test/CodeGen/X86/avx512vl-intrinsics.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=knl -mattr=+avx512vl --show-mc-encoding| FileCheck %s

; 256-bit

define <8 x i8> @test_cmp_d_256(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: test_cmp_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpeqd %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x1f,0xc1,0x00]
; CHECK-NEXT:    vpcmpltd %ymm1, %ymm0, %k1 ## encoding: [0x62,0xf3,0x7d,0x28,0x1f,0xc9,0x01]
; CHECK-NEXT:    vpcmpled %ymm1, %ymm0, %k2 ## encoding: [0x62,0xf3,0x7d,0x28,0x1f,0xd1,0x02]
; CHECK-NEXT:    vpcmpunordd %ymm1, %ymm0, %k3 ## encoding: [0x62,0xf3,0x7d,0x28,0x1f,0xd9,0x03]
; CHECK-NEXT:    vpcmpneqd %ymm1, %ymm0, %k4 ## encoding: [0x62,0xf3,0x7d,0x28,0x1f,0xe1,0x04]
; CHECK-NEXT:    vpcmpnltd %ymm1, %ymm0, %k5 ## encoding: [0x62,0xf3,0x7d,0x28,0x1f,0xe9,0x05]
; CHECK-NEXT:    vpcmpnled %ymm1, %ymm0, %k6 ## encoding: [0x62,0xf3,0x7d,0x28,0x1f,0xf1,0x06]
; CHECK-NEXT:    vpcmpordd %ymm1, %ymm0, %k7 ## encoding: [0x62,0xf3,0x7d,0x28,0x1f,0xf9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 0, i8 -1)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 1, i8 -1)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 2, i8 -1)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 3, i8 -1)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 4, i8 -1)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 5, i8 -1)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 6, i8 -1)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %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_d_256(<8 x i32> %a0, <8 x i32> %a1, i8 %mask) {
; CHECK-LABEL: test_mask_cmp_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpeqd %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1f,0xc1,0x00]
; CHECK-NEXT:    vpcmpltd %ymm1, %ymm0, %k2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1f,0xd1,0x01]
; CHECK-NEXT:    vpcmpled %ymm1, %ymm0, %k3 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1f,0xd9,0x02]
; CHECK-NEXT:    vpcmpunordd %ymm1, %ymm0, %k4 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1f,0xe1,0x03]
; CHECK-NEXT:    vpcmpneqd %ymm1, %ymm0, %k5 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1f,0xe9,0x04]
; CHECK-NEXT:    vpcmpnltd %ymm1, %ymm0, %k6 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1f,0xf1,0x05]
; CHECK-NEXT:    vpcmpnled %ymm1, %ymm0, %k7 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1f,0xf9,0x06]
; CHECK-NEXT:    vpcmpordd %ymm1, %ymm0, %k1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1f,0xc9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 0, i8 %mask)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 1, i8 %mask)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 2, i8 %mask)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 3, i8 %mask)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 4, i8 %mask)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 5, i8 %mask)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 6, i8 %mask)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 7, i8 %mask)
  %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
  ret <8 x i8> %vec7
}

declare i8 @llvm.x86.avx512.mask.cmp.d.256(<8 x i32>, <8 x i32>, i32, i8) nounwind readnone

define <8 x i8> @test_ucmp_d_256(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: test_ucmp_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpequd %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x28,0x1e,0xc1,0x00]
; CHECK-NEXT:    vpcmpltud %ymm1, %ymm0, %k1 ## encoding: [0x62,0xf3,0x7d,0x28,0x1e,0xc9,0x01]
; CHECK-NEXT:    vpcmpleud %ymm1, %ymm0, %k2 ## encoding: [0x62,0xf3,0x7d,0x28,0x1e,0xd1,0x02]
; CHECK-NEXT:    vpcmpunordud %ymm1, %ymm0, %k3 ## encoding: [0x62,0xf3,0x7d,0x28,0x1e,0xd9,0x03]
; CHECK-NEXT:    vpcmpnequd %ymm1, %ymm0, %k4 ## encoding: [0x62,0xf3,0x7d,0x28,0x1e,0xe1,0x04]
; CHECK-NEXT:    vpcmpnltud %ymm1, %ymm0, %k5 ## encoding: [0x62,0xf3,0x7d,0x28,0x1e,0xe9,0x05]
; CHECK-NEXT:    vpcmpnleud %ymm1, %ymm0, %k6 ## encoding: [0x62,0xf3,0x7d,0x28,0x1e,0xf1,0x06]
; CHECK-NEXT:    vpcmpordud %ymm1, %ymm0, %k7 ## encoding: [0x62,0xf3,0x7d,0x28,0x1e,0xf9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 0, i8 -1)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 1, i8 -1)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 2, i8 -1)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 3, i8 -1)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 4, i8 -1)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 5, i8 -1)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 6, i8 -1)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %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_d_256(<8 x i32> %a0, <8 x i32> %a1, i8 %mask) {
; CHECK-LABEL: test_mask_ucmp_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpequd %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1e,0xc1,0x00]
; CHECK-NEXT:    vpcmpltud %ymm1, %ymm0, %k2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1e,0xd1,0x01]
; CHECK-NEXT:    vpcmpleud %ymm1, %ymm0, %k3 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1e,0xd9,0x02]
; CHECK-NEXT:    vpcmpunordud %ymm1, %ymm0, %k4 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1e,0xe1,0x03]
; CHECK-NEXT:    vpcmpnequd %ymm1, %ymm0, %k5 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1e,0xe9,0x04]
; CHECK-NEXT:    vpcmpnltud %ymm1, %ymm0, %k6 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1e,0xf1,0x05]
; CHECK-NEXT:    vpcmpnleud %ymm1, %ymm0, %k7 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1e,0xf9,0x06]
; CHECK-NEXT:    vpcmpordud %ymm1, %ymm0, %k1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1e,0xc9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 0, i8 %mask)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 1, i8 %mask)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 2, i8 %mask)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 3, i8 %mask)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 4, i8 %mask)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 5, i8 %mask)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 6, i8 %mask)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32> %a0, <8 x i32> %a1, i32 7, i8 %mask)
  %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
  ret <8 x i8> %vec7
}

declare i8 @llvm.x86.avx512.mask.ucmp.d.256(<8 x i32>, <8 x i32>, i32, i8) nounwind readnone

define <8 x i8> @test_cmp_q_256(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: test_cmp_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpeqq %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x1f,0xc1,0x00]
; CHECK-NEXT:    vpcmpltq %ymm1, %ymm0, %k1 ## encoding: [0x62,0xf3,0xfd,0x28,0x1f,0xc9,0x01]
; CHECK-NEXT:    vpcmpleq %ymm1, %ymm0, %k2 ## encoding: [0x62,0xf3,0xfd,0x28,0x1f,0xd1,0x02]
; CHECK-NEXT:    vpcmpunordq %ymm1, %ymm0, %k3 ## encoding: [0x62,0xf3,0xfd,0x28,0x1f,0xd9,0x03]
; CHECK-NEXT:    vpcmpneqq %ymm1, %ymm0, %k4 ## encoding: [0x62,0xf3,0xfd,0x28,0x1f,0xe1,0x04]
; CHECK-NEXT:    vpcmpnltq %ymm1, %ymm0, %k5 ## encoding: [0x62,0xf3,0xfd,0x28,0x1f,0xe9,0x05]
; CHECK-NEXT:    vpcmpnleq %ymm1, %ymm0, %k6 ## encoding: [0x62,0xf3,0xfd,0x28,0x1f,0xf1,0x06]
; CHECK-NEXT:    vpcmpordq %ymm1, %ymm0, %k7 ## encoding: [0x62,0xf3,0xfd,0x28,0x1f,0xf9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 0, i8 -1)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 1, i8 -1)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 2, i8 -1)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 3, i8 -1)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 4, i8 -1)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 5, i8 -1)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 6, i8 -1)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %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_q_256(<4 x i64> %a0, <4 x i64> %a1, i8 %mask) {
; CHECK-LABEL: test_mask_cmp_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpeqq %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1f,0xc1,0x00]
; CHECK-NEXT:    vpcmpltq %ymm1, %ymm0, %k2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1f,0xd1,0x01]
; CHECK-NEXT:    vpcmpleq %ymm1, %ymm0, %k3 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1f,0xd9,0x02]
; CHECK-NEXT:    vpcmpunordq %ymm1, %ymm0, %k4 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1f,0xe1,0x03]
; CHECK-NEXT:    vpcmpneqq %ymm1, %ymm0, %k5 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1f,0xe9,0x04]
; CHECK-NEXT:    vpcmpnltq %ymm1, %ymm0, %k6 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1f,0xf1,0x05]
; CHECK-NEXT:    vpcmpnleq %ymm1, %ymm0, %k7 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1f,0xf9,0x06]
; CHECK-NEXT:    vpcmpordq %ymm1, %ymm0, %k1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1f,0xc9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 0, i8 %mask)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 1, i8 %mask)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 2, i8 %mask)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 3, i8 %mask)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 4, i8 %mask)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 5, i8 %mask)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 6, i8 %mask)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 7, i8 %mask)
  %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
  ret <8 x i8> %vec7
}

declare i8 @llvm.x86.avx512.mask.cmp.q.256(<4 x i64>, <4 x i64>, i32, i8) nounwind readnone

define <8 x i8> @test_ucmp_q_256(<4 x i64> %a0, <4 x i64> %a1) {
; CHECK-LABEL: test_ucmp_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpequq %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x28,0x1e,0xc1,0x00]
; CHECK-NEXT:    vpcmpltuq %ymm1, %ymm0, %k1 ## encoding: [0x62,0xf3,0xfd,0x28,0x1e,0xc9,0x01]
; CHECK-NEXT:    vpcmpleuq %ymm1, %ymm0, %k2 ## encoding: [0x62,0xf3,0xfd,0x28,0x1e,0xd1,0x02]
; CHECK-NEXT:    vpcmpunorduq %ymm1, %ymm0, %k3 ## encoding: [0x62,0xf3,0xfd,0x28,0x1e,0xd9,0x03]
; CHECK-NEXT:    vpcmpnequq %ymm1, %ymm0, %k4 ## encoding: [0x62,0xf3,0xfd,0x28,0x1e,0xe1,0x04]
; CHECK-NEXT:    vpcmpnltuq %ymm1, %ymm0, %k5 ## encoding: [0x62,0xf3,0xfd,0x28,0x1e,0xe9,0x05]
; CHECK-NEXT:    vpcmpnleuq %ymm1, %ymm0, %k6 ## encoding: [0x62,0xf3,0xfd,0x28,0x1e,0xf1,0x06]
; CHECK-NEXT:    vpcmporduq %ymm1, %ymm0, %k7 ## encoding: [0x62,0xf3,0xfd,0x28,0x1e,0xf9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 0, i8 -1)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 1, i8 -1)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 2, i8 -1)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 3, i8 -1)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 4, i8 -1)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 5, i8 -1)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 6, i8 -1)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %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_q_256(<4 x i64> %a0, <4 x i64> %a1, i8 %mask) {
; CHECK-LABEL: test_mask_ucmp_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpequq %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1e,0xc1,0x00]
; CHECK-NEXT:    vpcmpltuq %ymm1, %ymm0, %k2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1e,0xd1,0x01]
; CHECK-NEXT:    vpcmpleuq %ymm1, %ymm0, %k3 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1e,0xd9,0x02]
; CHECK-NEXT:    vpcmpunorduq %ymm1, %ymm0, %k4 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1e,0xe1,0x03]
; CHECK-NEXT:    vpcmpnequq %ymm1, %ymm0, %k5 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1e,0xe9,0x04]
; CHECK-NEXT:    vpcmpnltuq %ymm1, %ymm0, %k6 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1e,0xf1,0x05]
; CHECK-NEXT:    vpcmpnleuq %ymm1, %ymm0, %k7 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1e,0xf9,0x06]
; CHECK-NEXT:    vpcmporduq %ymm1, %ymm0, %k1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x1e,0xc9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 0, i8 %mask)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 1, i8 %mask)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 2, i8 %mask)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 3, i8 %mask)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 4, i8 %mask)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 5, i8 %mask)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 6, i8 %mask)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64> %a0, <4 x i64> %a1, i32 7, i8 %mask)
  %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
  ret <8 x i8> %vec7
}

declare i8 @llvm.x86.avx512.mask.ucmp.q.256(<4 x i64>, <4 x i64>, i32, i8) nounwind readnone

; 128-bit

define <8 x i8> @test_cmp_d_128(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: test_cmp_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpeqd %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x1f,0xc1,0x00]
; CHECK-NEXT:    vpcmpltd %xmm1, %xmm0, %k1 ## encoding: [0x62,0xf3,0x7d,0x08,0x1f,0xc9,0x01]
; CHECK-NEXT:    vpcmpled %xmm1, %xmm0, %k2 ## encoding: [0x62,0xf3,0x7d,0x08,0x1f,0xd1,0x02]
; CHECK-NEXT:    vpcmpunordd %xmm1, %xmm0, %k3 ## encoding: [0x62,0xf3,0x7d,0x08,0x1f,0xd9,0x03]
; CHECK-NEXT:    vpcmpneqd %xmm1, %xmm0, %k4 ## encoding: [0x62,0xf3,0x7d,0x08,0x1f,0xe1,0x04]
; CHECK-NEXT:    vpcmpnltd %xmm1, %xmm0, %k5 ## encoding: [0x62,0xf3,0x7d,0x08,0x1f,0xe9,0x05]
; CHECK-NEXT:    vpcmpnled %xmm1, %xmm0, %k6 ## encoding: [0x62,0xf3,0x7d,0x08,0x1f,0xf1,0x06]
; CHECK-NEXT:    vpcmpordd %xmm1, %xmm0, %k7 ## encoding: [0x62,0xf3,0x7d,0x08,0x1f,0xf9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 0, i8 -1)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 1, i8 -1)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 2, i8 -1)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 3, i8 -1)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 4, i8 -1)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 5, i8 -1)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 6, i8 -1)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %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_d_128(<4 x i32> %a0, <4 x i32> %a1, i8 %mask) {
; CHECK-LABEL: test_mask_cmp_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpeqd %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1f,0xc1,0x00]
; CHECK-NEXT:    vpcmpltd %xmm1, %xmm0, %k2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1f,0xd1,0x01]
; CHECK-NEXT:    vpcmpled %xmm1, %xmm0, %k3 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1f,0xd9,0x02]
; CHECK-NEXT:    vpcmpunordd %xmm1, %xmm0, %k4 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1f,0xe1,0x03]
; CHECK-NEXT:    vpcmpneqd %xmm1, %xmm0, %k5 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1f,0xe9,0x04]
; CHECK-NEXT:    vpcmpnltd %xmm1, %xmm0, %k6 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1f,0xf1,0x05]
; CHECK-NEXT:    vpcmpnled %xmm1, %xmm0, %k7 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1f,0xf9,0x06]
; CHECK-NEXT:    vpcmpordd %xmm1, %xmm0, %k1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1f,0xc9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 0, i8 %mask)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 1, i8 %mask)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 2, i8 %mask)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 3, i8 %mask)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 4, i8 %mask)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 5, i8 %mask)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 6, i8 %mask)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 7, i8 %mask)
  %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
  ret <8 x i8> %vec7
}

declare i8 @llvm.x86.avx512.mask.cmp.d.128(<4 x i32>, <4 x i32>, i32, i8) nounwind readnone

define <8 x i8> @test_ucmp_d_128(<4 x i32> %a0, <4 x i32> %a1) {
; CHECK-LABEL: test_ucmp_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpequd %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0x7d,0x08,0x1e,0xc1,0x00]
; CHECK-NEXT:    vpcmpltud %xmm1, %xmm0, %k1 ## encoding: [0x62,0xf3,0x7d,0x08,0x1e,0xc9,0x01]
; CHECK-NEXT:    vpcmpleud %xmm1, %xmm0, %k2 ## encoding: [0x62,0xf3,0x7d,0x08,0x1e,0xd1,0x02]
; CHECK-NEXT:    vpcmpunordud %xmm1, %xmm0, %k3 ## encoding: [0x62,0xf3,0x7d,0x08,0x1e,0xd9,0x03]
; CHECK-NEXT:    vpcmpnequd %xmm1, %xmm0, %k4 ## encoding: [0x62,0xf3,0x7d,0x08,0x1e,0xe1,0x04]
; CHECK-NEXT:    vpcmpnltud %xmm1, %xmm0, %k5 ## encoding: [0x62,0xf3,0x7d,0x08,0x1e,0xe9,0x05]
; CHECK-NEXT:    vpcmpnleud %xmm1, %xmm0, %k6 ## encoding: [0x62,0xf3,0x7d,0x08,0x1e,0xf1,0x06]
; CHECK-NEXT:    vpcmpordud %xmm1, %xmm0, %k7 ## encoding: [0x62,0xf3,0x7d,0x08,0x1e,0xf9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 0, i8 -1)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 1, i8 -1)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 2, i8 -1)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 3, i8 -1)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 4, i8 -1)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 5, i8 -1)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 6, i8 -1)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %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_d_128(<4 x i32> %a0, <4 x i32> %a1, i8 %mask) {
; CHECK-LABEL: test_mask_ucmp_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpequd %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1e,0xc1,0x00]
; CHECK-NEXT:    vpcmpltud %xmm1, %xmm0, %k2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1e,0xd1,0x01]
; CHECK-NEXT:    vpcmpleud %xmm1, %xmm0, %k3 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1e,0xd9,0x02]
; CHECK-NEXT:    vpcmpunordud %xmm1, %xmm0, %k4 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1e,0xe1,0x03]
; CHECK-NEXT:    vpcmpnequd %xmm1, %xmm0, %k5 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1e,0xe9,0x04]
; CHECK-NEXT:    vpcmpnltud %xmm1, %xmm0, %k6 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1e,0xf1,0x05]
; CHECK-NEXT:    vpcmpnleud %xmm1, %xmm0, %k7 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1e,0xf9,0x06]
; CHECK-NEXT:    vpcmpordud %xmm1, %xmm0, %k1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1e,0xc9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 0, i8 %mask)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 1, i8 %mask)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 2, i8 %mask)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 3, i8 %mask)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 4, i8 %mask)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 5, i8 %mask)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 6, i8 %mask)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32> %a0, <4 x i32> %a1, i32 7, i8 %mask)
  %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
  ret <8 x i8> %vec7
}

declare i8 @llvm.x86.avx512.mask.ucmp.d.128(<4 x i32>, <4 x i32>, i32, i8) nounwind readnone

define <8 x i8> @test_cmp_q_128(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: test_cmp_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpeqq %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x1f,0xc1,0x00]
; CHECK-NEXT:    vpcmpltq %xmm1, %xmm0, %k1 ## encoding: [0x62,0xf3,0xfd,0x08,0x1f,0xc9,0x01]
; CHECK-NEXT:    vpcmpleq %xmm1, %xmm0, %k2 ## encoding: [0x62,0xf3,0xfd,0x08,0x1f,0xd1,0x02]
; CHECK-NEXT:    vpcmpunordq %xmm1, %xmm0, %k3 ## encoding: [0x62,0xf3,0xfd,0x08,0x1f,0xd9,0x03]
; CHECK-NEXT:    vpcmpneqq %xmm1, %xmm0, %k4 ## encoding: [0x62,0xf3,0xfd,0x08,0x1f,0xe1,0x04]
; CHECK-NEXT:    vpcmpnltq %xmm1, %xmm0, %k5 ## encoding: [0x62,0xf3,0xfd,0x08,0x1f,0xe9,0x05]
; CHECK-NEXT:    vpcmpnleq %xmm1, %xmm0, %k6 ## encoding: [0x62,0xf3,0xfd,0x08,0x1f,0xf1,0x06]
; CHECK-NEXT:    vpcmpordq %xmm1, %xmm0, %k7 ## encoding: [0x62,0xf3,0xfd,0x08,0x1f,0xf9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 0, i8 -1)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 1, i8 -1)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 2, i8 -1)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 3, i8 -1)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 4, i8 -1)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 5, i8 -1)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 6, i8 -1)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %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_q_128(<2 x i64> %a0, <2 x i64> %a1, i8 %mask) {
; CHECK-LABEL: test_mask_cmp_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpeqq %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1f,0xc1,0x00]
; CHECK-NEXT:    vpcmpltq %xmm1, %xmm0, %k2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1f,0xd1,0x01]
; CHECK-NEXT:    vpcmpleq %xmm1, %xmm0, %k3 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1f,0xd9,0x02]
; CHECK-NEXT:    vpcmpunordq %xmm1, %xmm0, %k4 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1f,0xe1,0x03]
; CHECK-NEXT:    vpcmpneqq %xmm1, %xmm0, %k5 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1f,0xe9,0x04]
; CHECK-NEXT:    vpcmpnltq %xmm1, %xmm0, %k6 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1f,0xf1,0x05]
; CHECK-NEXT:    vpcmpnleq %xmm1, %xmm0, %k7 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1f,0xf9,0x06]
; CHECK-NEXT:    vpcmpordq %xmm1, %xmm0, %k1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1f,0xc9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 0, i8 %mask)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 1, i8 %mask)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 2, i8 %mask)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 3, i8 %mask)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 4, i8 %mask)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 5, i8 %mask)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 6, i8 %mask)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 7, i8 %mask)
  %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
  ret <8 x i8> %vec7
}

declare i8 @llvm.x86.avx512.mask.cmp.q.128(<2 x i64>, <2 x i64>, i32, i8) nounwind readnone

define <8 x i8> @test_ucmp_q_128(<2 x i64> %a0, <2 x i64> %a1) {
; CHECK-LABEL: test_ucmp_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpequq %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf3,0xfd,0x08,0x1e,0xc1,0x00]
; CHECK-NEXT:    vpcmpltuq %xmm1, %xmm0, %k1 ## encoding: [0x62,0xf3,0xfd,0x08,0x1e,0xc9,0x01]
; CHECK-NEXT:    vpcmpleuq %xmm1, %xmm0, %k2 ## encoding: [0x62,0xf3,0xfd,0x08,0x1e,0xd1,0x02]
; CHECK-NEXT:    vpcmpunorduq %xmm1, %xmm0, %k3 ## encoding: [0x62,0xf3,0xfd,0x08,0x1e,0xd9,0x03]
; CHECK-NEXT:    vpcmpnequq %xmm1, %xmm0, %k4 ## encoding: [0x62,0xf3,0xfd,0x08,0x1e,0xe1,0x04]
; CHECK-NEXT:    vpcmpnltuq %xmm1, %xmm0, %k5 ## encoding: [0x62,0xf3,0xfd,0x08,0x1e,0xe9,0x05]
; CHECK-NEXT:    vpcmpnleuq %xmm1, %xmm0, %k6 ## encoding: [0x62,0xf3,0xfd,0x08,0x1e,0xf1,0x06]
; CHECK-NEXT:    vpcmporduq %xmm1, %xmm0, %k7 ## encoding: [0x62,0xf3,0xfd,0x08,0x1e,0xf9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 0, i8 -1)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 1, i8 -1)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 2, i8 -1)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 3, i8 -1)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 4, i8 -1)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 5, i8 -1)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 6, i8 -1)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %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_q_128(<2 x i64> %a0, <2 x i64> %a1, i8 %mask) {
; CHECK-LABEL: test_mask_ucmp_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpequq %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1e,0xc1,0x00]
; CHECK-NEXT:    vpcmpltuq %xmm1, %xmm0, %k2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1e,0xd1,0x01]
; CHECK-NEXT:    vpcmpleuq %xmm1, %xmm0, %k3 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1e,0xd9,0x02]
; CHECK-NEXT:    vpcmpunorduq %xmm1, %xmm0, %k4 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1e,0xe1,0x03]
; CHECK-NEXT:    vpcmpnequq %xmm1, %xmm0, %k5 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1e,0xe9,0x04]
; CHECK-NEXT:    vpcmpnltuq %xmm1, %xmm0, %k6 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1e,0xf1,0x05]
; CHECK-NEXT:    vpcmpnleuq %xmm1, %xmm0, %k7 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1e,0xf9,0x06]
; CHECK-NEXT:    vpcmporduq %xmm1, %xmm0, %k1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x1e,0xc9,0x07]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    vpinsrb $0, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x00]
; CHECK-NEXT:    kmovw %k2, %eax ## encoding: [0xc5,0xf8,0x93,0xc2]
; CHECK-NEXT:    vpinsrb $2, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x02]
; CHECK-NEXT:    kmovw %k3, %eax ## encoding: [0xc5,0xf8,0x93,0xc3]
; CHECK-NEXT:    vpinsrb $4, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x04]
; CHECK-NEXT:    kmovw %k4, %eax ## encoding: [0xc5,0xf8,0x93,0xc4]
; CHECK-NEXT:    vpinsrb $6, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x06]
; CHECK-NEXT:    kmovw %k5, %eax ## encoding: [0xc5,0xf8,0x93,0xc5]
; CHECK-NEXT:    vpinsrb $8, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x08]
; CHECK-NEXT:    kmovw %k6, %eax ## encoding: [0xc5,0xf8,0x93,0xc6]
; CHECK-NEXT:    vpinsrb $10, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0a]
; CHECK-NEXT:    kmovw %k7, %eax ## encoding: [0xc5,0xf8,0x93,0xc7]
; CHECK-NEXT:    vpinsrb $12, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0c]
; CHECK-NEXT:    kmovw %k1, %eax ## encoding: [0xc5,0xf8,0x93,0xc1]
; CHECK-NEXT:    vpinsrb $14, %eax, %xmm0, %xmm0 ## encoding: [0xc4,0xe3,0x79,0x20,0xc0,0x0e]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res0 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 0, i8 %mask)
  %vec0 = insertelement <8 x i8> undef, i8 %res0, i32 0
  %res1 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 1, i8 %mask)
  %vec1 = insertelement <8 x i8> %vec0, i8 %res1, i32 1
  %res2 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 2, i8 %mask)
  %vec2 = insertelement <8 x i8> %vec1, i8 %res2, i32 2
  %res3 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 3, i8 %mask)
  %vec3 = insertelement <8 x i8> %vec2, i8 %res3, i32 3
  %res4 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 4, i8 %mask)
  %vec4 = insertelement <8 x i8> %vec3, i8 %res4, i32 4
  %res5 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 5, i8 %mask)
  %vec5 = insertelement <8 x i8> %vec4, i8 %res5, i32 5
  %res6 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 6, i8 %mask)
  %vec6 = insertelement <8 x i8> %vec5, i8 %res6, i32 6
  %res7 = call i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64> %a0, <2 x i64> %a1, i32 7, i8 %mask)
  %vec7 = insertelement <8 x i8> %vec6, i8 %res7, i32 7
  ret <8 x i8> %vec7
}

