xref: /aosp_15_r20/external/llvm/test/CodeGen/X86/avx512vl-intrinsics-upgrade.ll (revision 9880d6810fe72a1726cb53787c6711e909410d58)

; 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

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

define <8 x i32>@test_int_x86_avx512_pbroadcastd_256(<4 x i32> %x0, <8 x i32> %x1, i8 %mask, i32 * %y_ptr) {
; CHECK-LABEL: test_int_x86_avx512_pbroadcastd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpbroadcastd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x58,0xc8]
; CHECK-NEXT:    vpbroadcastd %xmm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x58,0xc0]
; CHECK-NEXT:    vpaddd (%rsi){1to8}, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x75,0x38,0xfe,0x0e]
; CHECK-NEXT:    vpaddd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfe,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %y_32  = load i32, i32 * %y_ptr
  %y = insertelement <4 x i32> undef, i32 %y_32, i32 0
  %res = call <8 x i32> @llvm.x86.avx512.pbroadcastd.256(<4 x i32> %y, <8 x i32> %x1, i8 -1)
  %res1 = call <8 x i32> @llvm.x86.avx512.pbroadcastd.256(<4 x i32> %x0, <8 x i32> %x1, i8 %mask)
  %res2 = call <8 x i32> @llvm.x86.avx512.pbroadcastd.256(<4 x 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.pbroadcastd.128(<4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_pbroadcastd_128(<4 x i32> %x0, <4 x i32> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_pbroadcastd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpbroadcastd %xmm0, %xmm2 ## encoding: [0x62,0xf2,0x7d,0x08,0x58,0xd0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpbroadcastd %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x58,0xc8]
; CHECK-NEXT:    vpbroadcastd %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x58,0xc0]
; CHECK-NEXT:    vpaddd %xmm1, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xc9]
; 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.pbroadcastd.128(<4 x i32> %x0, <4 x i32> %x1, i8 -1)
  %res1 = call <4 x i32> @llvm.x86.avx512.pbroadcastd.128(<4 x i32> %x0, <4 x i32> %x1, i8 %mask)
  %res2 = call <4 x i32> @llvm.x86.avx512.pbroadcastd.128(<4 x 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.pbroadcastq.256(<2 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_pbroadcastq_256(<2 x i64> %x0, <4 x i64> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_pbroadcastq_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpbroadcastq %xmm0, %ymm2 ## encoding: [0x62,0xf2,0xfd,0x28,0x59,0xd0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpbroadcastq %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x59,0xc8]
; CHECK-NEXT:    vpbroadcastq %xmm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x59,0xc0]
; CHECK-NEXT:    vpaddq %ymm1, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc9]
; CHECK-NEXT:    vpaddq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xd4,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.pbroadcastq.256(<2 x i64> %x0, <4 x i64> %x1,i8 -1)
  %res1 = call <4 x i64> @llvm.x86.avx512.pbroadcastq.256(<2 x i64> %x0, <4 x i64> %x1,i8 %mask)
  %res2 = call <4 x i64> @llvm.x86.avx512.pbroadcastq.256(<2 x 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.pbroadcastq.128(<2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_pbroadcastq_128(<2 x i64> %x0, <2 x i64> %x1, i8 %mask) {
; CHECK-LABEL: test_int_x86_avx512_pbroadcastq_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpbroadcastq %xmm0, %xmm2 ## encoding: [0x62,0xf2,0xfd,0x08,0x59,0xd0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpbroadcastq %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x59,0xc8]
; CHECK-NEXT:    vpbroadcastq %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0x89,0x59,0xc0]
; CHECK-NEXT:    vpaddq %xmm1, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc9]
; CHECK-NEXT:    vpaddq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xd4,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.pbroadcastq.128(<2 x i64> %x0, <2 x i64> %x1,i8 -1)
  %res1 = call <2 x i64> @llvm.x86.avx512.pbroadcastq.128(<2 x i64> %x0, <2 x i64> %x1,i8 %mask)
  %res2 = call <2 x i64> @llvm.x86.avx512.pbroadcastq.128(<2 x 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
}

