,0x5b,0xc8]
+; CHECK-NEXT: vcvtqq2ps %ymm0, %xmm0 ## encoding: [0x62,0xf1,0xfc,0x28,0x5b,0xc0]
+; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x74,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.256(<4 x i64> %x0, <4 x float> %x1, i8 %x2)
%res1 = call <4 x float> @llvm.x86.avx512.mask.cvtqq2ps.256(<4 x i64> %x0, <4 x float> %x1, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1351,11 +1757,11 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_cvtt_pd2qq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_pd2qq_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvttpd2qq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvttpd2qq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvttpd2qq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x7a,0xc8]
+; CHECK-NEXT: vcvttpd2qq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x7a,0xc0]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.cvttpd2qq.128(<2 x double> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.cvttpd2qq.128(<2 x double> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
@@ -1367,11 +1773,11 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_cvtt_pd2qq_256(<4 x double> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_pd2qq_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvttpd2qq %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvttpd2qq %ymm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvttpd2qq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x7a,0xc8]
+; CHECK-NEXT: vcvttpd2qq %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x7a,0xc0]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.cvttpd2qq.256(<4 x double> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.cvttpd2qq.256(<4 x double> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
@@ -1383,11 +1789,11 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_cvtt_pd2uqq_128(<2 x double> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_pd2uqq_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvttpd2uqq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvttpd2uqq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvttpd2uqq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x78,0xc8]
+; CHECK-NEXT: vcvttpd2uqq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x78,0xc0]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.128(<2 x double> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.128(<2 x double> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
@@ -1399,11 +1805,11 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_cvtt_pd2uqq_256(<4 x double> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_pd2uqq_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvttpd2uqq %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvttpd2uqq %ymm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvttpd2uqq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x78,0xc8]
+; CHECK-NEXT: vcvttpd2uqq %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x78,0xc0]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.256(<4 x double> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.cvttpd2uqq.256(<4 x double> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
@@ -1415,11 +1821,11 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_cvtt_ps2qq_128(<4 x float> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvttps2qq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvttps2qq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvttps2qq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x7a,0xc8]
+; CHECK-NEXT: vcvttps2qq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x7a,0xc0]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.cvttps2qq.128(<4 x float> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
@@ -1431,11 +1837,11 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_cvtt_ps2qq_256(<4 x float> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_ps2qq_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvttps2qq %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvttps2qq %xmm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvttps2qq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x7a,0xc8]
+; CHECK-NEXT: vcvttps2qq %xmm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x7a,0xc0]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.cvttps2qq.256(<4 x float> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.cvttps2qq.256(<4 x float> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
@@ -1447,11 +1853,11 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_mask_cvt_uqq2pd_128(<2 x i64> %x0, <2 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_uqq2pd_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtuqq2pd %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtuqq2pd %xmm0, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtuqq2pd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfe,0x09,0x7a,0xc8]
+; CHECK-NEXT: vcvtuqq2pd %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfe,0x08,0x7a,0xc0]
+; CHECK-NEXT: vaddpd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.cvtuqq2pd.128(<2 x i64> %x0, <2 x double> %x1, i8 %x2)
%res1 = call <2 x double> @llvm.x86.avx512.mask.cvtuqq2pd.128(<2 x i64> %x0, <2 x double> %x1, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1463,11 +1869,11 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask_cvt_uqq2pd_256(<4 x i64> %x0, <4 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_uqq2pd_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtuqq2pd %ymm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvtuqq2pd %ymm0, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtuqq2pd %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfe,0x29,0x7a,0xc8]
+; CHECK-NEXT: vcvtuqq2pd %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfe,0x28,0x7a,0xc0]
+; CHECK-NEXT: vaddpd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.cvtuqq2pd.256(<4 x i64> %x0, <4 x double> %x1, i8 %x2)
%res1 = call <4 x double> @llvm.x86.avx512.mask.cvtuqq2pd.256(<4 x i64> %x0, <4 x double> %x1, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1479,11 +1885,11 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask_cvt_uqq2ps_128(<2 x i64> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_uqq2ps_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtuqq2ps %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtuqq2ps %xmm0, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtuqq2ps %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xff,0x09,0x7a,0xc8]
+; CHECK-NEXT: vcvtuqq2ps %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xff,0x08,0x7a,0xc0]
+; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x74,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 %x2)
%res1 = call <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.128(<2 x i64> %x0, <4 x float> %x1, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1495,11 +1901,11 @@ declare <4 x float> @llvm.x86.avx512.mas
define <4 x float>@test_int_x86_avx512_mask_cvt_uqq2ps_256(<4 x i64> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_uqq2ps_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvtuqq2ps %ymm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvtuqq2ps %ymm0, %xmm0
-; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvtuqq2ps %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xff,0x29,0x7a,0xc8]
+; CHECK-NEXT: vcvtuqq2ps %ymm0, %xmm0 ## encoding: [0x62,0xf1,0xff,0x28,0x7a,0xc0]
+; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x74,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.256(<4 x i64> %x0, <4 x float> %x1, i8 %x2)
%res1 = call <4 x float> @llvm.x86.avx512.mask.cvtuqq2ps.256(<4 x i64> %x0, <4 x float> %x1, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1511,11 +1917,11 @@ declare <2 x i64> @llvm.x86.avx512.mask.