declare i8 @llvm.x86.avx512.mask.ucmp.q.128(<2 x i64>, <2 x i64>, i32, i8) nounwind readnone

define void @compr1(i8* %addr, <8 x double> %data, i8 %mask) {
; CHECK-LABEL: compr1:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vcompresspd %zmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0x8a,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.compress.store.pd.512(i8* %addr, <8 x double> %data, i8 %mask)
  ret void
}

declare void @llvm.x86.avx512.mask.compress.store.pd.512(i8* %addr, <8 x double> %data, i8 %mask)

define void @compr2(i8* %addr, <4 x double> %data, i8 %mask) {
; CHECK-LABEL: compr2:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vcompresspd %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x8a,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.compress.store.pd.256(i8* %addr, <4 x double> %data, i8 %mask)
  ret void
}

declare void @llvm.x86.avx512.mask.compress.store.pd.256(i8* %addr, <4 x double> %data, i8 %mask)

define void @compr3(i8* %addr, <4 x float> %data, i8 %mask) {
; CHECK-LABEL: compr3:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vcompressps %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x8a,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.compress.store.ps.128(i8* %addr, <4 x float> %data, i8 %mask)
  ret void
}

declare void @llvm.x86.avx512.mask.compress.store.ps.128(i8* %addr, <4 x float> %data, i8 %mask)

define <8 x double> @compr4(i8* %addr, <8 x double> %data, i8 %mask) {
; CHECK-LABEL: compr4:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vcompresspd %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xc9,0x8a,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x double> @llvm.x86.avx512.mask.compress.pd.512(<8 x double> %data, <8 x double> zeroinitializer, i8 %mask)
  ret <8 x double> %res
}

declare <8 x double> @llvm.x86.avx512.mask.compress.pd.512(<8 x double> %data, <8 x double> %src0, i8 %mask)

define <4 x double> @compr5(<4 x double> %data, <4 x double> %src0, i8 %mask) {
; CHECK-LABEL: compr5:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcompresspd %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x8a,0xc1]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.compress.pd.256( <4 x double> %data, <4 x double> %src0, i8 %mask)
  ret <4 x double> %res
}

declare <4 x double> @llvm.x86.avx512.mask.compress.pd.256(<4 x double> %data, <4 x double> %src0, i8 %mask)

define <4 x float> @compr6(<4 x float> %data, i8 %mask) {
; CHECK-LABEL: compr6:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcompressps %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x8a,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.compress.ps.128(<4 x float> %data, <4 x float>zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

declare <4 x float> @llvm.x86.avx512.mask.compress.ps.128(<4 x float> %data, <4 x float> %src0, i8 %mask)

define void @compr7(i8* %addr, <8 x double> %data) {
; CHECK-LABEL: compr7:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovupd %zmm0, (%rdi) ## encoding: [0x62,0xf1,0xfd,0x48,0x11,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.compress.store.pd.512(i8* %addr, <8 x double> %data, i8 -1)
  ret void
}

define <4 x float> @compr8(<4 x float> %data) {
; CHECK-LABEL: compr8:
; CHECK:       ## BB#0:
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.compress.ps.128(<4 x float> %data, <4 x float>zeroinitializer, i8 -1)
  ret <4 x float> %res
}

define void @compr9(i8* %addr, <8 x i64> %data, i8 %mask) {
; CHECK-LABEL: compr9:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpcompressq %zmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0x8b,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.compress.store.q.512(i8* %addr, <8 x i64> %data, i8 %mask)
  ret void
}

declare void @llvm.x86.avx512.mask.compress.store.q.512(i8* %addr, <8 x i64> %data, i8 %mask)

define <4 x i32> @compr10(<4 x i32> %data, i8 %mask) {
; CHECK-LABEL: compr10:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcompressd %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x8b,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.compress.d.128(<4 x i32> %data, <4 x i32>zeroinitializer, i8 %mask)
  ret <4 x i32> %res
}

declare <4 x i32> @llvm.x86.avx512.mask.compress.d.128(<4 x i32> %data, <4 x i32> %src0, i8 %mask)

; Expand

define <8 x double> @expand1(i8* %addr, <8 x double> %data, i8 %mask) {
; CHECK-LABEL: expand1:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vexpandpd (%rdi), %zmm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0x88,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x double> @llvm.x86.avx512.mask.expand.load.pd.512(i8* %addr, <8 x double> %data, i8 %mask)
  ret <8 x double> %res
}

declare <8 x double> @llvm.x86.avx512.mask.expand.load.pd.512(i8* %addr, <8 x double> %data, i8 %mask)

define <4 x double> @expand2(i8* %addr, <4 x double> %data, i8 %mask) {
; CHECK-LABEL: expand2:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vexpandpd (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x88,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.expand.load.pd.256(i8* %addr, <4 x double> %data, i8 %mask)
  ret <4 x double> %res
}

declare <4 x double> @llvm.x86.avx512.mask.expand.load.pd.256(i8* %addr, <4 x double> %data, i8 %mask)

define <4 x float> @expand3(i8* %addr, <4 x float> %data, i8 %mask) {
; CHECK-LABEL: expand3:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vexpandps (%rdi), %xmm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x88,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.expand.load.ps.128(i8* %addr, <4 x float> %data, i8 %mask)
  ret <4 x float> %res
}

declare <4 x float> @llvm.x86.avx512.mask.expand.load.ps.128(i8* %addr, <4 x float> %data, i8 %mask)

define <8 x double> @expand4(i8* %addr, <8 x double> %data, i8 %mask) {
; CHECK-LABEL: expand4:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vexpandpd %zmm0, %zmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xc9,0x88,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x double> @llvm.x86.avx512.mask.expand.pd.512(<8 x double> %data, <8 x double> zeroinitializer, i8 %mask)
  ret <8 x double> %res
}

declare <8 x double> @llvm.x86.avx512.mask.expand.pd.512(<8 x double> %data, <8 x double> %src0, i8 %mask)

define <4 x double> @expand5(<4 x double> %data, <4 x double> %src0, i8 %mask) {
; CHECK-LABEL: expand5:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vexpandpd %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x88,0xc8]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.expand.pd.256( <4 x double> %data, <4 x double> %src0, i8 %mask)
  ret <4 x double> %res
}

declare <4 x double> @llvm.x86.avx512.mask.expand.pd.256(<4 x double> %data, <4 x double> %src0, i8 %mask)

define <4 x float> @expand6(<4 x float> %data, i8 %mask) {
; CHECK-LABEL: expand6:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vexpandps %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x88,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.expand.ps.128(<4 x float> %data, <4 x float>zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

declare <4 x float> @llvm.x86.avx512.mask.expand.ps.128(<4 x float> %data, <4 x float> %src0, i8 %mask)

define <8 x double> @expand7(i8* %addr, <8 x double> %data) {
; CHECK-LABEL: expand7:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovupd (%rdi), %zmm0 ## encoding: [0x62,0xf1,0xfd,0x48,0x10,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x double> @llvm.x86.avx512.mask.expand.load.pd.512(i8* %addr, <8 x double> %data, i8 -1)
  ret <8 x double> %res
}

define <4 x float> @expand8(<4 x float> %data) {
; CHECK-LABEL: expand8:
; CHECK:       ## BB#0:
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.expand.ps.128(<4 x float> %data, <4 x float>zeroinitializer, i8 -1)
  ret <4 x float> %res
}

define <8 x i64> @expand9(i8* %addr, <8 x i64> %data, i8 %mask) {
; CHECK-LABEL: expand9:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpexpandq (%rdi), %zmm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x49,0x89,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i64> @llvm.x86.avx512.mask.expand.load.q.512(i8* %addr, <8 x i64> %data, i8 %mask)
  ret <8 x i64> %res
}

declare <8 x i64> @llvm.x86.avx512.mask.expand.load.q.512(i8* %addr, <8 x i64> %data, i8 %mask)

define <4 x i32> @expand10(<4 x i32> %data, i8 %mask) {
; CHECK-LABEL: expand10:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpexpandd %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x89,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.expand.d.128(<4 x i32> %data, <4 x i32>zeroinitializer, i8 %mask)
  ret <4 x i32> %res
}

declare <4 x i32> @llvm.x86.avx512.mask.expand.d.128(<4 x i32> %data, <4 x i32> %src0, i8 %mask)

define < 2 x i64> @test_mask_mul_epi32_rr_128(< 4 x i32> %a, < 4 x i32> %b) {
; CHECK-LABEL: test_mask_mul_epi32_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuldq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 -1)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epi32_rrk_128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmuldq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x28,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> %passThru, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epi32_rrkz_128(< 4 x i32> %a, < 4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmuldq %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epi32_rm_128(< 4 x i32> %a, < 4 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_mul_epi32_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuldq (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x28,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 4 x i32>, < 4 x i32>* %ptr_b
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 -1)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epi32_rmk_128(< 4 x i32> %a, < 4 x i32>* %ptr_b, < 2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuldq (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x28,0x0f]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 4 x i32>, < 4 x i32>* %ptr_b
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> %passThru, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epi32_rmkz_128(< 4 x i32> %a, < 4 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuldq (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x28,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 4 x i32>, < 4 x i32>* %ptr_b
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epi32_rmb_128(< 4 x i32> %a, i64* %ptr_b) {
; CHECK-LABEL: test_mask_mul_epi32_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuldq (%rdi){1to2}, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x18,0x28,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 2 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 2 x i64> %vecinit.i, < 2 x i64> undef, <2 x i32> zeroinitializer
  %b = bitcast < 2 x i64> %b64 to < 4 x i32>
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 -1)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epi32_rmbk_128(< 4 x i32> %a, i64* %ptr_b, < 2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuldq (%rdi){1to2}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x19,0x28,0x0f]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 2 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 2 x i64> %vecinit.i, < 2 x i64> undef, <2 x i32> zeroinitializer
  %b = bitcast < 2 x i64> %b64 to < 4 x i32>
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> %passThru, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epi32_rmbkz_128(< 4 x i32> %a, i64* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuldq (%rdi){1to2}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x99,0x28,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 2 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 2 x i64> %vecinit.i, < 2 x i64> undef, < 2 x i32> zeroinitializer
  %b = bitcast < 2 x i64> %b64 to < 4 x i32>
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 %mask)
  ret < 2 x i64> %res
}

declare < 2 x i64> @llvm.x86.avx512.mask.pmul.dq.128(< 4 x i32>, < 4 x i32>, < 2 x i64>, i8)

define < 4 x i64> @test_mask_mul_epi32_rr_256(< 8 x i32> %a, < 8 x i32> %b) {
; CHECK-LABEL: test_mask_mul_epi32_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuldq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 -1)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epi32_rrk_256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmuldq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x28,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> %passThru, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epi32_rrkz_256(< 8 x i32> %a, < 8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmuldq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epi32_rm_256(< 8 x i32> %a, < 8 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_mul_epi32_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuldq (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x28,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 8 x i32>, < 8 x i32>* %ptr_b
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 -1)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epi32_rmk_256(< 8 x i32> %a, < 8 x i32>* %ptr_b, < 4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuldq (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x28,0x0f]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 8 x i32>, < 8 x i32>* %ptr_b
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> %passThru, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epi32_rmkz_256(< 8 x i32> %a, < 8 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuldq (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x28,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 8 x i32>, < 8 x i32>* %ptr_b
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epi32_rmb_256(< 8 x i32> %a, i64* %ptr_b) {
; CHECK-LABEL: test_mask_mul_epi32_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuldq (%rdi){1to4}, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x38,0x28,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 4 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 4 x i64> %vecinit.i, < 4 x i64> undef, < 4 x i32> zeroinitializer
  %b = bitcast < 4 x i64> %b64 to < 8 x i32>
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 -1)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epi32_rmbk_256(< 8 x i32> %a, i64* %ptr_b, < 4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuldq (%rdi){1to4}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x39,0x28,0x0f]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 4 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 4 x i64> %vecinit.i, < 4 x i64> undef, < 4 x i32> zeroinitializer
  %b = bitcast < 4 x i64> %b64 to < 8 x i32>
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> %passThru, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epi32_rmbkz_256(< 8 x i32> %a, i64* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epi32_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuldq (%rdi){1to4}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xb9,0x28,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 4 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 4 x i64> %vecinit.i, < 4 x i64> undef, < 4 x i32> zeroinitializer
  %b = bitcast < 4 x i64> %b64 to < 8 x i32>
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 %mask)
  ret < 4 x i64> %res
}

declare < 4 x i64> @llvm.x86.avx512.mask.pmul.dq.256(< 8 x i32>, < 8 x i32>, < 4 x i64>, i8)

define < 2 x i64> @test_mask_mul_epu32_rr_128(< 4 x i32> %a, < 4 x i32> %b) {
; CHECK-LABEL: test_mask_mul_epu32_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuludq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xf4,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 -1)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epu32_rrk_128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmuludq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0xf4,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> %passThru, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epu32_rrkz_128(< 4 x i32> %a, < 4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmuludq %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0xf4,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epu32_rm_128(< 4 x i32> %a, < 4 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_mul_epu32_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuludq (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xf4,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 4 x i32>, < 4 x i32>* %ptr_b
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 -1)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epu32_rmk_128(< 4 x i32> %a, < 4 x i32>* %ptr_b, < 2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuludq (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0xf4,0x0f]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 4 x i32>, < 4 x i32>* %ptr_b
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> %passThru, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epu32_rmkz_128(< 4 x i32> %a, < 4 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuludq (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0xf4,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 4 x i32>, < 4 x i32>* %ptr_b
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epu32_rmb_128(< 4 x i32> %a, i64* %ptr_b) {
; CHECK-LABEL: test_mask_mul_epu32_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuludq (%rdi){1to2}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x18,0xf4,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 2 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 2 x i64> %vecinit.i, < 2 x i64> undef, <2 x i32> zeroinitializer
  %b = bitcast < 2 x i64> %b64 to < 4 x i32>
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 -1)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epu32_rmbk_128(< 4 x i32> %a, i64* %ptr_b, < 2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuludq (%rdi){1to2}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x19,0xf4,0x0f]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 2 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 2 x i64> %vecinit.i, < 2 x i64> undef, <2 x i32> zeroinitializer
  %b = bitcast < 2 x i64> %b64 to < 4 x i32>
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> %passThru, i8 %mask)
  ret < 2 x i64> %res
}

define < 2 x i64> @test_mask_mul_epu32_rmbkz_128(< 4 x i32> %a, i64* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuludq (%rdi){1to2}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x99,0xf4,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 2 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 2 x i64> %vecinit.i, < 2 x i64> undef, < 2 x i32> zeroinitializer
  %b = bitcast < 2 x i64> %b64 to < 4 x i32>
  %res = call < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32> %a, < 4 x i32> %b, < 2 x i64> zeroinitializer, i8 %mask)
  ret < 2 x i64> %res
}

declare < 2 x i64> @llvm.x86.avx512.mask.pmulu.dq.128(< 4 x i32>, < 4 x i32>, < 2 x i64>, i8)

define < 4 x i64> @test_mask_mul_epu32_rr_256(< 8 x i32> %a, < 8 x i32> %b) {
; CHECK-LABEL: test_mask_mul_epu32_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xf4,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 -1)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epu32_rrk_256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmuludq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xf4,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> %passThru, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epu32_rrkz_256(< 8 x i32> %a, < 8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0xf4,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epu32_rm_256(< 8 x i32> %a, < 8 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_mul_epu32_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuludq (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xf4,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 8 x i32>, < 8 x i32>* %ptr_b
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 -1)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epu32_rmk_256(< 8 x i32> %a, < 8 x i32>* %ptr_b, < 4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuludq (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xf4,0x0f]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 8 x i32>, < 8 x i32>* %ptr_b
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> %passThru, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epu32_rmkz_256(< 8 x i32> %a, < 8 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuludq (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0xf4,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load < 8 x i32>, < 8 x i32>* %ptr_b
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epu32_rmb_256(< 8 x i32> %a, i64* %ptr_b) {
; CHECK-LABEL: test_mask_mul_epu32_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmuludq (%rdi){1to4}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x38,0xf4,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 4 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 4 x i64> %vecinit.i, < 4 x i64> undef, < 4 x i32> zeroinitializer
  %b = bitcast < 4 x i64> %b64 to < 8 x i32>
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 -1)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epu32_rmbk_256(< 8 x i32> %a, i64* %ptr_b, < 4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuludq (%rdi){1to4}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x39,0xf4,0x0f]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 4 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 4 x i64> %vecinit.i, < 4 x i64> undef, < 4 x i32> zeroinitializer
  %b = bitcast < 4 x i64> %b64 to < 8 x i32>
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> %passThru, i8 %mask)
  ret < 4 x i64> %res
}

define < 4 x i64> @test_mask_mul_epu32_rmbkz_256(< 8 x i32> %a, i64* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_mul_epu32_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmuludq (%rdi){1to4}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xb9,0xf4,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement < 4 x i64> undef, i64 %q, i32 0
  %b64 = shufflevector < 4 x i64> %vecinit.i, < 4 x i64> undef, < 4 x i32> zeroinitializer
  %b = bitcast < 4 x i64> %b64 to < 8 x i32>
  %res = call < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32> %a, < 8 x i32> %b, < 4 x i64> zeroinitializer, i8 %mask)
  ret < 4 x i64> %res
}

declare < 4 x i64> @llvm.x86.avx512.mask.pmulu.dq.256(< 8 x i32>, < 8 x i32>, < 4 x i64>, i8)

define <4 x i32> @test_mask_add_epi32_rr_128(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_mask_add_epi32_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 -1)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_add_epi32_rrk_128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xfe,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_add_epi32_rrkz_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_add_epi32_rm_128(<4 x i32> %a, <4 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_add_epi32_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpaddd (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 -1)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_add_epi32_rmk_128(<4 x i32> %a, <4 x i32>* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpaddd (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xfe,0x0f]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_add_epi32_rmkz_128(<4 x i32> %a, <4 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpaddd (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xfe,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_add_epi32_rmb_128(<4 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_add_epi32_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpaddd (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x18,0xfe,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
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 -1)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_add_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpaddd (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x19,0xfe,0x0f]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; 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
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_add_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpaddd (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0xfe,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
  %res = call <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 %mask)
  ret <4 x i32> %res
}

declare <4 x i32> @llvm.x86.avx512.mask.padd.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32> @test_mask_sub_epi32_rr_128(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_mask_sub_epi32_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpsubd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfa,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 -1)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_sub_epi32_rrk_128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsubd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xfa,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_sub_epi32_rrkz_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsubd %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xfa,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_sub_epi32_rm_128(<4 x i32> %a, <4 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_sub_epi32_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpsubd (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfa,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 -1)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_sub_epi32_rmk_128(<4 x i32> %a, <4 x i32>* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsubd (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xfa,0x0f]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_sub_epi32_rmkz_128(<4 x i32> %a, <4 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsubd (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xfa,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_sub_epi32_rmb_128(<4 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_sub_epi32_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpsubd (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x18,0xfa,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
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 -1)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_sub_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsubd (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x19,0xfa,0x0f]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; 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
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask)
  ret <4 x i32> %res
}

define <4 x i32> @test_mask_sub_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsubd (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0xfa,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
  %res = call <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32> %a, <4 x i32> %b, <4 x i32> zeroinitializer, i8 %mask)
  ret <4 x i32> %res
}

declare <4 x i32> @llvm.x86.avx512.mask.psub.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <8 x i32> @test_mask_sub_epi32_rr_256(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: test_mask_sub_epi32_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpsubd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfa,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 -1)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_sub_epi32_rrk_256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsubd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xfa,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_sub_epi32_rrkz_256(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsubd %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xfa,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_sub_epi32_rm_256(<8 x i32> %a, <8 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_sub_epi32_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpsubd (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfa,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 -1)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_sub_epi32_rmk_256(<8 x i32> %a, <8 x i32>* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsubd (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xfa,0x0f]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_sub_epi32_rmkz_256(<8 x i32> %a, <8 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsubd (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xfa,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_sub_epi32_rmb_256(<8 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_sub_epi32_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpsubd (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x38,0xfa,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
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 -1)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_sub_epi32_rmbk_256(<8 x i32> %a, i32* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsubd (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x39,0xfa,0x0f]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; 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
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_sub_epi32_rmbkz_256(<8 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_sub_epi32_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsubd (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xb9,0xfa,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
  %res = call <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 %mask)
  ret <8 x i32> %res
}

declare <8 x i32> @llvm.x86.avx512.mask.psub.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32> @test_mask_add_epi32_rr_256(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: test_mask_add_epi32_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpaddd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 -1)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_add_epi32_rrk_256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpaddd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xfe,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_add_epi32_rrkz_256(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpaddd %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_add_epi32_rm_256(<8 x i32> %a, <8 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_add_epi32_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpaddd (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfe,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 -1)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_add_epi32_rmk_256(<8 x i32> %a, <8 x i32>* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpaddd (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xfe,0x0f]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_add_epi32_rmkz_256(<8 x i32> %a, <8 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpaddd (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xfe,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_add_epi32_rmb_256(<8 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_add_epi32_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpaddd (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x38,0xfe,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
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 -1)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_add_epi32_rmbk_256(<8 x i32> %a, i32* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpaddd (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x39,0xfe,0x0f]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; 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
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask)
  ret <8 x i32> %res
}

define <8 x i32> @test_mask_add_epi32_rmbkz_256(<8 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_add_epi32_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpaddd (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xb9,0xfe,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
  %res = call <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32> %a, <8 x i32> %b, <8 x i32> zeroinitializer, i8 %mask)
  ret <8 x i32> %res
}

declare <8 x i32> @llvm.x86.avx512.mask.padd.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define i8 @test_cmpps_256(<8 x float> %a, <8 x float> %b) {
; CHECK-LABEL: test_cmpps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vcmpleps %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf1,0x7c,0x28,0xc2,0xc1,0x02]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
   %res = call i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> %a, <8 x float> %b, i32 2, i8 -1)
   ret i8 %res
 }
 declare i8 @llvm.x86.avx512.mask.cmp.ps.256(<8 x float> , <8 x float> , i32, i8)