declare <4 x double> @llvm.x86.avx512.mask.broadcast.sd.pd.256(<2 x double>, <4 x double>, i8) nounwind readonly

define <4 x double> @test_x86_vbroadcast_sd_pd_256(<2 x double> %a0, <4 x double> %a1, i8 %mask ) {
; CHECK-LABEL: test_x86_vbroadcast_sd_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vbroadcastsd %xmm0, %ymm2 ## encoding: [0x62,0xf2,0xfd,0x28,0x19,0xd0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vbroadcastsd %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x19,0xc8]
; CHECK-NEXT:    vbroadcastsd %xmm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0xfd,0xa9,0x19,0xc0]
; CHECK-NEXT:    vaddpd %ymm1, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0xed,0x28,0x58,0xc9]
; 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.broadcast.sd.pd.256(<2 x double> %a0, <4 x double> zeroinitializer, i8 -1)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.broadcast.sd.pd.256(<2 x double> %a0, <4 x double> %a1, i8 %mask)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.broadcast.sd.pd.256(<2 x double> %a0, <4 x double> zeroinitializer, i8 %mask)
  %res3 = fadd <4 x double> %res, %res1
  %res4 = fadd <4 x double> %res2, %res3
  ret <4 x double> %res4
}

declare <8 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.256(<4 x float>, <8 x float>, i8) nounwind readonly

define <8 x float> @test_x86_vbroadcast_ss_ps_256(<4 x float> %a0, <8 x float> %a1, i8 %mask ) {
; CHECK-LABEL: test_x86_vbroadcast_ss_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vbroadcastss %xmm0, %ymm2 ## encoding: [0x62,0xf2,0x7d,0x28,0x18,0xd0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vbroadcastss %xmm0, %ymm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x29,0x18,0xc8]
; CHECK-NEXT:    vbroadcastss %xmm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0xa9,0x18,0xc0]
; CHECK-NEXT:    vaddps %ymm1, %ymm2, %ymm1 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xc9]
; 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.broadcast.ss.ps.256(<4 x float> %a0, <8 x float> zeroinitializer, i8 -1)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.256(<4 x float> %a0, <8 x float> %a1, i8 %mask)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.256(<4 x float> %a0, <8 x float> zeroinitializer, i8 %mask)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res2, %res3
  ret <8 x float> %res4
}

declare <4 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.128(<4 x float>, <4 x float>, i8) nounwind readonly

define <4 x float> @test_x86_vbroadcast_ss_ps_128(<4 x float> %a0, <4 x float> %a1, i8 %mask ) {
; CHECK-LABEL: test_x86_vbroadcast_ss_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vbroadcastss %xmm0, %xmm2 ## encoding: [0x62,0xf2,0x7d,0x08,0x18,0xd0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vbroadcastss %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf2,0x7d,0x09,0x18,0xc8]
; CHECK-NEXT:    vbroadcastss %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf2,0x7d,0x89,0x18,0xc0]
; CHECK-NEXT:    vaddps %xmm1, %xmm2, %xmm1 ## encoding: [0x62,0xf1,0x6c,0x08,0x58,0xc9]
; 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.broadcast.ss.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 -1)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.128(<4 x float> %a0, <4 x float> %a1, i8 %mask)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.broadcast.ss.ps.128(<4 x float> %a0, <4 x float> zeroinitializer, i8 %mask)
  %res3 = fadd <4 x float> %res, %res1
  %res4 = fadd <4 x float> %res2, %res3
  ret <4 x float> %res4
}