define <2 x i64>@test_int_x86_avx512_mask_cvtt_ps2uqq_128(<4 x float> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvttps2uqq %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vcvttps2uqq %xmm0, %xmm0
-; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvttps2uqq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x78,0xc8]
+; CHECK-NEXT: vcvttps2uqq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x78,0xc0]
+; CHECK-NEXT: vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0, <2 x i64> %x1, i8 %x2)
%res1 = call <2 x i64> @llvm.x86.avx512.mask.cvttps2uqq.128(<4 x float> %x0, <2 x i64> %x1, i8 -1)
%res2 = add <2 x i64> %res, %res1
@@ -1527,11 +1933,11 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_cvtt_ps2uqq_256(<4 x float> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_ps2uqq_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vcvttps2uqq %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vcvttps2uqq %xmm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vcvttps2uqq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x78,0xc8]
+; CHECK-NEXT: vcvttps2uqq %xmm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x78,0xc0]
+; CHECK-NEXT: vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.cvttps2uqq.256(<4 x float> %x0, <4 x i64> %x1, i8 %x2)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.cvttps2uqq.256(<4 x float> %x0, <4 x i64> %x1, i8 -1)
%res2 = add <4 x i64> %res, %res1
@@ -1540,12 +1946,14 @@ define <4 x i64>@test_int_x86_avx512_mas
declare <2 x double> @llvm.x86.avx512.mask.reduce.pd.128(<2 x double>, i32, <2 x double>, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_reduce_pd_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vreducepd {{.*}}{%k1}
-; CHECK: vreducepd
define <2 x double>@test_int_x86_avx512_mask_reduce_pd_128(<2 x double> %x0, <2 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_reduce_pd_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vreducepd $4, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x56,0xc8,0x04]
+; CHECK-NEXT: vreducepd $8, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0xfd,0x08,0x56,0xc0,0x08]
+; CHECK-NEXT: vaddpd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.reduce.pd.128(<2 x double> %x0, i32 4, <2 x double> %x2, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.reduce.pd.128(<2 x double> %x0, i32 8, <2 x double> %x2, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1554,12 +1962,14 @@ define <2 x double>@test_int_x86_avx512_
declare <4 x double> @llvm.x86.avx512.mask.reduce.pd.256(<4 x double>, i32, <4 x double>, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_reduce_pd_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vreducepd {{.*}}{%k1}
-; CHECK: vreducepd
define <4 x double>@test_int_x86_avx512_mask_reduce_pd_256(<4 x double> %x0, <4 x double> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_reduce_pd_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vreducepd $4, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x56,0xc8,0x04]
+; CHECK-NEXT: vreducepd $0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x56,0xc0,0x00]
+; CHECK-NEXT: vaddpd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.reduce.pd.256(<4 x double> %x0, i32 4, <4 x double> %x2, i8 %x3)
%res1 = call <4 x double> @llvm.x86.avx512.mask.reduce.pd.256(<4 x double> %x0, i32 0, <4 x double> %x2, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1567,12 +1977,15 @@ define <4 x double>@test_int_x86_avx512_
}
declare <4 x float> @llvm.x86.avx512.mask.reduce.ps.128(<4 x float>, i32, <4 x float>, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_reduce_ps_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vreduceps {{.*}}{%k1}
-; CHECK: vreduceps
+
define <4 x float>@test_int_x86_avx512_mask_reduce_ps_128(<4 x float> %x0, <4 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_reduce_ps_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vreduceps $4, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x56,0xc8,0x04]
+; CHECK-NEXT: vreduceps $88, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x56,0xc0,0x58]