define i8 @test_cmpps_128(<4 x float> %a, <4 x float> %b) {
; CHECK-LABEL: test_cmpps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vcmpleps %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf1,0x7c,0x08,0xc2,0xc1,0x02]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
   %res = call i8 @llvm.x86.avx512.mask.cmp.ps.128(<4 x float> %a, <4 x float> %b, i32 2, i8 -1)
   ret i8 %res
 }
 declare i8 @llvm.x86.avx512.mask.cmp.ps.128(<4 x float> , <4 x float> , i32, i8)

define i8 @test_cmppd_256(<4 x double> %a, <4 x double> %b) {
; CHECK-LABEL: test_cmppd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vcmplepd %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf1,0xfd,0x28,0xc2,0xc1,0x02]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
   %res = call i8 @llvm.x86.avx512.mask.cmp.pd.256(<4 x double> %a, <4 x double> %b, i32 2, i8 -1)
   ret i8 %res
 }
 declare i8 @llvm.x86.avx512.mask.cmp.pd.256(<4 x double> , <4 x double> , i32, i8)

define i8 @test_cmppd_128(<2 x double> %a, <2 x double> %b) {
; CHECK-LABEL: test_cmppd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vcmplepd %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf1,0xfd,0x08,0xc2,0xc1,0x02]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
   %res = call i8 @llvm.x86.avx512.mask.cmp.pd.128(<2 x double> %a, <2 x double> %b, i32 2, i8 -1)
   ret i8 %res
 }
 declare i8 @llvm.x86.avx512.mask.cmp.pd.128(<2 x double> , <2 x double> , i32, i8)

define <8 x float> @test_mm512_maskz_add_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_add_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vaddps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.add.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_mask_add_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_add_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vaddps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x58,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.add.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_add_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_add_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vaddps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.add.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 -1)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.avx512.mask.add.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <4 x float> @test_mm512_maskz_add_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_add_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vaddps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.add.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_mask_add_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_add_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vaddps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x58,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.add.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_add_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_add_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vaddps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.add.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 -1)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.avx512.mask.add.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <8 x float> @test_mm512_maskz_sub_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_sub_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vsubps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x5c,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.sub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_mask_sub_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_sub_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vsubps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x5c,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.sub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_sub_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_sub_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vsubps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x5c,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.sub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 -1)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.avx512.mask.sub.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <4 x float> @test_mm512_maskz_sub_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_sub_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vsubps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x5c,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.sub.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_mask_sub_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_sub_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vsubps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x5c,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.sub.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_sub_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_sub_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vsubps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x5c,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.sub.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 -1)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.avx512.mask.sub.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <8 x float> @test_mm512_maskz_mul_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_mul_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmulps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x59,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.mul.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_mask_mul_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_mul_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmulps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x59,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.mul.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_mul_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_mul_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmulps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x59,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.mul.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 -1)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.avx512.mask.mul.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <4 x float> @test_mm512_maskz_mul_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_mul_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmulps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x59,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.mul.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_mask_mul_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_mul_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmulps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x59,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.mul.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_mul_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_mul_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmulps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x59,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.mul.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 -1)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.avx512.mask.mul.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <8 x float> @test_mm512_maskz_div_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_div_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vdivps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x5e,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.div.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_mask_div_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_div_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vdivps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x5e,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.div.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_div_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_div_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vdivps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x5e,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.div.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 -1)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.avx512.mask.div.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <4 x float> @test_mm512_maskz_div_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_div_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vdivps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x5e,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.div.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_mask_div_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_div_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vdivps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x5e,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.div.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_div_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_div_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vdivps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x5e,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.div.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 -1)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.avx512.mask.div.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <8 x float> @test_mm512_maskz_max_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_max_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmaxps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x5f,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.max.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_mask_max_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_max_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmaxps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x5f,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.max.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_max_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_max_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmaxps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x5f,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.max.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 -1)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.avx512.mask.max.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <4 x float> @test_mm512_maskz_max_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_max_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmaxps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x5f,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.max.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_mask_max_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_max_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmaxps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x5f,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.max.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_max_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_max_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmaxps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x5f,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.max.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 -1)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.avx512.mask.max.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <8 x float> @test_mm512_maskz_min_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_min_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vminps %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x5d,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.min.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_mask_min_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_min_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vminps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x5d,0xd1]
; CHECK-NEXT:    vmovaps %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.min.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %src, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_mm512_min_ps_256(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_min_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vminps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x5d,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.min.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float>zeroinitializer, i8 -1)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.avx512.mask.min.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <4 x float> @test_mm512_maskz_min_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_maskz_min_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vminps %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x5d,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.min.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_mask_min_ps_128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask) {
; CHECK-LABEL: test_mm512_mask_min_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vminps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x5d,0xd1]
; CHECK-NEXT:    vmovaps %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.min.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float> %src, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_mm512_min_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_mm512_min_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vminps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x5d,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.min.ps.128(<4 x float> %a0, <4 x float> %a1, <4 x float>zeroinitializer, i8 -1)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.avx512.mask.min.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <4 x double> @test_sqrt_pd_256(<4 x double> %a0, i8 %mask) {
; CHECK-LABEL: test_sqrt_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vsqrtpd %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0x51,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.sqrt.pd.256(<4 x double> %a0,  <4 x double> zeroinitializer, i8 %mask)
  ret <4 x double> %res
}
declare <4 x double> @llvm.x86.avx512.mask.sqrt.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone

define <8 x float> @test_sqrt_ps_256(<8 x float> %a0, i8 %mask) {
; CHECK-LABEL: test_sqrt_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vsqrtps %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x51,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.sqrt.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

declare <8 x float> @llvm.x86.avx512.mask.sqrt.ps.256(<8 x float>, <8 x float>, i8) nounwind readnone

define <4 x double> @test_getexp_pd_256(<4 x double> %a0) {
; CHECK-LABEL: test_getexp_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vgetexppd %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x42,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.getexp.pd.256(<4 x double> %a0,  <4 x double> zeroinitializer, i8 -1)
  ret <4 x double> %res
}

declare <4 x double> @llvm.x86.avx512.mask.getexp.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone

define <8 x float> @test_getexp_ps_256(<8 x float> %a0) {
; CHECK-LABEL: test_getexp_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vgetexpps %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x42,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.getexp.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 -1)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.avx512.mask.getexp.ps.256(<8 x float>, <8 x float>, i8) nounwind readnone

declare <4 x i32> @llvm.x86.avx512.mask.pmaxs.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmaxs_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaxs_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmaxsd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x3d,0xd1]
; CHECK-NEXT:    vpmaxsd %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x3d,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.pmaxs.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2 ,i8 %mask)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmaxs.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %mask)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.pmaxs.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pmaxs_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaxs_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmaxsd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x3d,0xd1]
; CHECK-NEXT:    vpmaxsd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x3d,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.pmaxs.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmaxs.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.pmaxs.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pmaxs_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaxs_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmaxsq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x3d,0xd1]
; CHECK-NEXT:    vpmaxsq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x3d,0xc1]
; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.pmaxs.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmaxs.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.pmaxs.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pmaxs_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaxs_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmaxsq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x3d,0xd1]
; CHECK-NEXT:    vpmaxsq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x3d,0xc1]
; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.pmaxs.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmaxs.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %mask)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.pmaxu.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmaxu_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2,i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaxu_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmaxud %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x3f,0xd1]
; CHECK-NEXT:    vpmaxud %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x3f,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.pmaxu.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmaxu.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %mask)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.pmaxu.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pmaxu_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaxu_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmaxud %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x3f,0xd1]
; CHECK-NEXT:    vpmaxud %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x3f,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.pmaxu.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmaxu.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.pmaxu.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pmaxu_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaxu_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmaxuq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x3f,0xd1]
; CHECK-NEXT:    vpmaxuq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x3f,0xc1]
; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.pmaxu.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmaxu.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.pmaxu.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pmaxu_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmaxu_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmaxuq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x3f,0xd1]
; CHECK-NEXT:    vpmaxuq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x3f,0xc1]
; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.pmaxu.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmaxu.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %mask)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.pmins.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmins_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmins_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpminsd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x39,0xd1]
; CHECK-NEXT:    vpminsd %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x39,0xc1]
; 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.pmins.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmins.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %mask)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.pmins.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pmins_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmins_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpminsd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x39,0xd1]
; CHECK-NEXT:    vpminsd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x39,0xc1]
; 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.pmins.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmins.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.pmins.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pmins_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmins_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpminsq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x39,0xd1]
; CHECK-NEXT:    vpminsq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x39,0xc1]
; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.pmins.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmins.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.pmins.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pmins_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmins_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpminsq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x39,0xd1]
; CHECK-NEXT:    vpminsq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x39,0xc1]
; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.pmins.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmins.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %mask)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.pminu.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pminu_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pminu_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpminud %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x3b,0xd1]
; CHECK-NEXT:    vpminud %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x3b,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.pminu.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %mask)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pminu.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %mask)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.pminu.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pminu_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pminu_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpminud %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x3b,0xd1]
; CHECK-NEXT:    vpminud %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x3b,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.pminu.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pminu.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.pminu.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pminu_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pminu_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpminuq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x3b,0xd1]
; CHECK-NEXT:    vpminuq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x3b,0xc1]
; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.pminu.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pminu.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.pminu.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pminu_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pminu_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpminuq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x3b,0xd1]
; CHECK-NEXT:    vpminuq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x3b,0xc1]
; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.pminu.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %mask)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pminu.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %mask)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.vpermt2var.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_vpermt2var_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermt2var_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm1, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd9]
; CHECK-NEXT:    vpermt2d %xmm2, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x7e,0xda]
; CHECK-NEXT:    vpermt2d %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0x7e,0xca]
; CHECK-NEXT:    vpaddd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x65,0x08,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.vpermt2var.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.vpermt2var.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_maskz_vpermt2var_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vpermt2var_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm1, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd9]
; CHECK-NEXT:    vpermt2d %xmm2, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x7e,0xda]
; CHECK-NEXT:    vpermt2d %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0x7e,0xca]
; CHECK-NEXT:    vpaddd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x65,0x08,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.vpermt2var.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_vpermt2var_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermt2var_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm1, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd9]
; CHECK-NEXT:    vpermt2d %ymm2, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x7e,0xda]
; CHECK-NEXT:    vpermt2d %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0x7e,0xca]
; CHECK-NEXT:    vpaddd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.vpermt2var.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.vpermt2var.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_maskz_vpermt2var_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_maskz_vpermt2var_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm1, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd9]
; CHECK-NEXT:    vpermt2d %ymm2, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x7e,0xda]
; CHECK-NEXT:    vpermt2d %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0x7e,0xca]
; CHECK-NEXT:    vpaddd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.maskz.vpermt2var.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x double> @llvm.x86.avx512.mask.vpermi2var.pd.128(<2 x double>, <2 x i64>, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_vpermi2var_pd_128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermi2var_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm1, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd9]
; CHECK-NEXT:    vpermi2pd %xmm2, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x77,0xda]
; CHECK-NEXT:    vpermi2pd %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0xfd,0x08,0x77,0xca]
; CHECK-NEXT:    vaddpd %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.mask.vpermi2var.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 %x3)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.vpermi2var.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 -1)
  %res2 = fadd <2 x double> %res, %res1
  ret <2 x double> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.vpermi2var.pd.256(<4 x double>, <4 x i64>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_vpermi2var_pd_256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermi2var_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm1, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd9]
; CHECK-NEXT:    vpermi2pd %ymm2, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x77,0xda]
; CHECK-NEXT:    vpermi2pd %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0xfd,0x28,0x77,0xca]
; CHECK-NEXT:    vaddpd %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.vpermi2var.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.vpermi2var.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.vpermi2var.ps.128(<4 x float>, <4 x i32>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_vpermi2var_ps_128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermi2var_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm1, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd9]
; CHECK-NEXT:    vpermi2ps %xmm2, %xmm0, %xmm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x77,0xda]
; CHECK-NEXT:    vpermi2ps %xmm2, %xmm0, %xmm1 ## encoding: [0x62,0xf2,0x7d,0x08,0x77,0xca]
; CHECK-NEXT:    vaddps %xmm1, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.vpermi2var.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 %x3)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.vpermi2var.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.vpermi2var.ps.256(<8 x float>, <8 x i32>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_vpermi2var_ps_256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermi2var_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm1, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd9]
; CHECK-NEXT:    vpermi2ps %ymm2, %ymm0, %ymm3 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x77,0xda]
; CHECK-NEXT:    vpermi2ps %ymm2, %ymm0, %ymm1 ## encoding: [0x62,0xf2,0x7d,0x28,0x77,0xca]
; CHECK-NEXT:    vaddps %ymm1, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.vpermi2var.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.vpermi2var.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.pabs.q.128(<2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pabs_q_128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pabs_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpabsq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x1f,0xc8]
; CHECK-NEXT:    vpabsq %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x1f,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.pabs.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pabs.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.pabs.q.256(<4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pabs_q_256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pabs_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpabsq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x1f,0xc8]
; CHECK-NEXT:    vpabsq %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x1f,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.pabs.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pabs.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.pabs.d.128(<4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pabs_d_128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pabs_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpabsd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x1e,0xc8]
; CHECK-NEXT:    vpabsd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x1e,0xc0]
; 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.pabs.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pabs.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.pabs.d.256(<8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pabs_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pabs_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpabsd %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x1e,0xc8]
; CHECK-NEXT:    vpabsd %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x1e,0xc0]
; 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.pabs.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pabs.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x double> @llvm.x86.avx512.mask.scalef.pd.128(<2 x double>, <2 x double>, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_scalef_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_scalef_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vscalefpd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x2c,0xd1]
; CHECK-NEXT:    vscalefpd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x2c,0xc1]
; 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.scalef.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.scalef.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 -1)
  %res2 = fadd <2 x double> %res, %res1
  ret <2 x double> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.scalef.pd.256(<4 x double>, <4 x double>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_scalef_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_scalef_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vscalefpd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x2c,0xd1]
; CHECK-NEXT:    vscalefpd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x2c,0xc1]
; 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.scalef.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.scalef.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.scalef.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_scalef_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_scalef_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vscalefps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x2c,0xd1]
; CHECK-NEXT:    vscalefps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x2c,0xc1]
; 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.scalef.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.scalef.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.scalef.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_scalef_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_scalef_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vscalefps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x2c,0xd1]
; CHECK-NEXT:    vscalefps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x2c,0xc1]
; 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.scalef.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.scalef.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <16 x i8> @llvm.x86.avx512.mask.pmov.qb.128(<2 x i64>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmov_qb_128(<2 x i64> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovqb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x32,0xc1]
; CHECK-NEXT:    vpmovqb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x32,0xc2]
; CHECK-NEXT:    vpmovqb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x32,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmov.qb.128(<2 x i64> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmov.qb.128(<2 x i64> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmov.qb.128(<2 x i64> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmov.qb.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmov_qb_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qb_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovqb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x32,0x07]
; CHECK-NEXT:    vpmovqb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x32,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.qb.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.qb.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmovs.qb.128(<2 x i64>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmovs_qb_128(<2 x i64> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsqb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x22,0xc1]
; CHECK-NEXT:    vpmovsqb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x22,0xc2]
; CHECK-NEXT:    vpmovsqb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x22,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.qb.128(<2 x i64> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.qb.128(<2 x i64> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.qb.128(<2 x i64> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.qb.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovs_qb_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qb_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsqb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x22,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsqb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x22,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.qb.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.qb.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmovus.qb.128(<2 x i64>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmovus_qb_128(<2 x i64> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusqb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x12,0xc1]
; CHECK-NEXT:    vpmovusqb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x12,0xc2]
; CHECK-NEXT:    vpmovusqb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x12,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.qb.128(<2 x i64> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.qb.128(<2 x i64> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.qb.128(<2 x i64> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.qb.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovus_qb_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qb_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusqb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x12,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusqb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x12,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.qb.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.qb.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmov.qb.256(<4 x i64>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmov_qb_256(<4 x i64> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovqb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x32,0xc1]
; CHECK-NEXT:    vpmovqb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x32,0xc2]
; CHECK-NEXT:    vpmovqb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x32,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmov.qb.256(<4 x i64> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmov.qb.256(<4 x i64> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmov.qb.256(<4 x i64> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmov.qb.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmov_qb_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qb_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovqb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x32,0x07]
; CHECK-NEXT:    vpmovqb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x32,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.qb.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.qb.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmovs.qb.256(<4 x i64>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmovs_qb_256(<4 x i64> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsqb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x22,0xc1]
; CHECK-NEXT:    vpmovsqb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x22,0xc2]
; CHECK-NEXT:    vpmovsqb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x22,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.qb.256(<4 x i64> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.qb.256(<4 x i64> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.qb.256(<4 x i64> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.qb.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovs_qb_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qb_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsqb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x22,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsqb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x22,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.qb.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.qb.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmovus.qb.256(<4 x i64>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmovus_qb_256(<4 x i64> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusqb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x12,0xc1]
; CHECK-NEXT:    vpmovusqb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x12,0xc2]
; CHECK-NEXT:    vpmovusqb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x12,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.qb.256(<4 x i64> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.qb.256(<4 x i64> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.qb.256(<4 x i64> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.qb.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovus_qb_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qb_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusqb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x12,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusqb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x12,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.qb.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.qb.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmov.qw.128(<2 x i64>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmov_qw_128(<2 x i64> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qw_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovqw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x34,0xc1]
; CHECK-NEXT:    vpmovqw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x34,0xc2]
; CHECK-NEXT:    vpmovqw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x34,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmov.qw.128(<2 x i64> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmov.qw.128(<2 x i64> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmov.qw.128(<2 x i64> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmov.qw.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmov_qw_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qw_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovqw %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x34,0x07]
; CHECK-NEXT:    vpmovqw %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x34,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.qw.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.qw.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmovs.qw.128(<2 x i64>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmovs_qw_128(<2 x i64> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qw_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsqw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x24,0xc1]
; CHECK-NEXT:    vpmovsqw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x24,0xc2]
; CHECK-NEXT:    vpmovsqw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x24,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.qw.128(<2 x i64> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.qw.128(<2 x i64> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.qw.128(<2 x i64> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.qw.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovs_qw_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qw_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsqw %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x24,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsqw %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x24,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.qw.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.qw.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmovus.qw.128(<2 x i64>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmovus_qw_128(<2 x i64> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qw_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusqw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x14,0xc1]
; CHECK-NEXT:    vpmovusqw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x14,0xc2]
; CHECK-NEXT:    vpmovusqw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x14,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.qw.128(<2 x i64> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.qw.128(<2 x i64> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.qw.128(<2 x i64> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.qw.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovus_qw_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qw_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusqw %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x14,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusqw %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x14,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.qw.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.qw.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmov.qw.256(<4 x i64>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmov_qw_256(<4 x i64> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qw_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovqw %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x34,0xc1]
; CHECK-NEXT:    vpmovqw %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x34,0xc2]
; CHECK-NEXT:    vpmovqw %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x34,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmov.qw.256(<4 x i64> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmov.qw.256(<4 x i64> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmov.qw.256(<4 x i64> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmov.qw.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmov_qw_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qw_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovqw %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x34,0x07]
; CHECK-NEXT:    vpmovqw %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x34,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.qw.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.qw.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmovs.qw.256(<4 x i64>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmovs_qw_256(<4 x i64> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qw_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsqw %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x24,0xc1]
; CHECK-NEXT:    vpmovsqw %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x24,0xc2]
; CHECK-NEXT:    vpmovsqw %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x24,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.qw.256(<4 x i64> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.qw.256(<4 x i64> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.qw.256(<4 x i64> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.qw.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovs_qw_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qw_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsqw %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x24,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsqw %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x24,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.qw.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.qw.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmovus.qw.256(<4 x i64>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmovus_qw_256(<4 x i64> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qw_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusqw %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x14,0xc1]
; CHECK-NEXT:    vpmovusqw %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x14,0xc2]
; CHECK-NEXT:    vpmovusqw %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x14,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.qw.256(<4 x i64> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.qw.256(<4 x i64> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.qw.256(<4 x i64> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.qw.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovus_qw_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qw_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusqw %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x14,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusqw %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x14,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.qw.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.qw.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <4 x i32> @llvm.x86.avx512.mask.pmov.qd.128(<2 x i64>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmov_qd_128(<2 x i64> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovqd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x35,0xc1]
; CHECK-NEXT:    vpmovqd %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x35,0xc2]
; CHECK-NEXT:    vpmovqd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x35,0xc0]
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc1]
; CHECK-NEXT:    vpaddd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <4 x i32> @llvm.x86.avx512.mask.pmov.qd.128(<2 x i64> %x0, <4 x i32> %x1, i8 -1)
    %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmov.qd.128(<2 x i64> %x0, <4 x i32> %x1, i8 %x2)
    %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmov.qd.128(<2 x i64> %x0, <4 x i32> zeroinitializer, i8 %x2)
    %res3 = add <4 x i32> %res0, %res1
    %res4 = add <4 x i32> %res3, %res2
    ret <4 x i32> %res4
}

declare void @llvm.x86.avx512.mask.pmov.qd.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmov_qd_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qd_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovqd %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x35,0x07]
; CHECK-NEXT:    vpmovqd %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x35,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.qd.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.qd.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.128(<2 x i64>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmovs_qd_128(<2 x i64> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsqd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x25,0xc1]
; CHECK-NEXT:    vpmovsqd %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x25,0xc2]
; CHECK-NEXT:    vpmovsqd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x25,0xc0]
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc1]
; CHECK-NEXT:    vpaddd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.128(<2 x i64> %x0, <4 x i32> %x1, i8 -1)
    %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.128(<2 x i64> %x0, <4 x i32> %x1, i8 %x2)
    %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.128(<2 x i64> %x0, <4 x i32> zeroinitializer, i8 %x2)
    %res3 = add <4 x i32> %res0, %res1
    %res4 = add <4 x i32> %res3, %res2
    ret <4 x i32> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.qd.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovs_qd_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qd_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsqd %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x25,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsqd %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x25,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.qd.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.qd.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.128(<2 x i64>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmovus_qd_128(<2 x i64> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusqd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x15,0xc1]
; CHECK-NEXT:    vpmovusqd %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x15,0xc2]
; CHECK-NEXT:    vpmovusqd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x15,0xc0]
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc1]
; CHECK-NEXT:    vpaddd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.128(<2 x i64> %x0, <4 x i32> %x1, i8 -1)
    %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.128(<2 x i64> %x0, <4 x i32> %x1, i8 %x2)
    %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.128(<2 x i64> %x0, <4 x i32> zeroinitializer, i8 %x2)
    %res3 = add <4 x i32> %res0, %res1
    %res4 = add <4 x i32> %res3, %res2
    ret <4 x i32> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.qd.mem.128(i8* %ptr, <2 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovus_qd_mem_128(i8* %ptr, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qd_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusqd %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x15,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusqd %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x15,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.qd.mem.128(i8* %ptr, <2 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.qd.mem.128(i8* %ptr, <2 x i64> %x1, i8 %x2)
    ret void
}

declare <4 x i32> @llvm.x86.avx512.mask.pmov.qd.256(<4 x i64>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmov_qd_256(<4 x i64> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovqd %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x35,0xc1]
; CHECK-NEXT:    vpmovqd %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x35,0xc2]
; CHECK-NEXT:    vpmovqd %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x35,0xc0]
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc1]
; CHECK-NEXT:    vpaddd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <4 x i32> @llvm.x86.avx512.mask.pmov.qd.256(<4 x i64> %x0, <4 x i32> %x1, i8 -1)
    %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmov.qd.256(<4 x i64> %x0, <4 x i32> %x1, i8 %x2)
    %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmov.qd.256(<4 x i64> %x0, <4 x i32> zeroinitializer, i8 %x2)
    %res3 = add <4 x i32> %res0, %res1
    %res4 = add <4 x i32> %res3, %res2
    ret <4 x i32> %res4
}