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

define <4 x float>@test_int_x86_avx512_mask_movsldup_128(<4 x float> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movsldup_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovsldup %xmm0, %xmm2 ## encoding: [0x62,0xf1,0x7e,0x08,0x12,0xd0]
; CHECK-NEXT:    ## xmm2 = xmm0[0,0,2,2]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovsldup %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x09,0x12,0xc8]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[0,0,2,2]
; CHECK-NEXT:    vmovsldup %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7e,0x89,0x12,0xc0]
; CHECK-NEXT:    ## xmm0 {%k1} {z} = xmm0[0,0,2,2]
; 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.movsldup.128(<4 x float> %x0, <4 x float> %x1, i8 %x2)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.movsldup.128(<4 x float> %x0, <4 x float> %x1, i8 -1)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.movsldup.128(<4 x float> %x0, <4 x float> zeroinitializer, i8 %x2)
  %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.movsldup.256(<8 x float>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_movsldup_256(<8 x float> %x0, <8 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movsldup_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovsldup %ymm0, %ymm2 ## encoding: [0x62,0xf1,0x7e,0x28,0x12,0xd0]
; CHECK-NEXT:    ## ymm2 = ymm0[0,0,2,2,4,4,6,6]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovsldup %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x29,0x12,0xc8]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,0,2,2,4,4,6,6]
; CHECK-NEXT:    vmovsldup %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7e,0xa9,0x12,0xc0]
; CHECK-NEXT:    ## ymm0 {%k1} {z} = ymm0[0,0,2,2,4,4,6,6]
; CHECK-NEXT:    vaddps %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x74,0x28,0x58,0xca]
; 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.movsldup.256(<8 x float> %x0, <8 x float> %x1, i8 %x2)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.movsldup.256(<8 x float> %x0, <8 x float> %x1, i8 -1)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.movsldup.256(<8 x float> %x0, <8 x float> zeroinitializer, i8 %x2)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res2, %res3
  ret <8 x float> %res4
}

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

define <4 x float>@test_int_x86_avx512_mask_movshdup_128(<4 x float> %x0, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movshdup_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovshdup %xmm0, %xmm2 ## encoding: [0x62,0xf1,0x7e,0x08,0x16,0xd0]
; CHECK-NEXT:    ## xmm2 = xmm0[1,1,3,3]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovshdup %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x09,0x16,0xc8]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[1,1,3,3]
; CHECK-NEXT:    vmovshdup %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7e,0x89,0x16,0xc0]
; CHECK-NEXT:    ## xmm0 {%k1} {z} = xmm0[1,1,3,3]
; 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.movshdup.128(<4 x float> %x0, <4 x float> %x1, i8 %x2)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.movshdup.128(<4 x float> %x0, <4 x float> %x1, i8 -1)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.movshdup.128(<4 x float> %x0, <4 x float> zeroinitializer, i8 %x2)
  %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.movshdup.256(<8 x float>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_movshdup_256(<8 x float> %x0, <8 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movshdup_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovshdup %ymm0, %ymm2 ## encoding: [0x62,0xf1,0x7e,0x28,0x16,0xd0]
; CHECK-NEXT:    ## ymm2 = ymm0[1,1,3,3,5,5,7,7]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovshdup %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7e,0x29,0x16,0xc8]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[1,1,3,3,5,5,7,7]
; CHECK-NEXT:    vmovshdup %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7e,0xa9,0x16,0xc0]
; CHECK-NEXT:    ## ymm0 {%k1} {z} = ymm0[1,1,3,3,5,5,7,7]
; CHECK-NEXT:    vaddps %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0x74,0x28,0x58,0xca]
; 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.movshdup.256(<8 x float> %x0, <8 x float> %x1, i8 %x2)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.movshdup.256(<8 x float> %x0, <8 x float> %x1, i8 -1)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.movshdup.256(<8 x float> %x0, <8 x float> zeroinitializer, i8 %x2)
  %res3 = fadd <8 x float> %res, %res1
  %res4 = fadd <8 x float> %res2, %res3
  ret <8 x float> %res4
}
declare <2 x double> @llvm.x86.avx512.mask.movddup.128(<2 x double>, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_movddup_128(<2 x double> %x0, <2 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movddup_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovddup %xmm0, %xmm2 ## encoding: [0x62,0xf1,0xff,0x08,0x12,0xd0]
; CHECK-NEXT:    ## xmm2 = xmm0[0,0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovddup %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xff,0x09,0x12,0xc8]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[0,0]
; CHECK-NEXT:    vmovddup %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0xff,0x89,0x12,0xc0]
; CHECK-NEXT:    ## xmm0 {%k1} {z} = xmm0[0,0]
; CHECK-NEXT:    vaddpd %xmm2, %xmm1, %xmm1 ## encoding: [0x62,0xf1,0xf5,0x08,0x58,0xca]
; CHECK-NEXT:    vaddpd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x58,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.mask.movddup.128(<2 x double> %x0, <2 x double> %x1, i8 %x2)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.movddup.128(<2 x double> %x0, <2 x double> %x1, i8 -1)
  %res2 = call <2 x double> @llvm.x86.avx512.mask.movddup.128(<2 x double> %x0, <2 x double> zeroinitializer, i8 %x2)
  %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.movddup.256(<4 x double>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_movddup_256(<4 x double> %x0, <4 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_movddup_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovddup %ymm0, %ymm2 ## encoding: [0x62,0xf1,0xff,0x28,0x12,0xd0]
; CHECK-NEXT:    ## ymm2 = ymm0[0,0,2,2]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vmovddup %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xff,0x29,0x12,0xc8]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,0,2,2]
; CHECK-NEXT:    vmovddup %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xff,0xa9,0x12,0xc0]
; CHECK-NEXT:    ## ymm0 {%k1} {z} = ymm0[0,0,2,2]
; CHECK-NEXT:    vaddpd %ymm2, %ymm1, %ymm1 ## encoding: [0x62,0xf1,0xf5,0x28,0x58,0xca]
; 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.movddup.256(<4 x double> %x0, <4 x double> %x1, i8 %x2)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.movddup.256(<4 x double> %x0, <4 x double> %x1, i8 -1)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.movddup.256(<4 x double> %x0, <4 x double> zeroinitializer, i8 %x2)
  %res3 = fadd <4 x double> %res, %res1
  %res4 = fadd <4 x double> %res2, %res3
  ret <4 x double> %res4
}