+; CHECK-NEXT: vaddps %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x74,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.reduce.ps.128(<4 x float> %x0, i32 4, <4 x float> %x2, i8 %x3)
%res1 = call <4 x float> @llvm.x86.avx512.mask.reduce.ps.128(<4 x float> %x0, i32 88, <4 x float> %x2, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1581,12 +1994,14 @@ define <4 x float>@test_int_x86_avx512_m
declare <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float>, i32, <8 x float>, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_reduce_ps_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vreduceps {{.*}}{%k1}
-; CHECK: vreduceps
define <8 x float>@test_int_x86_avx512_mask_reduce_ps_256(<8 x float> %x0, <8 x float> %x2, i8 %x3) {
+; CHECK-LABEL: test_int_x86_avx512_mask_reduce_ps_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vreduceps $11, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x56,0xc8,0x0b]
+; CHECK-NEXT: vreduceps $11, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x56,0xc0,0x0b]
+; CHECK-NEXT: vaddps %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x74,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.reduce.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -1595,12 +2010,14 @@ define <8 x float>@test_int_x86_avx512_m
declare <2 x double> @llvm.x86.avx512.mask.range.pd.128(<2 x double>, <2 x double>, i32, <2 x double>, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_range_pd_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vrangepd {{.*}}{%k1}
-; CHECK: vrangepd
define <2 x double>@test_int_x86_avx512_mask_range_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_range_pd_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vrangepd $4, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x50,0xd1,0x04]
+; CHECK-NEXT: vrangepd $8, %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0xfd,0x08,0x50,0xc1,0x08]
+; CHECK-NEXT: vaddpd %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.range.pd.128(<2 x double> %x0, <2 x double> %x1, i32 4, <2 x double> %x3, i8 %x4)
%res1 = call <2 x double> @llvm.x86.avx512.mask.range.pd.128(<2 x double> %x0, <2 x double> %x1, i32 8, <2 x double> %x3, i8 -1)
%res2 = fadd <2 x double> %res, %res1
@@ -1609,12 +2026,14 @@ define <2 x double>@test_int_x86_avx512_
declare <4 x double> @llvm.x86.avx512.mask.range.pd.256(<4 x double>, <4 x double>, i32, <4 x double>, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_range_pd_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vrangepd {{.*}}{%k1}
-; CHECK: vrangepd
define <4 x double>@test_int_x86_avx512_mask_range_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_range_pd_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vrangepd $4, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x50,0xd1,0x04]
+; CHECK-NEXT: vrangepd $88, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x50,0xc1,0x58]
+; CHECK-NEXT: vaddpd %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.range.pd.256(<4 x double> %x0, <4 x double> %x1, i32 4, <4 x double> %x3, i8 %x4)
%res1 = call <4 x double> @llvm.x86.avx512.mask.range.pd.256(<4 x double> %x0, <4 x double> %x1, i32 88, <4 x double> %x3, i8 -1)
%res2 = fadd <4 x double> %res, %res1
@@ -1623,12 +2042,14 @@ define <4 x double>@test_int_x86_avx512_
declare <4 x float> @llvm.x86.avx512.mask.range.ps.128(<4 x float>, <4 x float>, i32, <4 x float>, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_range_ps_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vrangeps {{.*}}{%k1}
-; CHECK: vrangeps
define <4 x float>@test_int_x86_avx512_mask_range_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_range_ps_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vrangeps $4, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x50,0xd1,0x04]
+; CHECK-NEXT: vrangeps $88, %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x50,0xc1,0x58]
+; CHECK-NEXT: vaddps %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6c,0x08,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x float> @llvm.x86.avx512.mask.range.ps.128(<4 x float> %x0, <4 x float> %x1, i32 4, <4 x float> %x3, i8 %x4)