declare void @llvm.x86.avx512.mask.pmov.qd.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmov_qd_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_qd_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovqd %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x35,0x07]
; CHECK-NEXT:    vpmovqd %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x35,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.qd.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.qd.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.256(<4 x i64>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmovs_qd_256(<4 x i64> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsqd %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x25,0xc1]
; CHECK-NEXT:    vpmovsqd %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x25,0xc2]
; CHECK-NEXT:    vpmovsqd %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x25,0xc0]
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc1]
; CHECK-NEXT:    vpaddd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.256(<4 x i64> %x0, <4 x i32> %x1, i8 -1)
    %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.256(<4 x i64> %x0, <4 x i32> %x1, i8 %x2)
    %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmovs.qd.256(<4 x i64> %x0, <4 x i32> zeroinitializer, i8 %x2)
    %res3 = add <4 x i32> %res0, %res1
    %res4 = add <4 x i32> %res3, %res2
    ret <4 x i32> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.qd.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovs_qd_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_qd_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsqd %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x25,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsqd %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x25,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.qd.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.qd.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.256(<4 x i64>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmovus_qd_256(<4 x i64> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusqd %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x15,0xc1]
; CHECK-NEXT:    vpmovusqd %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x15,0xc2]
; CHECK-NEXT:    vpmovusqd %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x15,0xc0]
; CHECK-NEXT:    vpaddd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc1]
; CHECK-NEXT:    vpaddd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.256(<4 x i64> %x0, <4 x i32> %x1, i8 -1)
    %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.256(<4 x i64> %x0, <4 x i32> %x1, i8 %x2)
    %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmovus.qd.256(<4 x i64> %x0, <4 x i32> zeroinitializer, i8 %x2)
    %res3 = add <4 x i32> %res0, %res1
    %res4 = add <4 x i32> %res3, %res2
    ret <4 x i32> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.qd.mem.256(i8* %ptr, <4 x i64>, i8)

define void @test_int_x86_avx512_mask_pmovus_qd_mem_256(i8* %ptr, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_qd_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusqd %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x15,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusqd %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x15,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.qd.mem.256(i8* %ptr, <4 x i64> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.qd.mem.256(i8* %ptr, <4 x i64> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmov.db.128(<4 x i32>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmov_db_128(<4 x i32> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_db_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovdb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x31,0xc1]
; CHECK-NEXT:    vpmovdb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x31,0xc2]
; CHECK-NEXT:    vpmovdb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x31,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmov.db.128(<4 x i32> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmov.db.128(<4 x i32> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmov.db.128(<4 x i32> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmov.db.mem.128(i8* %ptr, <4 x i32>, i8)

define void @test_int_x86_avx512_mask_pmov_db_mem_128(i8* %ptr, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_db_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovdb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x31,0x07]
; CHECK-NEXT:    vpmovdb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x31,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.db.mem.128(i8* %ptr, <4 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.db.mem.128(i8* %ptr, <4 x i32> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmovs.db.128(<4 x i32>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmovs_db_128(<4 x i32> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_db_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsdb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x21,0xc1]
; CHECK-NEXT:    vpmovsdb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x21,0xc2]
; CHECK-NEXT:    vpmovsdb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x21,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.db.128(<4 x i32> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.db.128(<4 x i32> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.db.128(<4 x i32> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.db.mem.128(i8* %ptr, <4 x i32>, i8)

define void @test_int_x86_avx512_mask_pmovs_db_mem_128(i8* %ptr, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_db_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsdb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x21,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsdb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x21,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.db.mem.128(i8* %ptr, <4 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.db.mem.128(i8* %ptr, <4 x i32> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmovus.db.128(<4 x i32>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmovus_db_128(<4 x i32> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_db_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusdb %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x11,0xc1]
; CHECK-NEXT:    vpmovusdb %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x11,0xc2]
; CHECK-NEXT:    vpmovusdb %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x11,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.db.128(<4 x i32> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.db.128(<4 x i32> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.db.128(<4 x i32> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.db.mem.128(i8* %ptr, <4 x i32>, i8)

define void @test_int_x86_avx512_mask_pmovus_db_mem_128(i8* %ptr, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_db_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusdb %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x11,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusdb %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x11,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.db.mem.128(i8* %ptr, <4 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.db.mem.128(i8* %ptr, <4 x i32> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmov.db.256(<8 x i32>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmov_db_256(<8 x i32> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_db_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovdb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x31,0xc1]
; CHECK-NEXT:    vpmovdb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x31,0xc2]
; CHECK-NEXT:    vpmovdb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x31,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmov.db.256(<8 x i32> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmov.db.256(<8 x i32> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmov.db.256(<8 x i32> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmov.db.mem.256(i8* %ptr, <8 x i32>, i8)

define void @test_int_x86_avx512_mask_pmov_db_mem_256(i8* %ptr, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_db_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovdb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x31,0x07]
; CHECK-NEXT:    vpmovdb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x31,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.db.mem.256(i8* %ptr, <8 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.db.mem.256(i8* %ptr, <8 x i32> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmovs.db.256(<8 x i32>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmovs_db_256(<8 x i32> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_db_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsdb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x21,0xc1]
; CHECK-NEXT:    vpmovsdb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x21,0xc2]
; CHECK-NEXT:    vpmovsdb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x21,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.db.256(<8 x i32> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.db.256(<8 x i32> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovs.db.256(<8 x i32> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.db.mem.256(i8* %ptr, <8 x i32>, i8)

define void @test_int_x86_avx512_mask_pmovs_db_mem_256(i8* %ptr, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_db_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsdb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x21,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsdb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x21,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.db.mem.256(i8* %ptr, <8 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.db.mem.256(i8* %ptr, <8 x i32> %x1, i8 %x2)
    ret void
}

declare <16 x i8> @llvm.x86.avx512.mask.pmovus.db.256(<8 x i32>, <16 x i8>, i8)

define <16 x i8>@test_int_x86_avx512_mask_pmovus_db_256(<8 x i32> %x0, <16 x i8> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_db_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusdb %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x11,0xc1]
; CHECK-NEXT:    vpmovusdb %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x11,0xc2]
; CHECK-NEXT:    vpmovusdb %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x11,0xc0]
; CHECK-NEXT:    vpaddb %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc1]
; CHECK-NEXT:    vpaddb %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfc,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.db.256(<8 x i32> %x0, <16 x i8> %x1, i8 -1)
    %res1 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.db.256(<8 x i32> %x0, <16 x i8> %x1, i8 %x2)
    %res2 = call <16 x i8> @llvm.x86.avx512.mask.pmovus.db.256(<8 x i32> %x0, <16 x i8> zeroinitializer, i8 %x2)
    %res3 = add <16 x i8> %res0, %res1
    %res4 = add <16 x i8> %res3, %res2
    ret <16 x i8> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.db.mem.256(i8* %ptr, <8 x i32>, i8)

define void @test_int_x86_avx512_mask_pmovus_db_mem_256(i8* %ptr, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_db_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusdb %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x11,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusdb %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x11,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.db.mem.256(i8* %ptr, <8 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.db.mem.256(i8* %ptr, <8 x i32> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmov.dw.128(<4 x i32>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmov_dw_128(<4 x i32> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_dw_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovdw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x33,0xc1]
; CHECK-NEXT:    vpmovdw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x33,0xc2]
; CHECK-NEXT:    vpmovdw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x33,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmov.dw.128(<4 x i32> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmov.dw.128(<4 x i32> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmov.dw.128(<4 x i32> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmov.dw.mem.128(i8* %ptr, <4 x i32>, i8)

define void @test_int_x86_avx512_mask_pmov_dw_mem_128(i8* %ptr, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_dw_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovdw %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x33,0x07]
; CHECK-NEXT:    vpmovdw %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x33,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.dw.mem.128(i8* %ptr, <4 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.dw.mem.128(i8* %ptr, <4 x i32> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.128(<4 x i32>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmovs_dw_128(<4 x i32> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_dw_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsdw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x23,0xc1]
; CHECK-NEXT:    vpmovsdw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x23,0xc2]
; CHECK-NEXT:    vpmovsdw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x23,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.128(<4 x i32> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.128(<4 x i32> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.128(<4 x i32> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.dw.mem.128(i8* %ptr, <4 x i32>, i8)

define void @test_int_x86_avx512_mask_pmovs_dw_mem_128(i8* %ptr, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_dw_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsdw %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x23,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsdw %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x23,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.dw.mem.128(i8* %ptr, <4 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.dw.mem.128(i8* %ptr, <4 x i32> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.128(<4 x i32>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmovus_dw_128(<4 x i32> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_dw_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusdw %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x13,0xc1]
; CHECK-NEXT:    vpmovusdw %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0x89,0x13,0xc2]
; CHECK-NEXT:    vpmovusdw %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x08,0x13,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.128(<4 x i32> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.128(<4 x i32> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.128(<4 x i32> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.dw.mem.128(i8* %ptr, <4 x i32>, i8)

define void @test_int_x86_avx512_mask_pmovus_dw_mem_128(i8* %ptr, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_dw_mem_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusdw %xmm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x08,0x13,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusdw %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x13,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.dw.mem.128(i8* %ptr, <4 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.dw.mem.128(i8* %ptr, <4 x i32> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmov.dw.256(<8 x i32>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmov_dw_256(<8 x i32> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_dw_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovdw %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x33,0xc1]
; CHECK-NEXT:    vpmovdw %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x33,0xc2]
; CHECK-NEXT:    vpmovdw %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x33,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmov.dw.256(<8 x i32> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmov.dw.256(<8 x i32> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmov.dw.256(<8 x i32> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmov.dw.mem.256(i8* %ptr, <8 x i32>, i8)

define void @test_int_x86_avx512_mask_pmov_dw_mem_256(i8* %ptr, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmov_dw_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovdw %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x33,0x07]
; CHECK-NEXT:    vpmovdw %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x33,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmov.dw.mem.256(i8* %ptr, <8 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmov.dw.mem.256(i8* %ptr, <8 x i32> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.256(<8 x i32>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmovs_dw_256(<8 x i32> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_dw_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsdw %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x23,0xc1]
; CHECK-NEXT:    vpmovsdw %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x23,0xc2]
; CHECK-NEXT:    vpmovsdw %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x23,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.256(<8 x i32> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.256(<8 x i32> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovs.dw.256(<8 x i32> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmovs.dw.mem.256(i8* %ptr, <8 x i32>, i8)

define void @test_int_x86_avx512_mask_pmovs_dw_mem_256(i8* %ptr, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovs_dw_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovsdw %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x23,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovsdw %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x23,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovs.dw.mem.256(i8* %ptr, <8 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovs.dw.mem.256(i8* %ptr, <8 x i32> %x1, i8 %x2)
    ret void
}

declare <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.256(<8 x i32>, <8 x i16>, i8)

define <8 x i16>@test_int_x86_avx512_mask_pmovus_dw_256(<8 x i32> %x0, <8 x i16> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_dw_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovusdw %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x13,0xc1]
; CHECK-NEXT:    vpmovusdw %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7e,0xa9,0x13,0xc2]
; CHECK-NEXT:    vpmovusdw %ymm0, %xmm0 ## encoding: [0x62,0xf2,0x7e,0x28,0x13,0xc0]
; CHECK-NEXT:    vpaddw %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc1]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    %res0 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.256(<8 x i32> %x0, <8 x i16> %x1, i8 -1)
    %res1 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.256(<8 x i32> %x0, <8 x i16> %x1, i8 %x2)
    %res2 = call <8 x i16> @llvm.x86.avx512.mask.pmovus.dw.256(<8 x i32> %x0, <8 x i16> zeroinitializer, i8 %x2)
    %res3 = add <8 x i16> %res0, %res1
    %res4 = add <8 x i16> %res3, %res2
    ret <8 x i16> %res4
}

declare void @llvm.x86.avx512.mask.pmovus.dw.mem.256(i8* %ptr, <8 x i32>, i8)

define void @test_int_x86_avx512_mask_pmovus_dw_mem_256(i8* %ptr, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovus_dw_mem_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpmovusdw %ymm0, (%rdi) ## encoding: [0x62,0xf2,0x7e,0x28,0x13,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpmovusdw %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x13,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.pmovus.dw.mem.256(i8* %ptr, <8 x i32> %x1, i8 -1)
    call void @llvm.x86.avx512.mask.pmovus.dw.mem.256(i8* %ptr, <8 x i32> %x1, i8 %x2)
    ret void
}

declare <2 x double> @llvm.x86.avx512.mask.cvtdq2pd.128(<4 x i32>, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_cvt_dq2pd_128(<4 x i32> %x0, <2 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_dq2pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtdq2pd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x09,0xe6,0xc8]
; CHECK-NEXT:    vcvtdq2pd %xmm0, %xmm0 ## encoding: [0xc5,0xfa,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.cvtdq2pd.128(<4 x i32> %x0, <2 x double> %x1, i8 %x2)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.cvtdq2pd.128(<4 x i32> %x0, <2 x double> %x1, i8 -1)
  %res2 = fadd <2 x double> %res, %res1
  ret <2 x double> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.cvtdq2pd.256(<4 x i32>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_cvt_dq2pd_256(<4 x i32> %x0, <4 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_dq2pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtdq2pd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x29,0xe6,0xc8]
; CHECK-NEXT:    vcvtdq2pd %xmm0, %ymm0 ## encoding: [0xc5,0xfe,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.cvtdq2pd.256(<4 x i32> %x0, <4 x double> %x1, i8 %x2)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.cvtdq2pd.256(<4 x i32> %x0, <4 x double> %x1, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.cvtdq2ps.128(<4 x i32>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_cvt_dq2ps_128(<4 x i32> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_dq2ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtdq2ps %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x5b,0xc8]
; CHECK-NEXT:    vcvtdq2ps %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,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.cvtdq2ps.128(<4 x i32> %x0, <4 x float> %x1, i8 %x2)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.cvtdq2ps.128(<4 x i32> %x0, <4 x float> %x1, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.cvtdq2ps.256(<8 x i32>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_cvt_dq2ps_256(<8 x i32> %x0, <8 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_dq2ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtdq2ps %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x5b,0xc8]
; CHECK-NEXT:    vcvtdq2ps %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x5b,0xc0]
; 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.cvtdq2ps.256(<8 x i32> %x0, <8 x float> %x1, i8 %x2)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.cvtdq2ps.256(<8 x i32> %x0, <8 x float> %x1, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvtpd2dq.128(<2 x double>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvt_pd2dq_128(<2 x double> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2dq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtpd2dq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xff,0x09,0xe6,0xc8]
; CHECK-NEXT:    vcvtpd2dq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xff,0x08,0xe6,0xc0]
; 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.cvtpd2dq.128(<2 x double> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvtpd2dq.128(<2 x double> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvtpd2dq.256(<4 x double>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvt_pd2dq_256(<4 x double> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2dq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtpd2dq %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xff,0x29,0xe6,0xc8]
; CHECK-NEXT:    vcvtpd2dq %ymm0, %xmm0 ## encoding: [0x62,0xf1,0xff,0x28,0xe6,0xc0]
; 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.cvtpd2dq.256(<4 x double> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvtpd2dq.256(<4 x double> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.cvtpd2ps.256(<4 x double>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_cvt_pd2ps_256(<4 x double> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtpd2ps %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x5a,0xc8]
; CHECK-NEXT:    vcvtpd2ps %ymm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x5a,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.cvtpd2ps.256(<4 x double> %x0, <4 x float> %x1, i8 %x2)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.cvtpd2ps.256(<4 x double> %x0, <4 x float> %x1, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.cvtpd2ps(<2 x double>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_cvt_pd2ps(<2 x double> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2ps:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtpd2ps %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x5a,0xc8]
; CHECK-NEXT:    vcvtpd2ps %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x5a,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.cvtpd2ps(<2 x double> %x0, <4 x float> %x1, i8 %x2)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.cvtpd2ps(<2 x double> %x0, <4 x float> %x1, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvtpd2udq.128(<2 x double>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvt_pd2udq_128(<2 x double> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2udq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtpd2udq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfc,0x09,0x79,0xc8]
; CHECK-NEXT:    vcvtpd2udq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfc,0x08,0x79,0xc0]
; 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.cvtpd2udq.128(<2 x double> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvtpd2udq.128(<2 x double> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvtpd2udq.256(<4 x double>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvt_pd2udq_256(<4 x double> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_pd2udq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtpd2udq %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfc,0x29,0x79,0xc8]
; CHECK-NEXT:    vcvtpd2udq %ymm0, %xmm0 ## encoding: [0x62,0xf1,0xfc,0x28,0x79,0xc0]
; 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.cvtpd2udq.256(<4 x double> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvtpd2udq.256(<4 x double> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvtps2dq.128(<4 x float>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvt_ps2dq_128(<4 x float> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2dq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtps2dq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x5b,0xc8]
; CHECK-NEXT:    vcvtps2dq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x5b,0xc0]
; 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.cvtps2dq.128(<4 x float> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvtps2dq.128(<4 x float> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.cvtps2dq.256(<8 x float>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_cvt_ps2dq_256(<8 x float> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2dq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtps2dq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x5b,0xc8]
; CHECK-NEXT:    vcvtps2dq %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x5b,0xc0]
; 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.cvtps2dq.256(<8 x float> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.cvtps2dq.256(<8 x float> %x0, <8 x i32> %x1, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x double> @llvm.x86.avx512.mask.cvtps2pd.128(<4 x float>, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_cvt_ps2pd_128(<4 x float> %x0, <2 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtps2pd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x5a,0xc8]
; CHECK-NEXT:    vcvtps2pd %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x5a,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.cvtps2pd.128(<4 x float> %x0, <2 x double> %x1, i8 %x2)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.cvtps2pd.128(<4 x float> %x0, <2 x double> %x1, i8 -1)
  %res2 = fadd <2 x double> %res, %res1
  ret <2 x double> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.cvtps2pd.256(<4 x float>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_cvt_ps2pd_256(<4 x float> %x0, <4 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtps2pd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x5a,0xc8]
; CHECK-NEXT:    vcvtps2pd %xmm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x5a,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.cvtps2pd.256(<4 x float> %x0, <4 x double> %x1, i8 %x2)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.cvtps2pd.256(<4 x float> %x0, <4 x double> %x1, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvtps2udq.128(<4 x float>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvt_ps2udq_128(<4 x float> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2udq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtps2udq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x79,0xc8]
; CHECK-NEXT:    vcvtps2udq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x79,0xc0]
; 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.cvtps2udq.128(<4 x float> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvtps2udq.128(<4 x float> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.cvtps2udq.256(<8 x float>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_cvt_ps2udq_256(<8 x float> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_ps2udq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtps2udq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x79,0xc8]
; CHECK-NEXT:    vcvtps2udq %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x79,0xc0]
; 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.cvtps2udq.256(<8 x float> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.cvtps2udq.256(<8 x float> %x0, <8 x i32> %x1, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvttpd2dq.128(<2 x double>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvtt_pd2dq_128(<2 x double> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_pd2dq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvttpd2dq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0xe6,0xc8]
; CHECK-NEXT:    vcvttpd2dq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xe6,0xc0]
; 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.cvttpd2dq.128(<2 x double> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvttpd2dq.128(<2 x double> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvttpd2dq.256(<4 x double>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvtt_pd2dq_256(<4 x double> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_pd2dq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvttpd2dq %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xe6,0xc8]
; CHECK-NEXT:    vcvttpd2dq %ymm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xe6,0xc0]
; 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.cvttpd2dq.256(<4 x double> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvttpd2dq.256(<4 x double> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvttpd2udq.128(<2 x double>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvtt_pd2udq_128(<2 x double> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_pd2udq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvttpd2udq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfc,0x09,0x78,0xc8]
; CHECK-NEXT:    vcvttpd2udq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfc,0x08,0x78,0xc0]
; 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.cvttpd2udq.128(<2 x double> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvttpd2udq.128(<2 x double> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvttpd2udq.256(<4 x double>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvtt_pd2udq_256(<4 x double> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_pd2udq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvttpd2udq %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfc,0x29,0x78,0xc8]
; CHECK-NEXT:    vcvttpd2udq %ymm0, %xmm0 ## encoding: [0x62,0xf1,0xfc,0x28,0x78,0xc0]
; 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.cvttpd2udq.256(<4 x double> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvttpd2udq.256(<4 x double> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvttps2dq.128(<4 x float>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvtt_ps2dq_128(<4 x float> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_ps2dq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvttps2dq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x09,0x5b,0xc8]
; CHECK-NEXT:    vcvttps2dq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7e,0x08,0x5b,0xc0]
; 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.cvttps2dq.128(<4 x float> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvttps2dq.128(<4 x float> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.cvttps2dq.256(<8 x float>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_cvtt_ps2dq_256(<8 x float> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_ps2dq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvttps2dq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x29,0x5b,0xc8]
; CHECK-NEXT:    vcvttps2dq %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7e,0x28,0x5b,0xc0]
; 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.cvttps2dq.256(<8 x float> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.cvttps2dq.256(<8 x float> %x0, <8 x i32> %x1, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.cvttps2udq.128(<4 x float>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_cvtt_ps2udq_128(<4 x float> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_ps2udq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvttps2udq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x78,0xc8]
; CHECK-NEXT:    vcvttps2udq %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x78,0xc0]
; 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.cvttps2udq.128(<4 x float> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.cvttps2udq.128(<4 x float> %x0, <4 x i32> %x1, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.cvttps2udq.256(<8 x float>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_cvtt_ps2udq_256(<8 x float> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvtt_ps2udq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvttps2udq %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x78,0xc8]
; CHECK-NEXT:    vcvttps2udq %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x78,0xc0]
; 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.cvttps2udq.256(<8 x float> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.cvttps2udq.256(<8 x float> %x0, <8 x i32> %x1, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x double> @llvm.x86.avx512.mask.cvtudq2pd.128(<4 x i32>, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_cvt_udq2pd_128(<4 x i32> %x0, <2 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_udq2pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtudq2pd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x09,0x7a,0xc8]
; CHECK-NEXT:    vcvtudq2pd %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7e,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.cvtudq2pd.128(<4 x i32> %x0, <2 x double> %x1, i8 %x2)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.cvtudq2pd.128(<4 x i32> %x0, <2 x double> %x1, i8 -1)
  %res2 = fadd <2 x double> %res, %res1
  ret <2 x double> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.cvtudq2pd.256(<4 x i32>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_cvt_udq2pd_256(<4 x i32> %x0, <4 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_udq2pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtudq2pd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x29,0x7a,0xc8]
; CHECK-NEXT:    vcvtudq2pd %xmm0, %ymm0 ## encoding: [0x62,0xf1,0x7e,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.cvtudq2pd.256(<4 x i32> %x0, <4 x double> %x1, i8 %x2)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.cvtudq2pd.256(<4 x i32> %x0, <4 x double> %x1, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.cvtudq2ps.128(<4 x i32>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_cvt_udq2ps_128(<4 x i32> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_udq2ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtudq2ps %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7f,0x09,0x7a,0xc8]
; CHECK-NEXT:    vcvtudq2ps %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7f,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.cvtudq2ps.128(<4 x i32> %x0, <4 x float> %x1, i8 %x2)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.cvtudq2ps.128(<4 x i32> %x0, <4 x float> %x1, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.cvtudq2ps.256(<8 x i32>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_cvt_udq2ps_256(<8 x i32> %x0, <8 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_cvt_udq2ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtudq2ps %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7f,0x29,0x7a,0xc8]
; CHECK-NEXT:    vcvtudq2ps %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7f,0x28,0x7a,0xc0]
; 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.cvtudq2ps.256(<8 x i32> %x0, <8 x float> %x1, i8 %x2)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.cvtudq2ps.256(<8 x i32> %x0, <8 x float> %x1, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <2 x double> @llvm.x86.avx512.mask.rndscale.pd.128(<2 x double>, i32, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_rndscale_pd_128(<2 x double> %x0, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_rndscale_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrndscalepd $4, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x09,0xc8,0x04]
; CHECK-NEXT:    vrndscalepd $88, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0xfd,0x08,0x09,0xc0,0x58]
; 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.rndscale.pd.128(<2 x double> %x0, i32 4, <2 x double> %x2, i8 %x3)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.rndscale.pd.128(<2 x double> %x0, i32 88, <2 x double> %x2, i8 -1)
  %res2 = fadd <2 x double> %res, %res1
  ret <2 x double> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.rndscale.pd.256(<4 x double>, i32, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_rndscale_pd_256(<4 x double> %x0, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_rndscale_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrndscalepd $4, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x09,0xc8,0x04]
; CHECK-NEXT:    vrndscalepd $88, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x09,0xc0,0x58]
; 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.rndscale.pd.256(<4 x double> %x0, i32 4, <4 x double> %x2, i8 %x3)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.rndscale.pd.256(<4 x double> %x0, i32 88, <4 x double> %x2, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.rndscale.ps.128(<4 x float>, i32, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_rndscale_ps_128(<4 x float> %x0, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_rndscale_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrndscaleps $88, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x08,0xc8,0x58]
; CHECK-NEXT:    vrndscaleps $4, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x08,0xc0,0x04]
; 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.rndscale.ps.128(<4 x float> %x0, i32 88, <4 x float> %x2, i8 %x3)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.rndscale.ps.128(<4 x float> %x0, i32 4, <4 x float> %x2, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.rndscale.ps.256(<8 x float>, i32, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_rndscale_ps_256(<8 x float> %x0, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_rndscale_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrndscaleps $5, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x08,0xc8,0x05]
; CHECK-NEXT:    vrndscaleps $66, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x08,0xc0,0x42]
; 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.rndscale.ps.256(<8 x float> %x0, i32 5, <8 x float> %x2, i8 %x3)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.rndscale.ps.256(<8 x float> %x0, i32 66, <8 x float> %x2, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.shuf.f32x4.256(<8 x float>, <8 x float>, i32, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_shuf_f32x4_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_shuf_f32x4_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshuff32x4 $22, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x23,0xd1,0x16]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[0,1,2,3],ymm1[4,5,6,7]
; CHECK-NEXT:    vshuff32x4 $22, %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x23,0xd9,0x16]
; CHECK-NEXT:    ## ymm3 {%k1} {z} = ymm0[0,1,2,3],ymm1[4,5,6,7]
; CHECK-NEXT:    vshuff32x4 $22, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x23,0xc1,0x16]
; CHECK-NEXT:    ## ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xc0]
; 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.shuf.f32x4.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> %x3, i8 %x4)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.shuf.f32x4.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> %x3, i8 -1)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.shuf.f32x4.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> zeroinitializer, i8 %x4)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res2, %res3
  ret <8 x float> %res4
}

declare <4 x double> @llvm.x86.avx512.mask.shuf.f64x2.256(<4 x double>, <4 x double>, i32, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_shuf_f64x2_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_shuf_f64x2_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshuff64x2 $22, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x23,0xd1,0x16]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[0,1],ymm1[2,3]
; CHECK-NEXT:    vshuff64x2 $22, %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x23,0xd9,0x16]
; CHECK-NEXT:    ## ymm3 {%k1} {z} = ymm0[0,1],ymm1[2,3]
; CHECK-NEXT:    vshuff64x2 $22, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x23,0xc1,0x16]
; CHECK-NEXT:    ## ymm0 = ymm0[0,1],ymm1[2,3]
; 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.shuf.f64x2.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> %x3, i8 %x4)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.shuf.f64x2.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> %x3, i8 -1)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.shuf.f64x2.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> zeroinitializer, i8 %x4)
  %res3 = fadd <4 x double> %res, %res1
  %res4 = fadd <4 x double> %res2, %res3
  ret <4 x double> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.shuf.i32x4.256(<8 x i32>, <8 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_shuf_i32x4_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_shuf_i32x4_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshufi32x4 $22, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x43,0xd1,0x16]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[0,1,2,3],ymm1[4,5,6,7]
; CHECK-NEXT:    vshufi32x4 $22, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x43,0xc1,0x16]
; CHECK-NEXT:    ## ymm0 = ymm0[0,1,2,3],ymm1[4,5,6,7]
; 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.shuf.i32x4.256(<8 x i32> %x0, <8 x i32> %x1, i32 22, <8 x i32> %x3, i8 %x4)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.shuf.i32x4.256(<8 x i32> %x0, <8 x i32> %x1, i32 22, <8 x i32> %x3, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.shuf.i64x2.256(<4 x i64>, <4 x i64>, i32, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_shuf_i64x2_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_shuf_i64x2_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshufi64x2 $22, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x43,0xd1,0x16]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[0,1],ymm1[2,3]
; CHECK-NEXT:    vshufi64x2 $22, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x43,0xc1,0x16]
; CHECK-NEXT:    ## ymm0 = ymm0[0,1],ymm1[2,3]
; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.shuf.i64x2.256(<4 x i64> %x0, <4 x i64> %x1, i32 22, <4 x i64> %x3, i8 %x4)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.shuf.i64x2.256(<4 x i64> %x0, <4 x i64> %x1, i32 22, <4 x i64> %x3, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.vextractf32x4.256(<8 x float>, i32, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_vextractf32x4_256(<8 x float> %x0, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vextractf32x4_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vextractf32x4 $1, %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x19,0xc1,0x01]
; CHECK-NEXT:    vextractf32x4 $1, %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x19,0xc2,0x01]
; CHECK-NEXT:    vextractf32x4 $1, %ymm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x19,0xc0,0x01]
; CHECK-NEXT:    vaddps %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0x74,0x08,0x58,0xca]
; CHECK-NEXT:    vaddps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res  = call <4 x float> @llvm.x86.avx512.mask.vextractf32x4.256(<8 x float> %x0, i32 1, <4 x float> %x2, i8 %x3)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.vextractf32x4.256(<8 x float> %x0, i32 1, <4 x float> zeroinitializer, i8 %x3)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.vextractf32x4.256(<8 x float> %x0, i32 1, <4 x float> zeroinitializer, i8 -1)
  %res3 = fadd <4 x float> %res, %res1
  %res4 = fadd <4 x float> %res2, %res3
  ret <4 x float> %res4
}

declare <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double>, i32, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_getmant_pd_128(<2 x double> %x0, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_getmant_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vgetmantpd $11, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x26,0xc8,0x0b]
; CHECK-NEXT:    vgetmantpd $11, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0x89,0x26,0xd0,0x0b]
; CHECK-NEXT:    vgetmantpd $11, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0xfd,0x08,0x26,0xc0,0x0b]
; CHECK-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0x58,0xc0]
; 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.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> %x2, i8 %x3)
  %res2 = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> zeroinitializer, i8 %x3)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.getmant.pd.128(<2 x double> %x0, i32 11, <2 x double> %x2, i8 -1)
  %res3 = fadd <2 x double> %res, %res1
  %res4 = fadd <2 x double> %res2, %res3
  ret <2 x double> %res4
}

declare <4 x double> @llvm.x86.avx512.mask.getmant.pd.256(<4 x double>, i32, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_getmant_pd_256(<4 x double> %x0, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_getmant_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vgetmantpd $11, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x26,0xc8,0x0b]
; CHECK-NEXT:    vgetmantpd $11, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x26,0xc0,0x0b]
; 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.getmant.pd.256(<4 x double> %x0, i32 11, <4 x double> %x2, i8 %x3)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.getmant.pd.256(<4 x double> %x0, i32 11, <4 x double> %x2, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.getmant.ps.128(<4 x float>, i32, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_getmant_ps_128(<4 x float> %x0, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_getmant_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vgetmantps $11, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x26,0xc8,0x0b]
; CHECK-NEXT:    vgetmantps $11, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x26,0xc0,0x0b]
; 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.getmant.ps.128(<4 x float> %x0, i32 11, <4 x float> %x2, i8 %x3)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.getmant.ps.128(<4 x float> %x0, i32 11, <4 x float> %x2, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.getmant.ps.256(<8 x float>, i32, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_getmant_ps_256(<8 x float> %x0, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_getmant_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vgetmantps $11, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x26,0xc8,0x0b]
; CHECK-NEXT:    vgetmantps $11, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x26,0xc0,0x0b]
; 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.getmant.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 %x3)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.getmant.ps.256(<8 x float> %x0, i32 11, <8 x float> %x2, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <2 x double> @llvm.x86.avx512.mask.shuf.pd.128(<2 x double>, <2 x double>, i32, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_shuf_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_shuf_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshufpd $22, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0xc6,0xd1,0x16]
; CHECK-NEXT:    ## xmm2 {%k1} = xmm0[0],xmm1[1]
; CHECK-NEXT:    vshufpd $22, %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0xc6,0xd9,0x16]
; CHECK-NEXT:    ## xmm3 {%k1} {z} = xmm0[0],xmm1[1]
; CHECK-NEXT:    vshufpd $22, %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xc6,0xc1,0x16]
; CHECK-NEXT:    ## xmm0 = xmm0[0],xmm1[1]
; CHECK-NEXT:    vaddpd %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0x58,0xc0]
; 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.shuf.pd.128(<2 x double> %x0, <2 x double> %x1, i32 22, <2 x double> %x3, i8 %x4)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.shuf.pd.128(<2 x double> %x0, <2 x double> %x1, i32 22, <2 x double> %x3, i8 -1)
  %res2 = call <2 x double> @llvm.x86.avx512.mask.shuf.pd.128(<2 x double> %x0, <2 x double> %x1, i32 22, <2 x double> zeroinitializer, i8 %x4)
  %res3 = fadd <2 x double> %res, %res1
  %res4 = fadd <2 x double> %res2, %res3
  ret <2 x double> %res4
}

declare <4 x double> @llvm.x86.avx512.mask.shuf.pd.256(<4 x double>, <4 x double>, i32, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_shuf_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_shuf_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshufpd $22, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xc6,0xd1,0x16]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[0],ymm1[1],ymm0[3],ymm1[2]
; CHECK-NEXT:    vshufpd $22, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xc6,0xc1,0x16]
; CHECK-NEXT:    ## ymm0 = ymm0[0],ymm1[1],ymm0[3],ymm1[2]
; 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.shuf.pd.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> %x3, i8 %x4)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.shuf.pd.256(<4 x double> %x0, <4 x double> %x1, i32 22, <4 x double> %x3, i8 -1)
  %res2 = fadd <4 x double> %res, %res1
  ret <4 x double> %res2
}

declare <4 x float> @llvm.x86.avx512.mask.shuf.ps.128(<4 x float>, <4 x float>, i32, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_shuf_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_shuf_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshufps $22, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0xc6,0xd1,0x16]
; CHECK-NEXT:    ## xmm2 {%k1} = xmm0[2,1],xmm1[1,0]
; CHECK-NEXT:    vshufps $22, %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0xc6,0xc1,0x16]
; CHECK-NEXT:    ## xmm0 = xmm0[2,1],xmm1[1,0]
; 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.shuf.ps.128(<4 x float> %x0, <4 x float> %x1, i32 22, <4 x float> %x3, i8 %x4)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.shuf.ps.128(<4 x float> %x0, <4 x float> %x1, i32 22, <4 x float> %x3, i8 -1)
  %res2 = fadd <4 x float> %res, %res1
  ret <4 x float> %res2
}

declare <8 x float> @llvm.x86.avx512.mask.shuf.ps.256(<8 x float>, <8 x float>, i32, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_shuf_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_shuf_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshufps $22, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0xc6,0xd1,0x16]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[2,1],ymm1[1,0],ymm0[6,5],ymm1[5,4]
; CHECK-NEXT:    vshufps $22, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0xc6,0xc1,0x16]
; CHECK-NEXT:    ## ymm0 = ymm0[2,1],ymm1[1,0],ymm0[6,5],ymm1[5,4]
; 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.shuf.ps.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> %x3, i8 %x4)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.shuf.ps.256(<8 x float> %x0, <8 x float> %x1, i32 22, <8 x float> %x3, i8 -1)
  %res2 = fadd <8 x float> %res, %res1
  ret <8 x float> %res2
}

declare <4 x i32> @llvm.x86.avx512.mask.valign.d.128(<4 x i32>, <4 x i32>, i32, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_valign_d_128(<4 x i32> %x0, <4 x i32> %x1,<4 x i32> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_valign_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    valignd $22, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x03,0xd1,0x16]
; CHECK-NEXT:    valignd $22, %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0x89,0x03,0xd9,0x16]
; CHECK-NEXT:    valignd $22, %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x03,0xc1,0x16]
; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xc0]
; CHECK-NEXT:    vpaddd %xmm3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.valign.d.128(<4 x i32> %x0, <4 x i32> %x1, i32 22, <4 x i32> %x3, i8 %x4)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.valign.d.128(<4 x i32> %x0, <4 x i32> %x1, i32 22, <4 x i32> %x3, i8 -1)
    %res2 = call <4 x i32> @llvm.x86.avx512.mask.valign.d.128(<4 x i32> %x0, <4 x i32> %x1, i32 22, <4 x i32> zeroinitializer,i8 %x4)
  %res3 = add <4 x i32> %res, %res1
    %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.valign.d.256(<8 x i32>, <8 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_valign_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_valign_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    valignd $22, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x03,0xd1,0x16]
; CHECK-NEXT:    valignd $22, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x03,0xc1,0x16]
; 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.valign.d.256(<8 x i32> %x0, <8 x i32> %x1, i32 22, <8 x i32> %x3, i8 %x4)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.valign.d.256(<8 x i32> %x0, <8 x i32> %x1, i32 22, <8 x i32> %x3, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.valign.q.128(<2 x i64>, <2 x i64>, i32, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_valign_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_valign_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    valignq $22, %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x03,0xd1,0x16]
; CHECK-NEXT:    valignq $22, %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0xfd,0x08,0x03,0xc1,0x16]
; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.valign.q.128(<2 x i64> %x0, <2 x i64> %x1, i32 22, <2 x i64> %x3, i8 %x4)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.valign.q.128(<2 x i64> %x0, <2 x i64> %x1, i32 22, <2 x i64> %x3, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.valign.q.256(<4 x i64>, <4 x i64>, i32, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_valign_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_valign_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    valignq $22, %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x03,0xd1,0x16]
; CHECK-NEXT:    valignq $22, %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0xfd,0x28,0x03,0xc1,0x16]
; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.valign.q.256(<4 x i64> %x0, <4 x i64> %x1, i32 22, <4 x i64> %x3, i8 %x4)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.valign.q.256(<4 x i64> %x0, <4 x i64> %x1, i32 22, <4 x i64> %x3, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x double> @llvm.x86.avx512.mask.vpermilvar.pd.256(<4 x double>, <4 x i64>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_vpermilvar_pd_256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermilvar_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermilpd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x0d,0xd1]
; CHECK-NEXT:    vpermilpd %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x0d,0xd9]
; CHECK-NEXT:    vpermilpd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x0d,0xc1]
; CHECK-NEXT:    vaddpd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0x58,0xcb]
; CHECK-NEXT:    vaddpd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.vpermilvar.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.vpermilvar.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> zeroinitializer, i8 %x3)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.vpermilvar.pd.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 -1)
  %res3 = fadd <4 x double> %res, %res1
  %res4 = fadd <4 x double> %res2, %res3
  ret <4 x double> %res4
}

declare <2 x double> @llvm.x86.avx512.mask.vpermilvar.pd.128(<2 x double>, <2 x i64>, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_vpermilvar_pd_128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermilvar_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermilpd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x0d,0xd1]
; CHECK-NEXT:    vpermilpd %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x0d,0xd9]
; CHECK-NEXT:    vpermilpd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x0d,0xc1]
; CHECK-NEXT:    vaddpd %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0xed,0x08,0x58,0xcb]
; 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.vpermilvar.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 %x3)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.vpermilvar.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> zeroinitializer, i8 %x3)
  %res2 = call <2 x double> @llvm.x86.avx512.mask.vpermilvar.pd.128(<2 x double> %x0, <2 x i64> %x1, <2 x double> %x2, i8 -1)
  %res3 = fadd <2 x double> %res, %res1
  %res4 = fadd <2 x double> %res3, %res2
  ret <2 x double> %res4
}

declare <8 x float> @llvm.x86.avx512.mask.vpermilvar.ps.256(<8 x float>, <8 x i32>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_vpermilvar_ps_256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermilvar_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermilps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x0c,0xd1]
; CHECK-NEXT:    vpermilps %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x0c,0xd9]
; CHECK-NEXT:    vpermilps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x0c,0xc1]
; CHECK-NEXT:    vaddps %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xcb]
; 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.vpermilvar.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.vpermilvar.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> zeroinitializer, i8 %x3)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.vpermilvar.ps.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 -1)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res3, %res2
  ret <8 x float> %res4
}

declare <4 x float> @llvm.x86.avx512.mask.vpermilvar.ps.128(<4 x float>, <4 x i32>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_vpermilvar_ps_128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermilvar_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermilps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x0c,0xd1]
; CHECK-NEXT:    vpermilps %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x0c,0xd9]
; CHECK-NEXT:    vpermilps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x0c,0xc1]
; CHECK-NEXT:    vaddps %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6c,0x08,0x58,0xcb]
; CHECK-NEXT:    vaddps %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.vpermilvar.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 %x3)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.vpermilvar.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> zeroinitializer, i8 %x3)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.vpermilvar.ps.128(<4 x float> %x0, <4 x i32> %x1, <4 x float> %x2, i8 -1)
  %res3 = fadd <4 x float> %res, %res1
  %res4 = fadd <4 x float> %res2, %res3
  ret <4 x float> %res4
}

declare <8 x float> @llvm.x86.avx512.mask.insertf32x4.256(<8 x float>, <4 x float>, i32, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_insertf32x4_256(<8 x float> %x0, <4 x float> %x1, <8 x float> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_insertf32x4_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vinsertf32x4 $1, %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x18,0xd1,0x01]
; CHECK-NEXT:    vinsertf32x4 $1, %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x18,0xd9,0x01]
; CHECK-NEXT:    vinsertf32x4 $1, %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x18,0xc1,0x01]
; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xc0]
; 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.insertf32x4.256(<8 x float> %x0, <4 x float> %x1, i32 1, <8 x float> %x3, i8 %x4)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.insertf32x4.256(<8 x float> %x0, <4 x float> %x1, i32 1, <8 x float> %x3, i8 -1)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.insertf32x4.256(<8 x float> %x0, <4 x float> %x1, i32 1, <8 x float> zeroinitializer, i8 %x4)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res2, %res3
  ret <8 x float> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.inserti32x4.256(<8 x i32>, <4 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_inserti32x4_256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x3, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_inserti32x4_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vinserti32x4 $1, %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x38,0xd1,0x01]
; CHECK-NEXT:    vinserti32x4 $1, %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x38,0xd9,0x01]
; CHECK-NEXT:    vinserti32x4 $1, %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x38,0xc1,0x01]
; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xc0]
; CHECK-NEXT:    vpaddd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfe,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]

  %res = call <8 x i32> @llvm.x86.avx512.mask.inserti32x4.256(<8 x i32> %x0, <4 x i32> %x1, i32 1, <8 x i32> %x3, i8 %x4)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.inserti32x4.256(<8 x i32> %x0, <4 x i32> %x1, i32 1, <8 x i32> %x3, i8 -1)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.inserti32x4.256(<8 x i32> %x0, <4 x i32> %x1, i32 1, <8 x i32> zeroinitializer, i8 %x4)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res2, %res3
  ret <8 x i32> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.pternlog.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i32, i8)

define <4 x i32>@test_int_x86_avx512_mask_pternlog_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_pternlog_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
; CHECK-NEXT:    vpternlogd $33, %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf3,0x75,0x09,0x25,0xda,0x21]
; CHECK-NEXT:    vpternlogd $33, %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf3,0x75,0x08,0x25,0xc2,0x21]
; CHECK-NEXT:    vpaddd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x65,0x08,0xfe,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33, i8 %x4)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <4 x i32> @llvm.x86.avx512.maskz.pternlog.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i32, i8)

define <4 x i32>@test_int_x86_avx512_maskz_pternlog_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_maskz_pternlog_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
; CHECK-NEXT:    vpternlogd $33, %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0x89,0x25,0xda,0x21]
; CHECK-NEXT:    vpternlogd $33, %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf3,0x75,0x08,0x25,0xc2,0x21]
; CHECK-NEXT:    vpaddd %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x65,0x08,0xfe,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.maskz.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33, i8 %x4)
  %res1 = call <4 x i32> @llvm.x86.avx512.maskz.pternlog.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i32 33, i8 -1)
  %res2 = add <4 x i32> %res, %res1
  ret <4 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.pternlog.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i32, i8)

define <8 x i32>@test_int_x86_avx512_mask_pternlog_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_pternlog_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
; CHECK-NEXT:    vpternlogd $33, %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf3,0x75,0x29,0x25,0xda,0x21]
; CHECK-NEXT:    vpternlogd $33, %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0x75,0x28,0x25,0xc2,0x21]
; CHECK-NEXT:    vpaddd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfe,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33, i8 %x4)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <8 x i32> @llvm.x86.avx512.maskz.pternlog.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i32, i8)

define <8 x i32>@test_int_x86_avx512_maskz_pternlog_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_maskz_pternlog_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
; CHECK-NEXT:    vpternlogd $33, %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0xa9,0x25,0xda,0x21]
; CHECK-NEXT:    vpternlogd $33, %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0x75,0x28,0x25,0xc2,0x21]
; CHECK-NEXT:    vpaddd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfe,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.maskz.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33, i8 %x4)
  %res1 = call <8 x i32> @llvm.x86.avx512.maskz.pternlog.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i32 33, i8 -1)
  %res2 = add <8 x i32> %res, %res1
  ret <8 x i32> %res2
}

declare <2 x i64> @llvm.x86.avx512.mask.pternlog.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i32, i8)

define <2 x i64>@test_int_x86_avx512_mask_pternlog_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_pternlog_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
; CHECK-NEXT:    vpternlogq $33, %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf3,0xf5,0x09,0x25,0xda,0x21]
; CHECK-NEXT:    vpternlogq $33, %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf3,0xf5,0x08,0x25,0xc2,0x21]
; CHECK-NEXT:    vpaddq %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33, i8 %x4)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <2 x i64> @llvm.x86.avx512.maskz.pternlog.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i32, i8)

define <2 x i64>@test_int_x86_avx512_maskz_pternlog_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_maskz_pternlog_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
; CHECK-NEXT:    vpternlogq $33, %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0x89,0x25,0xda,0x21]
; CHECK-NEXT:    vpternlogq $33, %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf3,0xf5,0x08,0x25,0xc2,0x21]
; CHECK-NEXT:    vpaddq %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0xe5,0x08,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.maskz.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33, i8 %x4)
  %res1 = call <2 x i64> @llvm.x86.avx512.maskz.pternlog.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i32 33, i8 -1)
  %res2 = add <2 x i64> %res, %res1
  ret <2 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.mask.pternlog.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i32, i8)

define <4 x i64>@test_int_x86_avx512_mask_pternlog_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_pternlog_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
; CHECK-NEXT:    vpternlogq $33, %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf3,0xf5,0x29,0x25,0xda,0x21]
; CHECK-NEXT:    vpternlogq $33, %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0xf5,0x28,0x25,0xc2,0x21]
; CHECK-NEXT:    vpaddq %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33, i8 %x4)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

declare <4 x i64> @llvm.x86.avx512.maskz.pternlog.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i32, i8)

define <4 x i64>@test_int_x86_avx512_maskz_pternlog_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_maskz_pternlog_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
; CHECK-NEXT:    vpternlogq $33, %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0xa9,0x25,0xda,0x21]
; CHECK-NEXT:    vpternlogq $33, %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0xf5,0x28,0x25,0xc2,0x21]
; CHECK-NEXT:    vpaddq %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0xe5,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.maskz.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33, i8 %x4)
  %res1 = call <4 x i64> @llvm.x86.avx512.maskz.pternlog.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i32 33, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

define <4 x float> @test_x86_vcvtph2ps_128(<8 x i16> %a0) {
; CHECK-LABEL: test_x86_vcvtph2ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vcvtph2ps %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x13,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128(<8 x i16> %a0, <4 x float> zeroinitializer, i8 -1)
  ret <4 x float> %res
}

define <4 x float> @test_x86_vcvtph2ps_128_rrk(<8 x i16> %a0,<4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_x86_vcvtph2ps_128_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtph2ps %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x13,0xc8]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128(<8 x i16> %a0, <4 x float> %a1, i8 %mask)
  ret <4 x float> %res
}


define <4 x float> @test_x86_vcvtph2ps_128_rrkz(<8 x i16> %a0, i8 %mask) {
; CHECK-LABEL: test_x86_vcvtph2ps_128_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtph2ps %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x13,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128(<8 x i16> %a0, <4 x float> zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

declare <4 x float> @llvm.x86.avx512.mask.vcvtph2ps.128(<8 x i16>, <4 x float>, i8) nounwind readonly