%res1 = call <4 x float> @llvm.x86.avx512.mask.range.ps.128(<4 x float> %x0, <4 x float> %x1, i32 88, <4 x float> %x3, i8 -1)
%res2 = fadd <4 x float> %res, %res1
@@ -1637,12 +2058,14 @@ define <4 x float>@test_int_x86_avx512_m
declare <8 x float> @llvm.x86.avx512.mask.range.ps.256(<8 x float>, <8 x float>, i32, <8 x float>, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_range_ps_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vrangeps {{.*}}{%k1}
-; CHECK: vrangeps
define <8 x float>@test_int_x86_avx512_mask_range_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x3, i8 %x4) {
+; CHECK-LABEL: test_int_x86_avx512_mask_range_ps_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vrangeps $4, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x50,0xd1,0x04]
+; CHECK-NEXT: vrangeps $88, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x50,0xc1,0x58]
+; CHECK-NEXT: vaddps %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.range.ps.256(<8 x float> %x0, <8 x float> %x1, i32 4, <8 x float> %x3, i8 %x4)
%res1 = call <8 x float> @llvm.x86.avx512.mask.range.ps.256(<8 x float> %x0, <8 x float> %x1, i32 88, <8 x float> %x3, i8 -1)
%res2 = fadd <8 x float> %res, %res1
@@ -1654,13 +2077,13 @@ declare <2 x double> @llvm.x86.avx512.ma
define <2 x double>@test_int_x86_avx512_mask_vextractf64x2_256(<4 x double> %x0, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vextractf64x2_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vextractf64x2 $1, %ymm0, %xmm1 {%k1}
-; CHECK-NEXT: vextractf64x2 $1, %ymm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vextractf64x2 $1, %ymm0, %xmm0
-; CHECK-NEXT: vaddpd %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: vaddpd %xmm2, %xmm0, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vextractf64x2 $1, %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x19,0xc1,0x01]
+; CHECK-NEXT: vextractf64x2 $1, %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x19,0xc2,0x01]
+; CHECK-NEXT: vextractf64x2 $1, %ymm0, %xmm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x19,0xc0,0x01]
+; CHECK-NEXT: vaddpd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0x58,0xc0]
+; CHECK-NEXT: vaddpd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x58,0xc2]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x double> @llvm.x86.avx512.mask.vextractf64x2.256(<4 x double> %x0,i32 1, <2 x double> %x2, i8 %x3)
%res2 = call <2 x double> @llvm.x86.avx512.mask.vextractf64x2.256(<4 x double> %x0,i32 1, <2 x double> zeroinitializer, i8 %x3)
%res1 = call <2 x double> @llvm.x86.avx512.mask.vextractf64x2.256(<4 x double> %x0,i32 1, <2 x double> zeroinitializer, i8 -1)
@@ -1674,13 +2097,13 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask_insertf64x2_256(<4 x double> %x0, <2 x double> %x1, <4 x double> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_insertf64x2_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vinsertf64x2 $1, %xmm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vinsertf64x2 $1, %xmm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vinsertf64x2 $1, %xmm1, %ymm0, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vinsertf64x2 $1, %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x18,0xd1,0x01]
+; CHECK-NEXT: vinsertf64x2 $1, %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x18,0xd9,0x01]
+; CHECK-NEXT: vinsertf64x2 $1, %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x18,0xc1,0x01]
+; CHECK-NEXT: vaddpd %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0x58,0xc0]
+; CHECK-NEXT: vaddpd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x double> @llvm.x86.avx512.mask.insertf64x2.256(<4 x double> %x0, <2 x double> %x1, i32 1, <4 x double> %x3, i8 %x4)
%res1 = call <4 x double> @llvm.x86.avx512.mask.insertf64x2.256(<4 x double> %x0, <2 x double> %x1, i32 1, <4 x double> %x3, i8 -1)
%res2 = call <4 x double> @llvm.x86.avx512.mask.insertf64x2.256(<4 x double> %x0, <2 x double> %x1, i32 1, <4 x double> zeroinitializer, i8 %x4)
@@ -1694,13 +2117,13 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_inserti64x2_256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_inserti64x2_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vinserti64x2 $1, %xmm1, %ymm0, %ymm2 {%k1}
-; CHECK-NEXT: vinserti64x2 $1, %xmm1, %ymm0, %ymm3 {%k1} {z}
-; CHECK-NEXT: vinserti64x2 $1, %xmm1, %ymm0, %ymm0
-; CHECK-NEXT: vpaddq %ymm0, %ymm2, %ymm0
-; CHECK-NEXT: vpaddq %ymm3, %ymm0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vinserti64x2 $1, %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x38,0xd1,0x01]
+; CHECK-NEXT: vinserti64x2 $1, %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x38,0xd9,0x01]
+; CHECK-NEXT: vinserti64x2 $1, %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x38,0xc1,0x01]
+; CHECK-NEXT: vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
+; CHECK-NEXT: vpaddq %ymm3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xd4,0xc3]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.mask.inserti64x2.256(<4 x i64> %x0, <2 x i64> %x1, i32 1, <4 x i64> %x3, i8 %x4)
%res1 = call <4 x i64> @llvm.x86.avx512.mask.inserti64x2.256(<4 x i64> %x0, <2 x i64> %x1, i32 1, <4 x i64> %x3, i8 -1)
%res2 = call <4 x i64> @llvm.x86.avx512.mask.inserti64x2.256(<4 x i64> %x0, <2 x i64> %x1, i32 1, <4 x i64> zeroinitializer, i8 %x4)
@@ -1711,14 +2134,16 @@ define <4 x i64>@test_int_x86_avx512_mas
declare i8 @llvm.x86.avx512.mask.fpclass.ps.128(<4 x float>, i32, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_fpclass_ps_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vfpclassps
-; CHECK: {%k1}
-; CHECK: vfpclassps
-; CHECK: kmovb %k0
define i8 @test_int_x86_avx512_mask_fpclass_ps_128(<4 x float> %x0, i8 %x1) {
+; CHECK-LABEL: test_int_x86_avx512_mask_fpclass_ps_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vfpclassps $2, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x66,0xc0,0x02]
+; CHECK-NEXT: kmovb %k0, %ecx ## encoding: [0xc5,0xf9,0x93,0xc8]
+; CHECK-NEXT: vfpclassps $4, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x66,0xc0,0x04]
+; CHECK-NEXT: kmovb %k0, %eax ## encoding: [0xc5,0xf9,0x93,0xc0]
+; CHECK-NEXT: addb %cl, %al ## encoding: [0x00,0xc8]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.mask.fpclass.ps.128(<4 x float> %x0, i32 2, i8 %x1)
%res1 = call i8 @llvm.x86.avx512.mask.fpclass.ps.128(<4 x float> %x0, i32 4, i8 -1)
%res2 = add i8 %res, %res1
@@ -1727,14 +2152,16 @@ define i8 @test_int_x86_avx512_mask_fpcl
declare i8 @llvm.x86.avx512.mask.fpclass.ps.256(<8 x float>, i32, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_fpclass_ps_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vfpclassps
-; CHECK: {%k1}
-; CHECK: vfpclassps
-; CHECK: kmovb %k0
define i8 @test_int_x86_avx512_mask_fpclass_ps_256(<8 x float> %x0, i8 %x1) {
+; CHECK-LABEL: test_int_x86_avx512_mask_fpclass_ps_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vfpclassps $2, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x66,0xc0,0x02]
+; CHECK-NEXT: kmovb %k0, %ecx ## encoding: [0xc5,0xf9,0x93,0xc8]
+; CHECK-NEXT: vfpclassps $4, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x66,0xc0,0x04]
+; CHECK-NEXT: kmovb %k0, %eax ## encoding: [0xc5,0xf9,0x93,0xc0]
+; CHECK-NEXT: addb %cl, %al ## encoding: [0x00,0xc8]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.mask.fpclass.ps.256(<8 x float> %x0, i32 2, i8 %x1)
%res1 = call i8 @llvm.x86.avx512.mask.fpclass.ps.256(<8 x float> %x0, i32 4, i8 -1)
%res2 = add i8 %res, %res1
@@ -1743,14 +2170,16 @@ define i8 @test_int_x86_avx512_mask_fpcl
declare i8 @llvm.x86.avx512.mask.fpclass.pd.128(<2 x double>, i32, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_fpclass_pd_128
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vfpclasspd
-; CHECK: {%k1}
-; CHECK: vfpclasspd
-; CHECK: kmovb %k0
define i8 @test_int_x86_avx512_mask_fpclass_pd_128(<2 x double> %x0, i8 %x1) {
+; CHECK-LABEL: test_int_x86_avx512_mask_fpclass_pd_128:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vfpclasspd $4, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x66,0xc0,0x04]
+; CHECK-NEXT: kmovb %k0, %ecx ## encoding: [0xc5,0xf9,0x93,0xc8]
+; CHECK-NEXT: vfpclasspd $2, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x66,0xc0,0x02]