define <8 x float> @test_x86_vcvtph2ps_256(<8 x i16> %a0) {
; CHECK-LABEL: test_x86_vcvtph2ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vcvtph2ps %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x13,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256(<8 x i16> %a0, <8 x float> zeroinitializer, i8 -1)
  ret <8 x float> %res
}

define <8 x float> @test_x86_vcvtph2ps_256_rrk(<8 x i16> %a0,<8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_x86_vcvtph2ps_256_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtph2ps %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x13,0xc8]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256(<8 x i16> %a0, <8 x float> %a1, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_x86_vcvtph2ps_256_rrkz(<8 x i16> %a0, i8 %mask) {
; CHECK-LABEL: test_x86_vcvtph2ps_256_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtph2ps %xmm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x13,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256(<8 x i16> %a0, <8 x float> zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

declare <8 x float> @llvm.x86.avx512.mask.vcvtph2ps.256(<8 x i16>, <8 x float>, i8) nounwind readonly

define <8 x i16> @test_x86_vcvtps2ph_128(<4 x float> %a0, i8 %mask, <8 x i16> %src) {
; CHECK-LABEL: test_x86_vcvtps2ph_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtps2ph $2, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x1d,0xc1,0x02]
; CHECK-NEXT:    vcvtps2ph $2, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0x89,0x1d,0xc2,0x02]
; CHECK-NEXT:    vcvtps2ph $2, %xmm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x08,0x1d,0xc0,0x02]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0xc5,0xf1,0xfd,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res1 = call <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.128(<4 x float> %a0, i32 2, <8 x i16> zeroinitializer, i8 -1)
  %res2 = call <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.128(<4 x float> %a0, i32 2, <8 x i16> zeroinitializer, i8 %mask)
  %res3 = call <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.128(<4 x float> %a0, i32 2, <8 x i16> %src, i8 %mask)
  %res0 = add <8 x i16> %res1, %res2
  %res = add <8 x i16> %res3, %res0
  ret <8 x i16> %res
}

declare <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.128(<4 x float>, i32, <8 x i16>, i8) nounwind readonly

define <8 x i16> @test_x86_vcvtps2ph_256(<8 x float> %a0, i8 %mask, <8 x i16> %src) {
; CHECK-LABEL: test_x86_vcvtps2ph_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vcvtps2ph $2, %ymm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x1d,0xc1,0x02]
; CHECK-NEXT:    vcvtps2ph $2, %ymm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x1d,0xc2,0x02]
; CHECK-NEXT:    vcvtps2ph $2, %ymm0, %xmm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x1d,0xc0,0x02]
; CHECK-NEXT:    vpaddw %xmm2, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xfd,0xc2]
; CHECK-NEXT:    vpaddw %xmm0, %xmm1, %xmm0 ## encoding: [0xc5,0xf1,0xfd,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res1 = call <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.256(<8 x float> %a0, i32 2, <8 x i16> zeroinitializer, i8 -1)
  %res2 = call <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.256(<8 x float> %a0, i32 2, <8 x i16> zeroinitializer, i8 %mask)
  %res3 = call <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.256(<8 x float> %a0, i32 2, <8 x i16> %src, i8 %mask)
  %res0 = add <8 x i16> %res1, %res2
  %res = add <8 x i16> %res3, %res0
  ret <8 x i16> %res
}

declare <8 x i16> @llvm.x86.avx512.mask.vcvtps2ph.256(<8 x float>, i32, <8 x i16>, i8) nounwind readonly

define <8 x float> @test_rsqrt_ps_256_rr(<8 x float> %a0) {
; CHECK-LABEL: test_rsqrt_ps_256_rr:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vrsqrt14ps %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x4e,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.rsqrt14.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 -1)
  ret <8 x float> %res
}

define <8 x float> @test_rsqrt_ps_256_rrkz(<8 x float> %a0, i8 %mask) {
; CHECK-LABEL: test_rsqrt_ps_256_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrsqrt14ps %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x4e,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.rsqrt14.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_rsqrt_ps_256_rrk(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_rsqrt_ps_256_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrsqrt14ps %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x4e,0xc8]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.rsqrt14.ps.256(<8 x float> %a0, <8 x float> %a1, i8 %mask)
  ret <8 x float> %res
}

define <4 x float> @test_rsqrt_ps_128_rr(<4 x float> %a0) {
; CHECK-LABEL: test_rsqrt_ps_128_rr:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vrsqrt14ps %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x4e,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.rsqrt14.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 -1)
  ret <4 x float> %res
}

define <4 x float> @test_rsqrt_ps_128_rrkz(<4 x float> %a0, i8 %mask) {
; CHECK-LABEL: test_rsqrt_ps_128_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrsqrt14ps %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x4e,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.rsqrt14.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_rsqrt_ps_128_rrk(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_rsqrt_ps_128_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrsqrt14ps %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x4e,0xc8]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.rsqrt14.ps.128(<4 x float> %a0, <4 x float> %a1, i8 %mask)
  ret <4 x float> %res
}

declare <8 x float> @llvm.x86.avx512.rsqrt14.ps.256(<8 x float>, <8 x float>, i8) nounwind readnone
declare <4 x float> @llvm.x86.avx512.rsqrt14.ps.128(<4 x float>, <4 x float>, i8) nounwind readnone

define <8 x float> @test_rcp_ps_256_rr(<8 x float> %a0) {
; CHECK-LABEL: test_rcp_ps_256_rr:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vrcp14ps %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x4c,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.rcp14.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 -1)
  ret <8 x float> %res
}

define <8 x float> @test_rcp_ps_256_rrkz(<8 x float> %a0, i8 %mask) {
; CHECK-LABEL: test_rcp_ps_256_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrcp14ps %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x4c,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.rcp14.ps.256(<8 x float> %a0, <8 x float> zeroinitializer, i8 %mask)
  ret <8 x float> %res
}

define <8 x float> @test_rcp_ps_256_rrk(<8 x float> %a0, <8 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_rcp_ps_256_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrcp14ps %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x4c,0xc8]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.rcp14.ps.256(<8 x float> %a0, <8 x float> %a1, i8 %mask)
  ret <8 x float> %res
}

define <4 x float> @test_rcp_ps_128_rr(<4 x float> %a0) {
; CHECK-LABEL: test_rcp_ps_128_rr:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vrcp14ps %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x4c,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.rcp14.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 -1)
  ret <4 x float> %res
}

define <4 x float> @test_rcp_ps_128_rrkz(<4 x float> %a0, i8 %mask) {
; CHECK-LABEL: test_rcp_ps_128_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrcp14ps %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x4c,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.rcp14.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 %mask)
  ret <4 x float> %res
}

define <4 x float> @test_rcp_ps_128_rrk(<4 x float> %a0, <4 x float> %a1, i8 %mask) {
; CHECK-LABEL: test_rcp_ps_128_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrcp14ps %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x4c,0xc8]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.rcp14.ps.128(<4 x float> %a0, <4 x float> %a1, i8 %mask)
  ret <4 x float> %res
}

declare <8 x float> @llvm.x86.avx512.rcp14.ps.256(<8 x float>, <8 x float>, i8) nounwind readnone
declare <4 x float> @llvm.x86.avx512.rcp14.ps.128(<4 x float>, <4 x float>, i8) nounwind readnone

define <4 x double> @test_rsqrt_pd_256_rr(<4 x double> %a0) {
; CHECK-LABEL: test_rsqrt_pd_256_rr:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vrsqrt14pd %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x4e,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.rsqrt14.pd.256(<4 x double> %a0, <4 x double> zeroinitializer, i8 -1)
  ret <4 x double> %res
}

define <4 x double> @test_rsqrt_pd_256_rrkz(<4 x double> %a0, i8 %mask) {
; CHECK-LABEL: test_rsqrt_pd_256_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrsqrt14pd %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x4e,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.rsqrt14.pd.256(<4 x double> %a0, <4 x double> zeroinitializer, i8 %mask)
  ret <4 x double> %res
}

define <4 x double> @test_rsqrt_pd_256_rrk(<4 x double> %a0, <4 x double> %a1, i8 %mask) {
; CHECK-LABEL: test_rsqrt_pd_256_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrsqrt14pd %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x4e,0xc8]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.rsqrt14.pd.256(<4 x double> %a0, <4 x double> %a1, i8 %mask)
  ret <4 x double> %res
}

define <2 x double> @test_rsqrt_pd_128_rr(<2 x double> %a0) {
; CHECK-LABEL: test_rsqrt_pd_128_rr:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vrsqrt14pd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x4e,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.rsqrt14.pd.128(<2 x double> %a0, <2 x double> zeroinitializer, i8 -1)
  ret <2 x double> %res
}

define <2 x double> @test_rsqrt_pd_128_rrkz(<2 x double> %a0, i8 %mask) {
; CHECK-LABEL: test_rsqrt_pd_128_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrsqrt14pd %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x4e,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.rsqrt14.pd.128(<2 x double> %a0, <2 x double> zeroinitializer, i8 %mask)
  ret <2 x double> %res
}

define <2 x double> @test_rsqrt_pd_128_rrk(<2 x double> %a0, <2 x double> %a1, i8 %mask) {
; CHECK-LABEL: test_rsqrt_pd_128_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrsqrt14pd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x4e,0xc8]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.rsqrt14.pd.128(<2 x double> %a0, <2 x double> %a1, i8 %mask)
  ret <2 x double> %res
}

declare <4 x double> @llvm.x86.avx512.rsqrt14.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone
declare <2 x double> @llvm.x86.avx512.rsqrt14.pd.128(<2 x double>, <2 x double>, i8) nounwind readnone

define <4 x double> @test_rcp_pd_256_rr(<4 x double> %a0) {
; CHECK-LABEL: test_rcp_pd_256_rr:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vrcp14pd %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x4c,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double> %a0, <4 x double> zeroinitializer, i8 -1)
  ret <4 x double> %res
}

define <4 x double> @test_rcp_pd_256_rrkz(<4 x double> %a0, i8 %mask) {
; CHECK-LABEL: test_rcp_pd_256_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrcp14pd %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x4c,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double> %a0, <4 x double> zeroinitializer, i8 %mask)
  ret <4 x double> %res
}

define <4 x double> @test_rcp_pd_256_rrk(<4 x double> %a0, <4 x double> %a1, i8 %mask) {
; CHECK-LABEL: test_rcp_pd_256_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrcp14pd %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x4c,0xc8]
; CHECK-NEXT:    vmovaps %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double> %a0, <4 x double> %a1, i8 %mask)
  ret <4 x double> %res
}

define <2 x double> @test_rcp_pd_128_rr(<2 x double> %a0) {
; CHECK-LABEL: test_rcp_pd_128_rr:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vrcp14pd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x4c,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double> %a0, <2 x double> zeroinitializer, i8 -1)
  ret <2 x double> %res
}

define <2 x double> @test_rcp_pd_128_rrkz(<2 x double> %a0, i8 %mask) {
; CHECK-LABEL: test_rcp_pd_128_rrkz:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrcp14pd %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x4c,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double> %a0, <2 x double> zeroinitializer, i8 %mask)
  ret <2 x double> %res
}

define <2 x double> @test_rcp_pd_128_rrk(<2 x double> %a0, <2 x double> %a1, i8 %mask) {
; CHECK-LABEL: test_rcp_pd_128_rrk:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vrcp14pd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x4c,0xc8]
; CHECK-NEXT:    vmovaps %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double> %a0, <2 x double> %a1, i8 %mask)
  ret <2 x double> %res
}

declare <4 x double> @llvm.x86.avx512.rcp14.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone
declare <2 x double> @llvm.x86.avx512.rcp14.pd.128(<2 x double>, <2 x double>, i8) nounwind readnone

declare <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_broadcastf32x4_256(<4 x float> %x0, <8 x float> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_broadcastf32x4_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshuff32x4 $0, %ymm0, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x23,0xd0,0x00]
; CHECK-NEXT:    ## ymm2 {%k1} {z} = ymm0[0,1,2,3,0,1,2,3]
; CHECK-NEXT:    vshuff32x4 $0, %ymm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x23,0xc8,0x00]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,1,2,3,0,1,2,3]
; CHECK-NEXT:    vshuff32x4 $0, %ymm0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x23,0xc0,0x00]
; CHECK-NEXT:    ## ymm0 = ymm0[0,1,2,3,0,1,2,3]
; CHECK-NEXT:    vaddps %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x58,0xc1]
; CHECK-NEXT:    vaddps %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res1 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 -1)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> %x2, i8 %mask)
  %res3 = call <8 x float> @llvm.x86.avx512.mask.broadcastf32x4.256(<4 x float> %x0, <8 x float> zeroinitializer, i8 %mask)
  %res4 = fadd <8 x float> %res1, %res2
  %res5 = fadd <8 x float> %res3, %res4
  ret <8 x float> %res5
}

declare <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_broadcasti32x4_256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_broadcasti32x4_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    ## kill: %XMM0<def> %XMM0<kill> %YMM0<def>
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vshufi32x4 $0, %ymm0, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x43,0xd0,0x00]
; CHECK-NEXT:    ## ymm2 {%k1} {z} = ymm0[0,1,2,3,0,1,2,3]
; CHECK-NEXT:    vshufi32x4 $0, %ymm0, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x43,0xc8,0x00]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,1,2,3,0,1,2,3]
; CHECK-NEXT:    vshufi32x4 $0, %ymm0, %ymm0, %ymm0 ## encoding: [0x62,0xf3,0x7d,0x28,0x43,0xc0,0x00]
; CHECK-NEXT:    ## ymm0 = ymm0[0,1,2,3,0,1,2,3]
; CHECK-NEXT:    vpaddd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfe,0xc1]
; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 -1)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> %x2, i8 %mask)
  %res3 = call <8 x i32> @llvm.x86.avx512.mask.broadcasti32x4.256(<4 x i32> %x0, <8 x i32> zeroinitializer, i8 %mask)
  %res4 = add <8 x i32> %res1, %res2
  %res5 = add <8 x i32> %res3, %res4
  ret <8 x i32> %res5
}

declare <2 x i64> @llvm.x86.avx512.mask.psrl.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_psrl_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrlq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0xd3,0xd1]
; CHECK-NEXT:    vpsrlq %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0xd3,0xd9]
; CHECK-NEXT:    vpsrlq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xd3,0xc1]
; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc0]
; CHECK-NEXT:    vpaddq %xmm3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xd4,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.psrl.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psrl.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psrl.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.psrl.q.256(<4 x i64>, <2 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psrl_q_256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrlq %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xd3,0xd1]
; CHECK-NEXT:    vpsrlq %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0xd3,0xd9]
; CHECK-NEXT:    vpsrlq %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xd3,0xc1]
; 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.psrl.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psrl.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psrl.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.psrl.qi.128(<2 x i64>, i32, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_psrl_qi_128(<2 x i64> %x0, i32 %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_qi_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsrlq $255, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x09,0x73,0xd0,0xff]
; CHECK-NEXT:    vpsrlq $255, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0x89,0x73,0xd0,0xff]
; CHECK-NEXT:    vpsrlq $255, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x73,0xd0,0xff]
; CHECK-NEXT:    vpaddq %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0xd4,0xc0]
; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.psrl.qi.128(<2 x i64> %x0, i32 255, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psrl.qi.128(<2 x i64> %x0, i32 255, <2 x i64> %x2, i8 -1)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psrl.qi.128(<2 x i64> %x0, i32 255, <2 x i64> zeroinitializer, i8 %x3)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res2, %res3
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.psrl.qi.256(<4 x i64>, i32, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psrl_qi_256(<4 x i64> %x0, i32 %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_qi_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsrlq $255, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x29,0x73,0xd0,0xff]
; CHECK-NEXT:    vpsrlq $255, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0xa9,0x73,0xd0,0xff]
; CHECK-NEXT:    vpsrlq $255, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x73,0xd0,0xff]
; CHECK-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.psrl.qi.256(<4 x i64> %x0, i32 255, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psrl.qi.256(<4 x i64> %x0, i32 255, <4 x i64> %x2, i8 -1)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psrl.qi.256(<4 x i64> %x0, i32 255, <4 x i64> zeroinitializer, i8 %x3)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res2, %res3
  ret <4 x i64> %res4
}
declare <4 x i32> @llvm.x86.avx512.mask.psrl.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)
define <4 x i32>@test_int_x86_avx512_mask_psrl_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrld %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xd2,0xd1]
; CHECK-NEXT:    vpsrld %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xd2,0xd9]
; CHECK-NEXT:    vpsrld %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xd2,0xc1]
; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xc0]
; CHECK-NEXT:    vpaddd %xmm3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.psrl.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psrl.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psrl.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psrl.d.256(<8 x i32>, <4 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psrl_d_256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrld %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xd2,0xd1]
; CHECK-NEXT:    vpsrld %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xd2,0xd9]
; CHECK-NEXT:    vpsrld %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xd2,0xc1]
; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xc0]
; CHECK-NEXT:    vpaddd %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x65,0x28,0xfe,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.psrl.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psrl.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psrl.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res2, %res3
  ret <8 x i32> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.psrl.di.128(<4 x i32>, i32, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_psrl_di_128(<4 x i32> %x0, i32 %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_di_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsrld $255, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x09,0x72,0xd0,0xff]
; CHECK-NEXT:    vpsrld $255, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0x89,0x72,0xd0,0xff]
; CHECK-NEXT:    vpsrld $255, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x72,0xd0,0xff]
; CHECK-NEXT:    vpaddd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfe,0xc0]
; 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.psrl.di.128(<4 x i32> %x0, i32 255, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psrl.di.128(<4 x i32> %x0, i32 255, <4 x i32> %x2, i8 -1)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psrl.di.128(<4 x i32> %x0, i32 255, <4 x i32> zeroinitializer, i8 %x3)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res2, %res3
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psrl.di.256(<8 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psrl_di_256(<8 x i32> %x0, i32 %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_di_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsrld $255, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x29,0x72,0xd0,0xff]
; CHECK-NEXT:    vpsrld $255, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xa9,0x72,0xd0,0xff]
; CHECK-NEXT:    vpsrld $255, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x72,0xd0,0xff]
; CHECK-NEXT:    vpaddd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfe,0xc0]
; 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.psrl.di.256(<8 x i32> %x0, i32 255, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psrl.di.256(<8 x i32> %x0, i32 255, <8 x i32> %x2, i8 -1)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psrl.di.256(<8 x i32> %x0, i32 255, <8 x i32> zeroinitializer, i8 %x3)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res2, %res3
  ret <8 x i32> %res4
}

declare <16 x i32> @llvm.x86.avx512.mask.psrl.di.512(<16 x i32>, i32, <16 x i32>, i16)

define <16 x i32>@test_int_x86_avx512_mask_psrl_di_512(<16 x i32> %x0, i32 %x1, <16 x i32> %x2, i16 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrl_di_512:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsrld $255, %zmm0, %zmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x49,0x72,0xd0,0xff]
; CHECK-NEXT:    vpsrld $255, %zmm0, %zmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xc9,0x72,0xd0,0xff]
; CHECK-NEXT:    vpsrld $255, %zmm0, %zmm0 ## encoding: [0x62,0xf1,0x7d,0x48,0x72,0xd0,0xff]
; CHECK-NEXT:    vpaddd %zmm0, %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x75,0x48,0xfe,0xc0]
; CHECK-NEXT:    vpaddd %zmm0, %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x6d,0x48,0xfe,0xc0]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <16 x i32> @llvm.x86.avx512.mask.psrl.di.512(<16 x i32> %x0, i32 255, <16 x i32> %x2, i16 %x3)
  %res1 = call <16 x i32> @llvm.x86.avx512.mask.psrl.di.512(<16 x i32> %x0, i32 255, <16 x i32> %x2, i16 -1)
  %res2 = call <16 x i32> @llvm.x86.avx512.mask.psrl.di.512(<16 x i32> %x0, i32 255, <16 x i32> zeroinitializer, i16 %x3)
  %res3 = add <16 x i32> %res, %res1
  %res4 = add <16 x i32> %res2, %res3
  ret <16 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.psrlv2.di(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_psrlv2_di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrlv2_di:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrlvq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x45,0xd1]
; CHECK-NEXT:    vpsrlvq %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x45,0xd9]
; CHECK-NEXT:    vpsrlvq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x45,0xc1]
; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xcb]
; 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.psrlv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psrlv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psrlv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.psrlv4.di(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psrlv4_di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrlv4_di:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x45,0xd1]
; CHECK-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x45,0xd9]
; CHECK-NEXT:    vpsrlvq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x45,0xc1]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xcb]
; 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.psrlv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psrlv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psrlv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.psrlv4.si(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_psrlv4_si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrlv4_si:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrlvd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x45,0xd1]
; CHECK-NEXT:    vpsrlvd %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x45,0xd9]
; CHECK-NEXT:    vpsrlvd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x45,0xc1]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xcb]
; 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.psrlv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psrlv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psrlv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psrlv8.si(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psrlv8_si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrlv8_si:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x45,0xd1]
; CHECK-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x45,0xd9]
; CHECK-NEXT:    vpsrlvd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x45,0xc1]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xcb]
; 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.psrlv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psrlv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psrlv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.psra.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_psra_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrad %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xe2,0xd1]
; CHECK-NEXT:    vpsrad %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xe2,0xd9]
; CHECK-NEXT:    vpsrad %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xe2,0xc1]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xcb]
; 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.psra.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psra.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psra.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psra.d.256(<8 x i32>, <4 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psra_d_256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsrad %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xe2,0xd1]
; CHECK-NEXT:    vpsrad %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xe2,0xd9]
; CHECK-NEXT:    vpsrad %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xe2,0xc1]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xcb]
; 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.psra.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psra.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psra.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.psra.di.128(<4 x i32>, i32, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_psra_di_128(<4 x i32> %x0, i32 %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_di_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsrad $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x09,0x72,0xe0,0x03]
; CHECK-NEXT:    vpsrad $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0x89,0x72,0xe0,0x03]
; CHECK-NEXT:    vpsrad $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x72,0xe0,0x03]
; 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.psra.di.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psra.di.128(<4 x i32> %x0, i32 3, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psra.di.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psra.di.256(<8 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psra_di_256(<8 x i32> %x0, i32 %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_di_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsrad $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x29,0x72,0xe0,0x03]
; CHECK-NEXT:    vpsrad $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xa9,0x72,0xe0,0x03]
; CHECK-NEXT:    vpsrad $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x72,0xe0,0x03]
; 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.psra.di.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psra.di.256(<8 x i32> %x0, i32 3, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psra.di.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.psra.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_psra_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsraq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0xe2,0xd1]
; CHECK-NEXT:    vpsraq %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0xe2,0xd9]
; CHECK-NEXT:    vpsraq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xe2,0xc1]
; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xcb]
; 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.psra.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psra.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psra.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.psra.q.256(<4 x i64>, <2 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psra_q_256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsraq %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xe2,0xd1]
; CHECK-NEXT:    vpsraq %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0xe2,0xd9]
; CHECK-NEXT:    vpsraq %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xe2,0xc1]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xcb]
; 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.psra.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psra.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psra.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.psra.qi.128(<2 x i64>, i32, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_psra_qi_128(<2 x i64> %x0, i32 %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_qi_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsraq $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x09,0x72,0xe0,0x03]
; CHECK-NEXT:    vpsraq $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0x89,0x72,0xe0,0x03]
; CHECK-NEXT:    vpsraq $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x72,0xe0,0x03]
; 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.psra.qi.128(<2 x i64> %x0, i32 3, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psra.qi.128(<2 x i64> %x0, i32 3, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psra.qi.128(<2 x i64> %x0, i32 3, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.psra.qi.256(<4 x i64>, i32, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psra_qi_256(<4 x i64> %x0, i32 %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psra_qi_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsraq $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x29,0x72,0xe0,0x03]
; CHECK-NEXT:    vpsraq $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0xa9,0x72,0xe0,0x03]
; CHECK-NEXT:    vpsraq $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x72,0xe0,0x03]
; 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.psra.qi.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psra.qi.256(<4 x i64> %x0, i32 3, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psra.qi.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}


declare <4 x i32> @llvm.x86.avx512.mask.psll.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_psll_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpslld %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xf2,0xd1]
; CHECK-NEXT:    vpslld %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xf2,0xd9]
; CHECK-NEXT:    vpslld %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xf2,0xc1]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xcb]
; 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.psll.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psll.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psll.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psll.d.256(<8 x i32>, <4 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psll_d_256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpslld %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xf2,0xd1]
; CHECK-NEXT:    vpslld %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xf2,0xd9]
; CHECK-NEXT:    vpslld %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xf2,0xc1]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xcb]
; 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.psll.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psll.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psll.d.256(<8 x i32> %x0, <4 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.psll.di.128(<4 x i32>, i32, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_psll_di_128(<4 x i32> %x0, i32 %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_di_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpslld $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x09,0x72,0xf0,0x03]
; CHECK-NEXT:    vpslld $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0x89,0x72,0xf0,0x03]
; CHECK-NEXT:    vpslld $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x72,0xf0,0x03]
; 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.psll.di.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psll.di.128(<4 x i32> %x0, i32 3, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psll.di.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psll.di.256(<8 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psll_di_256(<8 x i32> %x0, i32 %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_di_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpslld $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x29,0x72,0xf0,0x03]
; CHECK-NEXT:    vpslld $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xa9,0x72,0xf0,0x03]
; CHECK-NEXT:    vpslld $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x72,0xf0,0x03]
; 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.psll.di.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psll.di.256(<8 x i32> %x0, i32 3, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psll.di.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.psll.q.256(<4 x i64>, <2 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psll_q_256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsllq %xmm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xf3,0xd1]
; CHECK-NEXT:    vpsllq %xmm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0xf3,0xd9]
; CHECK-NEXT:    vpsllq %xmm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xf3,0xc1]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xcb]
; 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.psll.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psll.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psll.q.256(<4 x i64> %x0, <2 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.psll.qi.128(<2 x i64>, i32, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_psll_qi_128(<2 x i64> %x0, i32 %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_qi_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsllq $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x09,0x73,0xf0,0x03]
; CHECK-NEXT:    vpsllq $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0x89,0x73,0xf0,0x03]
; CHECK-NEXT:    vpsllq $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x73,0xf0,0x03]
; 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.psll.qi.128(<2 x i64> %x0, i32 3, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psll.qi.128(<2 x i64> %x0, i32 3, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psll.qi.128(<2 x i64> %x0, i32 3, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.psll.qi.256(<4 x i64>, i32, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psll_qi_256(<4 x i64> %x0, i32 %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psll_qi_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpsllq $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x29,0x73,0xf0,0x03]
; CHECK-NEXT:    vpsllq $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0xa9,0x73,0xf0,0x03]
; CHECK-NEXT:    vpsllq $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x73,0xf0,0x03]
; 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.psll.qi.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psll.qi.256(<4 x i64> %x0, i32 3, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psll.qi.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
 ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.psrav4.si(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_psrav4_si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrav4_si:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsravd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x46,0xd1]
; CHECK-NEXT:    vpsravd %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x46,0xd9]
; CHECK-NEXT:    vpsravd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x46,0xc1]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xcb]
; 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.psrav4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psrav4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psrav4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psrav8.si(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psrav8_si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrav8_si:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsravd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x46,0xd1]
; CHECK-NEXT:    vpsravd %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x46,0xd9]
; CHECK-NEXT:    vpsravd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x46,0xc1]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xcb]
; 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.psrav8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psrav8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psrav8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

define <8 x i32>@test_int_x86_avx512_mask_psrav8_si_const() {
; CHECK-LABEL: test_int_x86_avx512_mask_psrav8_si_const:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqa32 {{.*#+}} ymm0 = [2,9,4294967284,23,4294967270,37,4294967256,51]
; CHECK-NEXT:    ## encoding: [0x62,0xf1,0x7d,0x28,0x6f,0x05,A,A,A,A]
; CHECK-NEXT:    ## fixup A - offset: 6, value: LCPI371_0-4, kind: reloc_riprel_4byte
; CHECK-NEXT:    vpsravd {{.*}}(%rip), %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x46,0x05,A,A,A,A]
; CHECK-NEXT:    ## fixup A - offset: 6, value: LCPI371_1-4, kind: reloc_riprel_4byte
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.psrav8.si(<8 x i32> <i32 2, i32 9, i32 -12, i32 23, i32 -26, i32 37, i32 -40, i32 51>, <8 x i32> <i32 1, i32 18, i32 35, i32 52, i32 69, i32 15, i32 32, i32 49>, <8 x i32> zeroinitializer, i8 -1)
  ret <8 x i32> %res
}

declare <2 x i64> @llvm.x86.avx512.mask.psrav.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_psrav_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrav_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsravq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x46,0xd1]
; CHECK-NEXT:    vpsravq %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x46,0xd9]
; CHECK-NEXT:    vpsravq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x46,0xc1]
; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xcb]
; 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.psrav.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psrav.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psrav.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

define <2 x i64>@test_int_x86_avx512_mask_psrav_q_128_const(i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrav_q_128_const:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqa64 {{.*#+}} xmm0 = [2,18446744073709551607]
; CHECK-NEXT:    ## encoding: [0x62,0xf1,0xfd,0x08,0x6f,0x05,A,A,A,A]
; CHECK-NEXT:    ## fixup A - offset: 6, value: LCPI373_0-4, kind: reloc_riprel_4byte
; CHECK-NEXT:    vpsravq {{.*}}(%rip), %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x46,0x05,A,A,A,A]
; CHECK-NEXT:    ## fixup A - offset: 6, value: LCPI373_1-4, kind: reloc_riprel_4byte
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.psrav.q.128(<2 x i64> <i64 2, i64 -9>, <2 x i64> <i64 1, i64 90>, <2 x i64> zeroinitializer, i8 -1)
  ret <2 x i64> %res
}

declare <4 x i64> @llvm.x86.avx512.mask.psrav.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psrav_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psrav_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsravq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x46,0xd1]
; CHECK-NEXT:    vpsravq %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x46,0xd9]
; CHECK-NEXT:    vpsravq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x46,0xc1]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xcb]
; 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.psrav.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psrav.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psrav.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.psllv2.di(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_psllv2_di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psllv2_di:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsllvq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x47,0xd1]
; CHECK-NEXT:    vpsllvq %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x47,0xd9]
; CHECK-NEXT:    vpsllvq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x47,0xc1]
; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xcb]
; 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.psllv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.psllv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.psllv2.di(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.psllv4.di(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_psllv4_di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psllv4_di:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsllvq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x47,0xd1]
; CHECK-NEXT:    vpsllvq %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x47,0xd9]
; CHECK-NEXT:    vpsllvq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x47,0xc1]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xcb]
; 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.psllv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.psllv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.psllv4.di(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.psllv4.si(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_psllv4_si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psllv4_si:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsllvd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x47,0xd1]
; CHECK-NEXT:    vpsllvd %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x47,0xd9]
; CHECK-NEXT:    vpsllvd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x47,0xc1]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xcb]
; 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.psllv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.psllv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.psllv4.si(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.psllv8.si(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_psllv8_si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_psllv8_si:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpsllvd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x47,0xd1]
; CHECK-NEXT:    vpsllvd %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x47,0xd9]
; CHECK-NEXT:    vpsllvd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x47,0xc1]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xcb]
; 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.psllv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.psllv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.psllv8.si(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.prorv.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_prorv_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prorv_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vprorvd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x14,0xd1]
; CHECK-NEXT:    vprorvd %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x14,0xd9]
; CHECK-NEXT:    vprorvd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x14,0xc1]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xcb]
; 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.prorv.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.prorv.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.prorv.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.prorv.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_prorv_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prorv_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vprorvd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x14,0xd1]
; CHECK-NEXT:    vprorvd %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x14,0xd9]
; CHECK-NEXT:    vprorvd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x14,0xc1]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xcb]
; 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.prorv.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.prorv.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.prorv.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.prorv.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_prorv_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prorv_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vprorvq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x14,0xd1]
; CHECK-NEXT:    vprorvq %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x14,0xd9]
; CHECK-NEXT:    vprorvq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x14,0xc1]
; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xcb]
; 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.prorv.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.prorv.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.prorv.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.prorv.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_prorv_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prorv_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vprorvq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x14,0xd1]
; CHECK-NEXT:    vprorvq %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x14,0xd9]
; CHECK-NEXT:    vprorvq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x14,0xc1]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xcb]
; 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.prorv.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.prorv.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.prorv.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.prol.d.128(<4 x i32>, i32, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_prol_d_128(<4 x i32> %x0, i32 %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prol_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vprold $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x09,0x72,0xc8,0x03]
; CHECK-NEXT:    vprold $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0x89,0x72,0xc8,0x03]
; CHECK-NEXT:    vprold $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x72,0xc8,0x03]
; 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.prol.d.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.prol.d.128(<4 x i32> %x0, i32 3, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.prol.d.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.prol.d.256(<8 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_prol_d_256(<8 x i32> %x0, i32 %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prol_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vprold $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x29,0x72,0xc8,0x03]
; CHECK-NEXT:    vprold $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xa9,0x72,0xc8,0x03]
; CHECK-NEXT:    vprold $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x72,0xc8,0x03]
; 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.prol.d.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.prol.d.256(<8 x i32> %x0, i32 3, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.prol.d.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.prol.q.128(<2 x i64>, i32, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_prol_q_128(<2 x i64> %x0, i32 %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prol_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vprolq $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x09,0x72,0xc8,0x03]
; CHECK-NEXT:    vprolq $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0x89,0x72,0xc8,0x03]
; CHECK-NEXT:    vprolq $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x72,0xc8,0x03]
; 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.prol.q.128(<2 x i64> %x0, i32 3, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.prol.q.128(<2 x i64> %x0, i32 3, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.prol.q.128(<2 x i64> %x0, i32 3, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.prol.q.256(<4 x i64>, i32, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_prol_q_256(<4 x i64> %x0, i32 %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prol_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vprolq $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x29,0x72,0xc8,0x03]
; CHECK-NEXT:    vprolq $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0xa9,0x72,0xc8,0x03]
; CHECK-NEXT:    vprolq $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x72,0xc8,0x03]
; 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.prol.q.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.prol.q.256(<4 x i64> %x0, i32 3, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.prol.q.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.prolv.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_prolv_d_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prolv_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vprolvd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x15,0xd1]
; CHECK-NEXT:    vprolvd %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x15,0xd9]
; CHECK-NEXT:    vprolvd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x15,0xc1]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xcb]
; 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.prolv.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.prolv.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.prolv.d.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.prolv.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_prolv_d_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prolv_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vprolvd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x15,0xd1]
; CHECK-NEXT:    vprolvd %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x15,0xd9]
; CHECK-NEXT:    vprolvd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x15,0xc1]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xcb]
; 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.prolv.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.prolv.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.prolv.d.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.prolv.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_prolv_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prolv_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vprolvq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x15,0xd1]
; CHECK-NEXT:    vprolvq %xmm1, %xmm0, %xmm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x15,0xd9]
; CHECK-NEXT:    vprolvq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf2,0xfd,0x08,0x15,0xc1]
; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xcb]
; 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.prolv.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.prolv.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.prolv.q.128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.prolv.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_prolv_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_prolv_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vprolvq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x15,0xd1]
; CHECK-NEXT:    vprolvq %ymm1, %ymm0, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x15,0xd9]
; CHECK-NEXT:    vprolvq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf2,0xfd,0x28,0x15,0xc1]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xcb]
; 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.prolv.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.prolv.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.prolv.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.pror.d.128(<4 x i32>, i32, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pror_d_128(<4 x i32> %x0, i32 %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pror_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vprord $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x09,0x72,0xc0,0x03]
; CHECK-NEXT:    vprord $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0x89,0x72,0xc0,0x03]
; CHECK-NEXT:    vprord $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x72,0xc0,0x03]
; 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.pror.d.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pror.d.128(<4 x i32> %x0, i32 3, <4 x i32> zeroinitializer, i8 %x3)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.pror.d.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.pror.d.256(<8 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pror_d_256(<8 x i32> %x0, i32 %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pror_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vprord $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x75,0x29,0x72,0xc0,0x03]
; CHECK-NEXT:    vprord $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0x6d,0xa9,0x72,0xc0,0x03]
; CHECK-NEXT:    vprord $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x72,0xc0,0x03]
; 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.pror.d.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pror.d.256(<8 x i32> %x0, i32 3, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.pror.d.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.pror.q.128(<2 x i64>, i32, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pror_q_128(<2 x i64> %x0, i32 %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pror_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vprorq $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x09,0x72,0xc0,0x03]
; CHECK-NEXT:    vprorq $3, %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0x89,0x72,0xc0,0x03]
; CHECK-NEXT:    vprorq $3, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x72,0xc0,0x03]
; 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.pror.q.128(<2 x i64> %x0, i32 3, <2 x i64> %x2, i8 %x3)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pror.q.128(<2 x i64> %x0, i32 3, <2 x i64> zeroinitializer, i8 %x3)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.pror.q.128(<2 x i64> %x0, i32 3, <2 x i64> %x2, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.pror.q.256(<4 x i64>, i32, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pror_q_256(<4 x i64> %x0, i32 %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pror_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vprorq $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xf5,0x29,0x72,0xc0,0x03]
; CHECK-NEXT:    vprorq $3, %ymm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf1,0xed,0xa9,0x72,0xc0,0x03]
; CHECK-NEXT:    vprorq $3, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x72,0xc0,0x03]
; 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.pror.q.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pror.q.256(<4 x i64> %x0, i32 3, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.pror.q.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.pmovzxb.d.128(<16 x i8>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmovzxb_d_128(<16 x i8> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxb_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxbd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x31,0xc8]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
; CHECK-NEXT:    vpmovzxbd %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x31,0xd0]
; CHECK-NEXT:    ## xmm2 {%k1} {z} = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
; CHECK-NEXT:    vpmovzxbd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x31,0xc0]
; CHECK-NEXT:    ## xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
; 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.pmovzxb.d.128(<16 x i8> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmovzxb.d.128(<16 x i8> %x0, <4 x i32> zeroinitializer, i8 %x2)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmovzxb.d.128(<16 x i8> %x0, <4 x i32> %x1, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.pmovzxb.d.256(<16 x i8>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pmovzxb_d_256(<16 x i8> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxb_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxbd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x31,0xc8]
; CHECK-NEXT:    ## ymm1 {%k1} = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero
; CHECK-NEXT:    vpmovzxbd %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x31,0xd0]
; CHECK-NEXT:    ## ymm2 {%k1} {z} = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero
; CHECK-NEXT:    vpmovzxbd %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x31,0xc0]
; CHECK-NEXT:    ## ymm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[5],zero,zero,zero,xmm0[6],zero,zero,zero,xmm0[7],zero,zero,zero
; 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.pmovzxb.d.256(<16 x i8> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmovzxb.d.256(<16 x i8> %x0, <8 x i32> zeroinitializer, i8 %x2)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.pmovzxb.d.256(<16 x i8> %x0, <8 x i32> %x1, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.pmovzxb.q.128(<16 x i8>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pmovzxb_q_128(<16 x i8> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxb_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxbq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x32,0xc8]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT:    vpmovzxbq %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x32,0xd0]
; CHECK-NEXT:    ## xmm2 {%k1} {z} = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT:    vpmovzxbq %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x32,0xc0]
; CHECK-NEXT:    ## xmm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero
; 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.pmovzxb.q.128(<16 x i8> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmovzxb.q.128(<16 x i8> %x0, <2 x i64> zeroinitializer, i8 %x2)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.pmovzxb.q.128(<16 x i8> %x0, <2 x i64> %x1, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.pmovzxb.q.256(<16 x i8>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pmovzxb_q_256(<16 x i8> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxb_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxbq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x32,0xc8]
; CHECK-NEXT:    ## ymm1 {%k1} = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero,xmm0[2],zero,zero,zero,zero,zero,zero,zero,xmm0[3],zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT:    vpmovzxbq %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x32,0xd0]
; CHECK-NEXT:    ## ymm2 {%k1} {z} = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero,xmm0[2],zero,zero,zero,zero,zero,zero,zero,xmm0[3],zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT:    vpmovzxbq %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x32,0xc0]
; CHECK-NEXT:    ## ymm0 = xmm0[0],zero,zero,zero,zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,zero,zero,zero,xmm0[2],zero,zero,zero,zero,zero,zero,zero,xmm0[3],zero,zero,zero,zero,zero,zero,zero
; 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.pmovzxb.q.256(<16 x i8> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmovzxb.q.256(<16 x i8> %x0, <4 x i64> zeroinitializer, i8 %x2)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.pmovzxb.q.256(<16 x i8> %x0, <4 x i64> %x1, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.pmovzxd.q.128(<4 x i32>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pmovzxd_q_128(<4 x i32> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxd_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxdq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x35,0xc8]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[0],zero,xmm0[1],zero
; CHECK-NEXT:    vpmovzxdq %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x35,0xd0]
; CHECK-NEXT:    ## xmm2 {%k1} {z} = xmm0[0],zero,xmm0[1],zero
; CHECK-NEXT:    vpmovzxdq %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x35,0xc0]
; CHECK-NEXT:    ## xmm0 = xmm0[0],zero,xmm0[1],zero
; 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.pmovzxd.q.128(<4 x i32> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmovzxd.q.128(<4 x i32> %x0, <2 x i64> zeroinitializer, i8 %x2)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.pmovzxd.q.128(<4 x i32> %x0, <2 x i64> %x1, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.pmovzxd.q.256(<4 x i32>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pmovzxd_q_256(<4 x i32> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxd_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxdq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x35,0xc8]
; CHECK-NEXT:    ## ymm1 {%k1} = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; CHECK-NEXT:    vpmovzxdq %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x35,0xd0]
; CHECK-NEXT:    ## ymm2 {%k1} {z} = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; CHECK-NEXT:    vpmovzxdq %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x35,0xc0]
; CHECK-NEXT:    ## ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; 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.pmovzxd.q.256(<4 x i32> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmovzxd.q.256(<4 x i32> %x0, <4 x i64> zeroinitializer, i8 %x2)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.pmovzxd.q.256(<4 x i32> %x0, <4 x i64> %x1, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.pmovzxw.d.128(<8 x i16>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmovzxw_d_128(<8 x i16> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxw_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxwd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x33,0xc8]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; CHECK-NEXT:    vpmovzxwd %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x33,0xd0]
; CHECK-NEXT:    ## xmm2 {%k1} {z} = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; CHECK-NEXT:    vpmovzxwd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x33,0xc0]
; CHECK-NEXT:    ## xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; 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.pmovzxw.d.128(<8 x i16> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmovzxw.d.128(<8 x i16> %x0, <4 x i32> zeroinitializer, i8 %x2)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmovzxw.d.128(<8 x i16> %x0, <4 x i32> %x1, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.pmovzxw.d.256(<8 x i16>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pmovzxw_d_256(<8 x i16> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxw_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxwd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x33,0xc8]
; CHECK-NEXT:    ## ymm1 {%k1} = 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:    vpmovzxwd %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x33,0xd0]
; CHECK-NEXT:    ## ymm2 {%k1} {z} = 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:    vpmovzxwd %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x33,0xc0]
; 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
; 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.pmovzxw.d.256(<8 x i16> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmovzxw.d.256(<8 x i16> %x0, <8 x i32> zeroinitializer, i8 %x2)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.pmovzxw.d.256(<8 x i16> %x0, <8 x i32> %x1, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.pmovzxw.q.128(<8 x i16>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pmovzxw_q_128(<8 x i16> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxw_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxwq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x34,0xc8]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero
; CHECK-NEXT:    vpmovzxwq %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x34,0xd0]
; CHECK-NEXT:    ## xmm2 {%k1} {z} = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero
; CHECK-NEXT:    vpmovzxwq %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x34,0xc0]
; CHECK-NEXT:    ## xmm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero
; 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.pmovzxw.q.128(<8 x i16> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmovzxw.q.128(<8 x i16> %x0, <2 x i64> zeroinitializer, i8 %x2)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.pmovzxw.q.128(<8 x i16> %x0, <2 x i64> %x1, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.pmovzxw.q.256(<8 x i16>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pmovzxw_q_256(<8 x i16> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovzxw_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovzxwq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x34,0xc8]
; CHECK-NEXT:    ## ymm1 {%k1} = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
; CHECK-NEXT:    vpmovzxwq %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x34,0xd0]
; CHECK-NEXT:    ## ymm2 {%k1} {z} = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
; CHECK-NEXT:    vpmovzxwq %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x34,0xc0]
; CHECK-NEXT:    ## ymm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
; 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.pmovzxw.q.256(<8 x i16> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmovzxw.q.256(<8 x i16> %x0, <4 x i64> zeroinitializer, i8 %x2)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.pmovzxw.q.256(<8 x i16> %x0, <4 x i64> %x1, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.pmovsxb.d.128(<16 x i8>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmovsxb_d_128(<16 x i8> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxb_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsxbd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x21,0xc8]
; CHECK-NEXT:    vpmovsxbd %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x21,0xd0]
; CHECK-NEXT:    vpmovsxbd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x21,0xc0]
; 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.pmovsxb.d.128(<16 x i8> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmovsxb.d.128(<16 x i8> %x0, <4 x i32> zeroinitializer, i8 %x2)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmovsxb.d.128(<16 x i8> %x0, <4 x i32> %x1, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.pmovsxb.d.256(<16 x i8>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pmovsxb_d_256(<16 x i8> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxb_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsxbd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x21,0xc8]
; CHECK-NEXT:    vpmovsxbd %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x21,0xd0]
; CHECK-NEXT:    vpmovsxbd %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x21,0xc0]
; 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.pmovsxb.d.256(<16 x i8> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmovsxb.d.256(<16 x i8> %x0, <8 x i32> zeroinitializer, i8 %x2)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.pmovsxb.d.256(<16 x i8> %x0, <8 x i32> %x1, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.pmovsxb.q.128(<16 x i8>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pmovsxb_q_128(<16 x i8> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxb_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsxbq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x22,0xc8]
; CHECK-NEXT:    vpmovsxbq %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x22,0xd0]
; CHECK-NEXT:    vpmovsxbq %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x22,0xc0]
; 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.pmovsxb.q.128(<16 x i8> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmovsxb.q.128(<16 x i8> %x0, <2 x i64> zeroinitializer, i8 %x2)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.pmovsxb.q.128(<16 x i8> %x0, <2 x i64> %x1, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.pmovsxb.q.256(<16 x i8>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pmovsxb_q_256(<16 x i8> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxb_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsxbq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x22,0xc8]
; CHECK-NEXT:    vpmovsxbq %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x22,0xd0]
; CHECK-NEXT:    vpmovsxbq %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x22,0xc0]
; 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.pmovsxb.q.256(<16 x i8> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmovsxb.q.256(<16 x i8> %x0, <4 x i64> zeroinitializer, i8 %x2)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.pmovsxb.q.256(<16 x i8> %x0, <4 x i64> %x1, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.pmovsxw.d.128(<8 x i16>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pmovsxw_d_128(<8 x i16> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxw_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsxwd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x23,0xc8]
; CHECK-NEXT:    vpmovsxwd %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x23,0xd0]
; CHECK-NEXT:    vpmovsxwd %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x23,0xc0]
; 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.pmovsxw.d.128(<8 x i16> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pmovsxw.d.128(<8 x i16> %x0, <4 x i32> zeroinitializer, i8 %x2)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.pmovsxw.d.128(<8 x i16> %x0, <4 x i32> %x1, i8 -1)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res3, %res2
  ret <4 x i32> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.pmovsxw.d.256(<8 x i16>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pmovsxw_d_256(<8 x i16> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxw_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsxwd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x23,0xc8]
; CHECK-NEXT:    vpmovsxwd %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x23,0xd0]
; CHECK-NEXT:    vpmovsxwd %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x23,0xc0]
; 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.pmovsxw.d.256(<8 x i16> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pmovsxw.d.256(<8 x i16> %x0, <8 x i32> zeroinitializer, i8 %x2)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.pmovsxw.d.256(<8 x i16> %x0, <8 x i32> %x1, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.pmovsxw.q.128(<8 x i16>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pmovsxw_q_128(<8 x i16> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxw_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsxwq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x24,0xc8]
; CHECK-NEXT:    vpmovsxwq %xmm0, %xmm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x24,0xd0]
; CHECK-NEXT:    vpmovsxwq %xmm0, %xmm0 ## encoding: [0x62,0xf2,0x7d,0x08,0x24,0xc0]
; 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.pmovsxw.q.128(<8 x i16> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pmovsxw.q.128(<8 x i16> %x0, <2 x i64> zeroinitializer, i8 %x2)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.pmovsxw.q.128(<8 x i16> %x0, <2 x i64> %x1, i8 -1)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res3, %res2
  ret <2 x i64> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.pmovsxw.q.256(<8 x i16>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pmovsxw_q_256(<8 x i16> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_pmovsxw_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpmovsxwq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x24,0xc8]
; CHECK-NEXT:    vpmovsxwq %xmm0, %ymm2 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x24,0xd0]
; CHECK-NEXT:    vpmovsxwq %xmm0, %ymm0 ## encoding: [0x62,0xf2,0x7d,0x28,0x24,0xc0]
; 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.pmovsxw.q.256(<8 x i16> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pmovsxw.q.256(<8 x i16> %x0, <4 x i64> zeroinitializer, i8 %x2)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.pmovsxw.q.256(<8 x i16> %x0, <4 x i64> %x1, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <4 x double> @llvm.x86.avx512.mask.permvar.df.256(<4 x double>, <4 x i64>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_permvar_df_256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_permvar_df_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermpd %ymm0, %ymm1, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0x16,0xd0]
; CHECK-NEXT:    vpermpd %ymm0, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xa9,0x16,0xd8]
; CHECK-NEXT:    vpermpd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0x16,0xc0]
; CHECK-NEXT:    vaddpd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0x58,0xcb]
; 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.permvar.df.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 %x3)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.permvar.df.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> zeroinitializer, i8 %x3)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.permvar.df.256(<4 x double> %x0, <4 x i64> %x1, <4 x double> %x2, i8 -1)
  %res3 = fadd <4 x double> %res, %res1
  %res4 = fadd <4 x double> %res3, %res2
  ret <4 x double> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.permvar.di.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_permvar_di_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_permvar_di_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermq %ymm0, %ymm1, %ymm2 {%k1} ## encoding: [0x62,0xf2,0xf5,0x29,0x36,0xd0]
; CHECK-NEXT:    vpermq %ymm0, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0xf5,0xa9,0x36,0xd8]
; CHECK-NEXT:    vpermq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0xf5,0x28,0x36,0xc0]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xcb]
; 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.permvar.di.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.permvar.di.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.permvar.di.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res3, %res2
  ret <4 x i64> %res4
}