+; CHECK-NEXT: kmovb %k0, %eax ## encoding: [0xc5,0xf9,0x93,0xc0]
+; CHECK-NEXT: addb %cl, %al ## encoding: [0x00,0xc8]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.mask.fpclass.pd.128(<2 x double> %x0, i32 4, i8 %x1)
%res1 = call i8 @llvm.x86.avx512.mask.fpclass.pd.128(<2 x double> %x0, i32 2, i8 -1)
%res2 = add i8 %res, %res1
@@ -1759,14 +2188,16 @@ define i8 @test_int_x86_avx512_mask_fpcl
declare i8 @llvm.x86.avx512.mask.fpclass.pd.256(<4 x double>, i32, i8)
-; CHECK-LABEL: @test_int_x86_avx512_mask_fpclass_pd_256
-; CHECK-NOT: call
-; CHECK: kmov
-; CHECK: vfpclasspd
-; CHECK: {%k1}
-; CHECK: vfpclasspd
-; CHECK: kmovb %k0
define i8 @test_int_x86_avx512_mask_fpclass_pd_256(<4 x double> %x0, i8 %x1) {
+; CHECK-LABEL: test_int_x86_avx512_mask_fpclass_pd_256:
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vfpclasspd $2, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x66,0xc0,0x02]
+; CHECK-NEXT: kmovb %k0, %ecx ## encoding: [0xc5,0xf9,0x93,0xc8]
+; CHECK-NEXT: vfpclasspd $4, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x66,0xc0,0x04]
+; CHECK-NEXT: kmovb %k0, %eax ## encoding: [0xc5,0xf9,0x93,0xc0]
+; CHECK-NEXT: addb %cl, %al ## encoding: [0x00,0xc8]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.mask.fpclass.pd.256(<4 x double> %x0, i32 2, i8 %x1)
%res1 = call i8 @llvm.x86.avx512.mask.fpclass.pd.256(<4 x double> %x0, i32 4, i8 -1)
%res2 = add i8 %res, %res1
@@ -1778,13 +2209,13 @@ declare <8 x float> @llvm.x86.avx512.mas
define <8 x float>@test_int_x86_avx512_mask_broadcastf32x2_256(<4 x float> %x0, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x2_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vbroadcastf32x2 %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vbroadcastf32x2 %xmm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vbroadcastf32x2 %xmm0, %ymm0
-; CHECK-NEXT: vaddps %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vaddps %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vbroadcastf32x2 %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x19,0xc8]
+; CHECK-NEXT: vbroadcastf32x2 %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x19,0xd0]
+; CHECK-NEXT: vbroadcastf32x2 %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x19,0xc0]
+; CHECK-NEXT: vaddps %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x74,0x28,0x58,0xca]
+; CHECK-NEXT: vaddps %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x74,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x2.256(<4 x float> %x0, <8 x float> %x2, i8 %x3)
%res1 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x2.256(<4 x float> %x0, <8 x float> zeroinitializer, i8 %x3)
%res2 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x2.256(<4 x float> %x0, <8 x float> %x2, i8 -1)
@@ -1798,13 +2229,13 @@ declare <8 x i32> @llvm.x86.avx512.mask.
define <8 x i32>@test_int_x86_avx512_mask_broadcasti32x2_256(<4 x i32> %x0, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x2_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vbroadcasti32x2 %xmm0, %ymm1 {%k1}
-; CHECK-NEXT: vbroadcasti32x2 %xmm0, %ymm2 {%k1} {z}
-; CHECK-NEXT: vbroadcasti32x2 %xmm0, %ymm0
-; CHECK-NEXT: vpaddd %ymm2, %ymm1, %ymm1
-; CHECK-NEXT: vpaddd %ymm0, %ymm1, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vbroadcasti32x2 %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x59,0xc8]
+; CHECK-NEXT: vbroadcasti32x2 %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x59,0xd0]
+; CHECK-NEXT: vbroadcasti32x2 %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x59,0xc0]
+; CHECK-NEXT: vpaddd %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x75,0x28,0xfe,0xca]
+; CHECK-NEXT: vpaddd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfe,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x2.256(<4 x i32> %x0, <8 x i32> %x2, i8 %x3)
%res1 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x2.256(<4 x i32> %x0, <8 x i32> zeroinitializer, i8 %x3)
%res2 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x2.256(<4 x i32> %x0, <8 x i32> %x2, i8 -1)
@@ -1818,13 +2249,13 @@ declare <4 x i32> @llvm.x86.avx512.mask.
define <4 x i32>@test_int_x86_avx512_mask_broadcasti32x2_128(<4 x i32> %x0, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x2_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k1
-; CHECK-NEXT: vbroadcasti32x2 %xmm0, %xmm1 {%k1}
-; CHECK-NEXT: vbroadcasti32x2 %xmm0, %xmm2 {%k1} {z}
-; CHECK-NEXT: vbroadcasti32x2 %xmm0, %xmm0
-; CHECK-NEXT: vpaddd %xmm2, %xmm1, %xmm1
-; CHECK-NEXT: vpaddd %xmm0, %xmm1, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vbroadcasti32x2 %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x59,0xc8]
+; CHECK-NEXT: vbroadcasti32x2 %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x59,0xd0]
+; CHECK-NEXT: vbroadcasti32x2 %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x59,0xc0]
+; CHECK-NEXT: vpaddd %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0x75,0x08,0xfe,0xca]
+; CHECK-NEXT: vpaddd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfe,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i32> @llvm.x86.avx512.mask.broadcasti32x2.128(<4 x i32> %x0, <4 x i32> %x2, i8 %x3)
%res1 = call <4 x i32> @llvm.x86.avx512.mask.broadcasti32x2.128(<4 x i32> %x0, <4 x i32> zeroinitializer, i8 %x3)
%res2 = call <4 x i32> @llvm.x86.avx512.mask.broadcasti32x2.128(<4 x i32> %x0, <4 x i32> %x2, i8 -1)
@@ -1838,9 +2269,9 @@ declare i8 @llvm.x86.avx512.cvtd2mask.12
define i8 at test_int_x86_avx512_cvtd2mask_128(<4 x i32> %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtd2mask_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpmovd2m %xmm0, %k0
-; CHECK-NEXT: kmovb %k0, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpmovd2m %xmm0, %k0 ## encoding: [0x62,0xf2,0x7e,0x08,0x39,0xc0]
+; CHECK-NEXT: kmovb %k0, %eax ## encoding: [0xc5,0xf9,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.cvtd2mask.128(<4 x i32> %x0)
ret i8 %res
}
@@ -1850,9 +2281,9 @@ declare i8 @llvm.x86.avx512.cvtd2mask.25
define i8 at test_int_x86_avx512_cvtd2mask_256(<8 x i32> %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtd2mask_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpmovd2m %ymm0, %k0
-; CHECK-NEXT: kmovb %k0, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpmovd2m %ymm0, %k0 ## encoding: [0x62,0xf2,0x7e,0x28,0x39,0xc0]
+; CHECK-NEXT: kmovb %k0, %eax ## encoding: [0xc5,0xf9,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.cvtd2mask.256(<8 x i32> %x0)
ret i8 %res
}
@@ -1862,9 +2293,9 @@ declare i8 @llvm.x86.avx512.cvtq2mask.12
define i8 at test_int_x86_avx512_cvtq2mask_128(<2 x i64> %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtq2mask_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpmovq2m %xmm0, %k0
-; CHECK-NEXT: kmovb %k0, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpmovq2m %xmm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x08,0x39,0xc0]
+; CHECK-NEXT: kmovb %k0, %eax ## encoding: [0xc5,0xf9,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.cvtq2mask.128(<2 x i64> %x0)
ret i8 %res
}
@@ -1874,9 +2305,9 @@ declare i8 @llvm.x86.avx512.cvtq2mask.25
define i8 at test_int_x86_avx512_cvtq2mask_256(<4 x i64> %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtq2mask_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: vpmovq2m %ymm0, %k0
-; CHECK-NEXT: kmovb %k0, %eax
-; CHECK-NEXT: retq
+; CHECK-NEXT: vpmovq2m %ymm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x28,0x39,0xc0]
+; CHECK-NEXT: kmovb %k0, %eax ## encoding: [0xc5,0xf9,0x93,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call i8 @llvm.x86.avx512.cvtq2mask.256(<4 x i64> %x0)
ret i8 %res
}
@@ -1886,9 +2317,9 @@ declare <4 x i32> @llvm.x86.avx512.cvtma
define <4 x i32>@test_int_x86_avx512_cvtmask2d_128(i8 %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtmask2d_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k0
-; CHECK-NEXT: vpmovm2d %k0, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k0 ## encoding: [0xc5,0xf9,0x92,0xc7]
+; CHECK-NEXT: vpmovm2d %k0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x38,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i32> @llvm.x86.avx512.cvtmask2d.128(i8 %x0)
ret <4 x i32> %res
}
@@ -1898,9 +2329,9 @@ declare <8 x i32> @llvm.x86.avx512.cvtma
define <8 x i32>@test_int_x86_avx512_cvtmask2d_256(i8 %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtmask2d_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k0