declare <8 x float> @llvm.x86.avx512.mask.permvar.sf.256(<8 x float>, <8 x i32>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_permvar_sf_256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_permvar_sf_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermps %ymm0, %ymm1, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0x16,0xd0]
; CHECK-NEXT:    vpermps %ymm0, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xa9,0x16,0xd8]
; CHECK-NEXT:    vpermps %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0x16,0xc0]
; CHECK-NEXT:    vaddps %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xcb]
; 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.permvar.sf.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 %x3)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.permvar.sf.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> zeroinitializer, i8 %x3)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.permvar.sf.256(<8 x float> %x0, <8 x i32> %x1, <8 x float> %x2, i8 -1)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res3, %res2
  ret <8 x float> %res4
}

declare <8 x i32> @llvm.x86.avx512.mask.permvar.si.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_permvar_si_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_permvar_si_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermd %ymm0, %ymm1, %ymm2 {%k1} ## encoding: [0x62,0xf2,0x75,0x29,0x36,0xd0]
; CHECK-NEXT:    vpermd %ymm0, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf2,0x75,0xa9,0x36,0xd8]
; CHECK-NEXT:    vpermd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf2,0x75,0x28,0x36,0xc0]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xcb]
; 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.permvar.si.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.permvar.si.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> zeroinitializer, i8 %x3)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.permvar.si.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 -1)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res3, %res2
  ret <8 x i32> %res4
}

declare <2 x double> @llvm.x86.avx512.mask.fixupimm.pd.128(<2 x double>, <2 x double>, <2 x i64>, i32, i8)

define <2 x double>@test_int_x86_avx512_mask_fixupimm_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x i64> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
; CHECK-NEXT:    vfixupimmpd $5, %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf3,0xf5,0x09,0x54,0xda,0x05]
; CHECK-NEXT:    vpxord %xmm4, %xmm4, %xmm4 ## encoding: [0x62,0xf1,0x5d,0x08,0xef,0xe4]
; CHECK-NEXT:    vfixupimmpd $4, %xmm2, %xmm1, %xmm4 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0x89,0x54,0xe2,0x04]
; CHECK-NEXT:    vfixupimmpd $3, %xmm2, %xmm1, %xmm0 ## encoding: [0x62,0xf3,0xf5,0x08,0x54,0xc2,0x03]
; CHECK-NEXT:    vaddpd %xmm4, %xmm3, %xmm1 ## encoding: [0x62,0xf1,0xe5,0x08,0x58,0xcc]
; 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.fixupimm.pd.128(<2 x double> %x0, <2 x double> %x1,<2 x i64> %x2, i32 5, i8 %x4)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.fixupimm.pd.128(<2 x double> zeroinitializer, <2 x double> %x1, <2 x i64> %x2, i32 4, i8 %x4)
  %res2 = call <2 x double> @llvm.x86.avx512.mask.fixupimm.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x i64> %x2, i32 3, i8 -1)
  %res3 = fadd <2 x double> %res, %res1
  %res4 = fadd <2 x double> %res3, %res2
  ret <2 x double> %res4
}

declare <2 x double> @llvm.x86.avx512.maskz.fixupimm.pd.128(<2 x double>, <2 x double>, <2 x i64>, i32, i8)

define <2 x double>@test_int_x86_avx512_maskz_fixupimm_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x i64> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
; CHECK-NEXT:    vfixupimmpd $5, %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0x89,0x54,0xda,0x05]
; CHECK-NEXT:    vpxord %xmm2, %xmm2, %xmm2 ## encoding: [0x62,0xf1,0x6d,0x08,0xef,0xd2]
; CHECK-NEXT:    vfixupimmpd $3, %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0x89,0x54,0xc2,0x03]
; 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.fixupimm.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x i64> %x2, i32 5, i8 %x4)
  %res1 = call <2 x double> @llvm.x86.avx512.maskz.fixupimm.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x i64> zeroinitializer, i32 3, i8 %x4)
  ;%res2 = call <2 x double> @llvm.x86.avx512.maskz.fixupimm.pd.128(<2 x double> %x0, <2 x double> %x1, <2 x i64> %x2, i32 4, i8 -1)
  %res3 = fadd <2 x double> %res, %res1
  ;%res4 = fadd <2 x double> %res3, %res2
  ret <2 x double> %res3
}

declare <4 x double> @llvm.x86.avx512.mask.fixupimm.pd.256(<4 x double>, <4 x double>, <4 x i64>, i32, i8)

define <4 x double>@test_int_x86_avx512_mask_fixupimm_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x i64> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
; CHECK-NEXT:    vfixupimmpd $4, %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf3,0xf5,0x29,0x54,0xda,0x04]
; CHECK-NEXT:    vpxord %ymm4, %ymm4, %ymm4 ## encoding: [0x62,0xf1,0x5d,0x28,0xef,0xe4]
; CHECK-NEXT:    vfixupimmpd $5, %ymm2, %ymm1, %ymm4 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0xa9,0x54,0xe2,0x05]
; CHECK-NEXT:    vfixupimmpd $3, %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0xf5,0x28,0x54,0xc2,0x03]
; CHECK-NEXT:    vaddpd %ymm4, %ymm3, %ymm1 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xcc]
; 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.fixupimm.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x i64> %x2, i32 4, i8 %x4)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.fixupimm.pd.256(<4 x double> zeroinitializer, <4 x double> %x1, <4 x i64> %x2 , i32 5, i8 %x4)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.fixupimm.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x i64> %x2, i32 3, i8 -1)
  %res3 = fadd <4 x double> %res, %res1
  %res4 = fadd <4 x double> %res3, %res2
  ret <4 x double> %res4
}

declare <4 x double> @llvm.x86.avx512.maskz.fixupimm.pd.256(<4 x double>, <4 x double>, <4 x i64>, i32, i8)

define <4 x double>@test_int_x86_avx512_maskz_fixupimm_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x i64> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
; CHECK-NEXT:    vfixupimmpd $5, %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0xa9,0x54,0xda,0x05]
; CHECK-NEXT:    vpxord %ymm4, %ymm4, %ymm4 ## encoding: [0x62,0xf1,0x5d,0x28,0xef,0xe4]
; CHECK-NEXT:    vmovaps %ymm0, %ymm5 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xe8]
; CHECK-NEXT:    vfixupimmpd $4, %ymm4, %ymm1, %ymm5 {%k1} {z} ## encoding: [0x62,0xf3,0xf5,0xa9,0x54,0xec,0x04]
; CHECK-NEXT:    vfixupimmpd $3, %ymm2, %ymm1, %ymm0 ## encoding: [0x62,0xf3,0xf5,0x28,0x54,0xc2,0x03]
; CHECK-NEXT:    vaddpd %ymm5, %ymm3, %ymm1 ## encoding: [0x62,0xf1,0xe5,0x28,0x58,0xcd]
; 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.maskz.fixupimm.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x i64> %x2, i32 5, i8 %x4)
  %res1 = call <4 x double> @llvm.x86.avx512.maskz.fixupimm.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x i64> zeroinitializer, i32 4, i8 %x4)
  %res2 = call <4 x double> @llvm.x86.avx512.maskz.fixupimm.pd.256(<4 x double> %x0, <4 x double> %x1, <4 x i64> %x2, i32 3, i8 -1)
  %res3 = fadd <4 x double> %res, %res1
  %res4 = fadd <4 x double> %res3, %res2
  ret <4 x double> %res4
}

declare <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float>, <4 x float>, <4 x i32>, i32, i8)

define <4 x float>@test_int_x86_avx512_mask_fixupimm_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm3 {%k1} ## encoding: [0x62,0xf3,0x75,0x09,0x54,0xda,0x05]
; CHECK-NEXT:    vmovaps %xmm0, %xmm4 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xe0]
; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm4 ## encoding: [0x62,0xf3,0x75,0x08,0x54,0xe2,0x05]
; CHECK-NEXT:    vpxord %xmm2, %xmm2, %xmm2 ## encoding: [0x62,0xf1,0x6d,0x08,0xef,0xd2]
; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm0 {%k1} ## encoding: [0x62,0xf3,0x75,0x09,0x54,0xc2,0x05]
; CHECK-NEXT:    vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
; CHECK-NEXT:    vaddps %xmm4, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x58,0xc4]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 5, i8 %x4)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> zeroinitializer, i32 5, i8 %x4)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 5, i8 -1)
  %res3 = fadd <4 x float> %res, %res1
  %res4 = fadd <4 x float> %res3, %res2
  ret <4 x float> %res4
}

declare <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float>, <4 x float>, <4 x i32>, i32, i8)

define <4 x float>@test_int_x86_avx512_maskz_fixupimm_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xd8]
; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm3 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0x89,0x54,0xda,0x05]
; CHECK-NEXT:    vmovaps %xmm0, %xmm4 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0xe0]
; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm4 ## encoding: [0x62,0xf3,0x75,0x08,0x54,0xe2,0x05]
; CHECK-NEXT:    vpxord %xmm2, %xmm2, %xmm2 ## encoding: [0x62,0xf1,0x6d,0x08,0xef,0xd2]
; CHECK-NEXT:    vfixupimmps $5, %xmm2, %xmm1, %xmm0 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0x89,0x54,0xc2,0x05]
; CHECK-NEXT:    vaddps %xmm0, %xmm3, %xmm0 ## encoding: [0x62,0xf1,0x64,0x08,0x58,0xc0]
; CHECK-NEXT:    vaddps %xmm4, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x58,0xc4]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 5, i8 %x4)
  %res1 = call <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> zeroinitializer, i32 5, i8 %x4)
  %res2 = call <4 x float> @llvm.x86.avx512.maskz.fixupimm.ps.128(<4 x float> %x0, <4 x float> %x1, <4 x i32> %x2, i32 5, i8 -1)
  %res3 = fadd <4 x float> %res, %res1
  %res4 = fadd <4 x float> %res3, %res2
  ret <4 x float> %res4
}

declare <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float>, <8 x float>, <8 x i32>, i32, i8)

define <8 x float>@test_int_x86_avx512_mask_fixupimm_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_mask_fixupimm_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm3 {%k1} ## encoding: [0x62,0xf3,0x75,0x29,0x54,0xda,0x05]
; CHECK-NEXT:    vmovaps %ymm0, %ymm4 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xe0]
; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm4 ## encoding: [0x62,0xf3,0x75,0x28,0x54,0xe2,0x05]
; CHECK-NEXT:    vpxord %ymm2, %ymm2, %ymm2 ## encoding: [0x62,0xf1,0x6d,0x28,0xef,0xd2]
; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm0 {%k1} ## encoding: [0x62,0xf3,0x75,0x29,0x54,0xc2,0x05]
; CHECK-NEXT:    vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
; CHECK-NEXT:    vaddps %ymm4, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x58,0xc4]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 5, i8 %x4)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> zeroinitializer, i32 5, i8 %x4)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 5, i8 -1)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res3, %res2
  ret <8 x float> %res4
}

declare <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float>, <8 x float>, <8 x i32>, i32, i8)

define <8 x float>@test_int_x86_avx512_maskz_fixupimm_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i8 %x4) {
; CHECK-LABEL: test_int_x86_avx512_maskz_fixupimm_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovaps %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xd8]
; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm3 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0xa9,0x54,0xda,0x05]
; CHECK-NEXT:    vmovaps %ymm0, %ymm4 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0xe0]
; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm4 ## encoding: [0x62,0xf3,0x75,0x28,0x54,0xe2,0x05]
; CHECK-NEXT:    vpxord %ymm2, %ymm2, %ymm2 ## encoding: [0x62,0xf1,0x6d,0x28,0xef,0xd2]
; CHECK-NEXT:    vfixupimmps $5, %ymm2, %ymm1, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0x75,0xa9,0x54,0xc2,0x05]
; CHECK-NEXT:    vaddps %ymm0, %ymm3, %ymm0 ## encoding: [0x62,0xf1,0x64,0x28,0x58,0xc0]
; CHECK-NEXT:    vaddps %ymm4, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x58,0xc4]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 5, i8 %x4)
  %res1 = call <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> zeroinitializer, i32 5, i8 %x4)
  %res2 = call <8 x float> @llvm.x86.avx512.maskz.fixupimm.ps.256(<8 x float> %x0, <8 x float> %x1, <8 x i32> %x2, i32 5, i8 -1)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res3, %res2
  ret <8 x float> %res4
}

declare i8 @llvm.x86.avx512.ptestm.d.128(<4 x i32>, <4 x i32>,i8)

define i8@test_int_x86_avx512_ptestm_d_128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestm_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vptestmd %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
; CHECK-NEXT:    vptestmd %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf2,0x7d,0x08,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    addb %cl, %al ## encoding: [0x00,0xc8]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call i8 @llvm.x86.avx512.ptestm.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call i8 @llvm.x86.avx512.ptestm.d.128(<4 x i32> %x0, <4 x i32> %x1, i8-1)
  %res2 = add i8 %res, %res1
  ret i8 %res2
}

declare i8 @llvm.x86.avx512.ptestm.d.256(<8 x i32>, <8 x i32>, i8)

define i8@test_int_x86_avx512_ptestm_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestm_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vptestmd %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
; CHECK-NEXT:    vptestmd %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf2,0x7d,0x28,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    addb %cl, %al ## encoding: [0x00,0xc8]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call i8 @llvm.x86.avx512.ptestm.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call i8 @llvm.x86.avx512.ptestm.d.256(<8 x i32> %x0, <8 x i32> %x1, i8-1)
  %res2 = add i8 %res, %res1
  ret i8 %res2
}

declare i8 @llvm.x86.avx512.ptestm.q.128(<2 x i64>, <2 x i64>, i8)

define i8@test_int_x86_avx512_ptestm_q_128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestm_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vptestmq %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
; CHECK-NEXT:    vptestmq %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf2,0xfd,0x08,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    addb %cl, %al ## encoding: [0x00,0xc8]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call i8 @llvm.x86.avx512.ptestm.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call i8 @llvm.x86.avx512.ptestm.q.128(<2 x i64> %x0, <2 x i64> %x1, i8-1)
  %res2 = add i8 %res, %res1
  ret i8 %res2
}

declare i8 @llvm.x86.avx512.ptestm.q.256(<4 x i64>, <4 x i64>, i8)

define i8@test_int_x86_avx512_ptestm_q_256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestm_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vptestmq %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
; CHECK-NEXT:    vptestmq %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf2,0xfd,0x28,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    addb %cl, %al ## encoding: [0x00,0xc8]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call i8 @llvm.x86.avx512.ptestm.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call i8 @llvm.x86.avx512.ptestm.q.256(<4 x i64> %x0, <4 x i64> %x1, i8-1)
  %res2 = add i8 %res, %res1
  ret i8 %res2
}

declare i8 @llvm.x86.avx512.ptestnm.d.128(<4 x i32>, <4 x i32>, i8 %x2)

define i8@test_int_x86_avx512_ptestnm_d_128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestnm_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vptestnmd %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf2,0x7e,0x09,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
; CHECK-NEXT:    vptestnmd %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf2,0x7e,0x08,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    addb %cl, %al ## encoding: [0x00,0xc8]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call i8 @llvm.x86.avx512.ptestnm.d.128(<4 x i32> %x0, <4 x i32> %x1, i8 %x2)
  %res1 = call i8 @llvm.x86.avx512.ptestnm.d.128(<4 x i32> %x0, <4 x i32> %x1, i8-1)
  %res2 = add i8 %res, %res1
  ret i8 %res2
}

declare i8 @llvm.x86.avx512.ptestnm.d.256(<8 x i32>, <8 x i32>, i8 %x2)

define i8@test_int_x86_avx512_ptestnm_d_256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestnm_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vptestnmd %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf2,0x7e,0x29,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
; CHECK-NEXT:    vptestnmd %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf2,0x7e,0x28,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    addb %cl, %al ## encoding: [0x00,0xc8]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call i8 @llvm.x86.avx512.ptestnm.d.256(<8 x i32> %x0, <8 x i32> %x1, i8 %x2)
  %res1 = call i8 @llvm.x86.avx512.ptestnm.d.256(<8 x i32> %x0, <8 x i32> %x1, i8-1)
  %res2 = add i8 %res, %res1
  ret i8 %res2
}

declare i8 @llvm.x86.avx512.ptestnm.q.128(<2 x i64>, <2 x i64>, i8 %x2)

define i8@test_int_x86_avx512_ptestnm_q_128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestnm_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vptestnmq %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfe,0x09,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
; CHECK-NEXT:    vptestnmq %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x08,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    addb %cl, %al ## encoding: [0x00,0xc8]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call i8 @llvm.x86.avx512.ptestnm.q.128(<2 x i64> %x0, <2 x i64> %x1, i8 %x2)
  %res1 = call i8 @llvm.x86.avx512.ptestnm.q.128(<2 x i64> %x0, <2 x i64> %x1, i8-1)
  %res2 = add i8 %res, %res1
  ret i8 %res2
}

declare i8 @llvm.x86.avx512.ptestnm.q.256(<4 x i64>, <4 x i64>, i8 %x2)

define i8@test_int_x86_avx512_ptestnm_q_256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_ptestnm_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vptestnmq %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfe,0x29,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %ecx ## encoding: [0xc5,0xf8,0x93,0xc8]
; CHECK-NEXT:    vptestnmq %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf2,0xfe,0x28,0x27,0xc1]
; CHECK-NEXT:    kmovw %k0, %eax ## encoding: [0xc5,0xf8,0x93,0xc0]
; CHECK-NEXT:    addb %cl, %al ## encoding: [0x00,0xc8]
; CHECK-NEXT:    ## kill: %AL<def> %AL<kill> %EAX<kill>
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call i8 @llvm.x86.avx512.ptestnm.q.256(<4 x i64> %x0, <4 x i64> %x1, i8 %x2)
  %res1 = call i8 @llvm.x86.avx512.ptestnm.q.256(<4 x i64> %x0, <4 x i64> %x1, i8-1)
  %res2 = add i8 %res, %res1
  ret i8 %res2
}

declare <8 x i32> @llvm.x86.avx512.mask.pbroadcast.d.gpr.256(i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pbroadcast_d_gpr_256(i32 %x0, <8 x i32> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pbroadcast_d_gpr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpbroadcastd %edi, %ymm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x7c,0xc7]
; CHECK-NEXT:    vpbroadcastd %edi, %ymm1 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x7c,0xcf]
; CHECK-NEXT:    vpbroadcastd %edi, %ymm2 ## encoding: [0x62,0xf2,0x7d,0x28,0x7c,0xd7]
; CHECK-NEXT:    vpaddd %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xc0]
; 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.pbroadcast.d.gpr.256(i32 %x0, <8 x i32> %x1, i8 -1)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.pbroadcast.d.gpr.256(i32 %x0, <8 x i32> %x1, i8 %mask)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.pbroadcast.d.gpr.256(i32 %x0, <8 x i32> zeroinitializer, i8 %mask)
  %res3 = add <8 x i32> %res, %res1
  %res4 = add <8 x i32> %res2, %res3
  ret <8 x i32> %res4
}

declare <4 x i32> @llvm.x86.avx512.mask.pbroadcast.d.gpr.128(i32, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_pbroadcast_d_gpr_128(i32 %x0, <4 x i32> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pbroadcast_d_gpr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpbroadcastd %edi, %xmm0 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x7c,0xc7]
; CHECK-NEXT:    vpbroadcastd %edi, %xmm1 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x7c,0xcf]
; CHECK-NEXT:    vpbroadcastd %edi, %xmm2 ## encoding: [0x62,0xf2,0x7d,0x08,0x7c,0xd7]
; CHECK-NEXT:    vpaddd %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xc0]
; 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.pbroadcast.d.gpr.128(i32 %x0, <4 x i32> %x1, i8 -1)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.pbroadcast.d.gpr.128(i32 %x0, <4 x i32> %x1, i8 %mask)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.pbroadcast.d.gpr.128(i32 %x0, <4 x i32> zeroinitializer, i8 %mask)
  %res3 = add <4 x i32> %res, %res1
  %res4 = add <4 x i32> %res2, %res3
  ret <4 x i32> %res4
}

declare <4 x i64> @llvm.x86.avx512.mask.pbroadcast.q.gpr.256(i64, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_pbroadcast_q_gpr_256(i64 %x0, <4 x i64> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pbroadcast_q_gpr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpbroadcastq %rdi, %ymm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x7c,0xc7]
; CHECK-NEXT:    vpbroadcastq %rdi, %ymm1 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x7c,0xcf]
; CHECK-NEXT:    vpbroadcastq %rdi, %ymm2 ## encoding: [0x62,0xf2,0xfd,0x28,0x7c,0xd7]
; CHECK-NEXT:    vpaddq %ymm0, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc0]
; 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.pbroadcast.q.gpr.256(i64 %x0, <4 x i64> %x1,i8 -1)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.pbroadcast.q.gpr.256(i64 %x0, <4 x i64> %x1,i8 %mask)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.pbroadcast.q.gpr.256(i64 %x0, <4 x i64> zeroinitializer,i8 %mask)
  %res3 = add <4 x i64> %res, %res1
  %res4 = add <4 x i64> %res2, %res3
  ret <4 x i64> %res4
}

declare <2 x i64> @llvm.x86.avx512.mask.pbroadcast.q.gpr.128(i64, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_pbroadcast_q_gpr_128(i64 %x0, <2 x i64> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_mask_pbroadcast_q_gpr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpbroadcastq %rdi, %xmm0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x7c,0xc7]
; CHECK-NEXT:    vpbroadcastq %rdi, %xmm1 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x7c,0xcf]
; CHECK-NEXT:    vpbroadcastq %rdi, %xmm2 ## encoding: [0x62,0xf2,0xfd,0x08,0x7c,0xd7]
; CHECK-NEXT:    vpaddq %xmm0, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc0]
; 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.pbroadcast.q.gpr.128(i64 %x0, <2 x i64> %x1,i8 -1)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.pbroadcast.q.gpr.128(i64 %x0, <2 x i64> %x1,i8 %mask)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.pbroadcast.q.gpr.128(i64 %x0, <2 x i64> zeroinitializer,i8 %mask)
  %res3 = add <2 x i64> %res, %res1
  %res4 = add <2 x i64> %res2, %res3
  ret <2 x i64> %res4
}