-; CHECK-NEXT: vpmovm2d %k0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k0 ## encoding: [0xc5,0xf9,0x92,0xc7]
+; CHECK-NEXT: vpmovm2d %k0, %ymm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x38,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <8 x i32> @llvm.x86.avx512.cvtmask2d.256(i8 %x0)
ret <8 x i32> %res
}
@@ -1910,9 +2341,9 @@ declare <2 x i64> @llvm.x86.avx512.cvtma
define <2 x i64>@test_int_x86_avx512_cvtmask2q_128(i8 %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtmask2q_128:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k0
-; CHECK-NEXT: vpmovm2q %k0, %xmm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k0 ## encoding: [0xc5,0xf9,0x92,0xc7]
+; CHECK-NEXT: vpmovm2q %k0, %xmm0 ## encoding: [0x62,0xf2,0xfe,0x08,0x38,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <2 x i64> @llvm.x86.avx512.cvtmask2q.128(i8 %x0)
ret <2 x i64> %res
}
@@ -1922,9 +2353,9 @@ declare <4 x i64> @llvm.x86.avx512.cvtma
define <4 x i64>@test_int_x86_avx512_cvtmask2q_256(i8 %x0) {
; CHECK-LABEL: test_int_x86_avx512_cvtmask2q_256:
; CHECK: ## BB#0:
-; CHECK-NEXT: kmovb %edi, %k0
-; CHECK-NEXT: vpmovm2q %k0, %ymm0
-; CHECK-NEXT: retq
+; CHECK-NEXT: kmovb %edi, %k0 ## encoding: [0xc5,0xf9,0x92,0xc7]
+; CHECK-NEXT: vpmovm2q %k0, %ymm0 ## encoding: [0x62,0xf2,0xfe,0x28,0x38,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res = call <4 x i64> @llvm.x86.avx512.cvtmask2q.256(i8 %x0)
ret <4 x i64> %res
}
@@ -1932,12 +2363,17 @@ declare <4 x double> @llvm.x86.avx512.ma
define <4 x double>@test_int_x86_avx512_mask_broadcastf64x2_256(<2 x double> %x0, <4 x double> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf64x2_256:
-; CHECK: kmovb %edi, %k1
-; CHECK: vshuff64x2 $0, %ymm0, %ymm0, %ymm2 {%k1} {z}
-; CHECK: vshuff64x2 $0, %ymm0, %ymm0, %ymm1 {%k1}
-; CHECK: vshuff64x2 $0, %ymm0, %ymm0, %ymm0
-; CHECK: vaddpd %ymm1, %ymm0, %ymm0
-; CHECK: vaddpd %ymm0, %ymm2, %ymm0
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vshuff64x2 $0, %ymm0, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x23,0xd0,0x00]
+; CHECK-NEXT: ## ymm2 = ymm0[0,1,0,1]
+; CHECK-NEXT: vshuff64x2 $0, %ymm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x23,0xc8,0x00]
+; CHECK-NEXT: ## ymm1 = ymm0[0,1,0,1]
+; CHECK-NEXT: vshuff64x2 $0, %ymm0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x23,0xc0,0x00]
+; CHECK-NEXT: ## ymm0 = ymm0[0,1,0,1]
+; CHECK-NEXT: vaddpd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x58,0xc1]
+; CHECK-NEXT: vaddpd %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0x58,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res1 = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> %x2, i8 -1)
%res2 = call <4 x double> @llvm.x86.avx512.mask.broadcastf64x2.256(<2 x double> %x0, <4 x double> %x2, i8 %mask)
@@ -1951,12 +2387,17 @@ declare <4 x i64> @llvm.x86.avx512.mask.
define <4 x i64>@test_int_x86_avx512_mask_broadcasti64x2_256(<2 x i64> %x0, <4 x i64> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti64x2_256:
-; CHECK: kmovb %edi, %k1
-; CHECK: vshufi64x2 $0, %ymm0, %ymm0, %ymm2 {%k1} {z}
-; CHECK: vshufi64x2 $0, %ymm0, %ymm0, %ymm1 {%k1}
-; CHECK: vshufi64x2 $0, %ymm0, %ymm0, %ymm0
-; CHECK: vpaddq %ymm1, %ymm0, %ymm0
-; CHECK: vpaddq %ymm0, %ymm2, %ymm0
+; CHECK: ## BB#0:
+; CHECK-NEXT: kmovb %edi, %k1 ## encoding: [0xc5,0xf9,0x92,0xcf]
+; CHECK-NEXT: vshufi64x2 $0, %ymm0, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x43,0xd0,0x00]
+; CHECK-NEXT: ## ymm2 = ymm0[0,1,0,1]
+; CHECK-NEXT: vshufi64x2 $0, %ymm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x43,0xc8,0x00]
+; CHECK-NEXT: ## ymm1 = ymm0[0,1,0,1]
+; CHECK-NEXT: vshufi64x2 $0, %ymm0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x43,0xc0,0x00]
+; CHECK-NEXT: ## ymm0 = ymm0[0,1,0,1]
+; CHECK-NEXT: vpaddq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xd4,0xc1]
+; CHECK-NEXT: vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
+; CHECK-NEXT: retq ## encoding: [0xc3]
%res1 = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> %x2, i8 -1)
%res2 = call <4 x i64> @llvm.x86.avx512.mask.broadcasti64x2.256(<2 x i64> %x0, <4 x i64> %x2, i8 %mask)
More information about the llvm-commits
mailing list