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

define <4 x double>@test_int_x86_avx512_mask_vpermil_pd_256(<4 x double> %x0, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermil_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpermilpd $6, %ymm0, %ymm2 ## encoding: [0x62,0xf3,0xfd,0x28,0x05,0xd0,0x06]
; CHECK-NEXT:    ## ymm2 = ymm0[0,1,3,2]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermilpd $6, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x05,0xc8,0x06]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[0,1,3,2]
; CHECK-NEXT:    vpermilpd $6, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x05,0xc0,0x06]
; CHECK-NEXT:    ## ymm0 {%k1} {z} = ymm0[0,1,3,2]
; CHECK-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0x58,0xc0]
; 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.vpermil.pd.256(<4 x double> %x0, i32 22, <4 x double> %x2, i8 %x3)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.vpermil.pd.256(<4 x double> %x0, i32 22, <4 x double> zeroinitializer, i8 %x3)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.vpermil.pd.256(<4 x double> %x0, i32 22, <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.vpermil.pd.128(<2 x double>, i32, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_vpermil_pd_128(<2 x double> %x0, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermil_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpermilpd $1, %xmm0, %xmm2 ## encoding: [0x62,0xf3,0xfd,0x08,0x05,0xd0,0x01]
; CHECK-NEXT:    ## xmm2 = xmm0[1,0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermilpd $1, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x09,0x05,0xc8,0x01]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[1,0]
; CHECK-NEXT:    vpermilpd $1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0x89,0x05,0xc0,0x01]
; CHECK-NEXT:    ## xmm0 {%k1} {z} = xmm0[1,0]
; CHECK-NEXT:    vaddpd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0xf5,0x08,0x58,0xc0]
; CHECK-NEXT:    vaddpd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x58,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.mask.vpermil.pd.128(<2 x double> %x0, i32 1, <2 x double> %x2, i8 %x3)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.vpermil.pd.128(<2 x double> %x0, i32 1, <2 x double> zeroinitializer, i8 %x3)
  %res2 = call <2 x double> @llvm.x86.avx512.mask.vpermil.pd.128(<2 x double> %x0, i32 1, <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.vpermil.ps.256(<8 x float>, i32, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_vpermil_ps_256(<8 x float> %x0, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermil_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpermilps $22, %ymm0, %ymm2 ## encoding: [0x62,0xf3,0x7d,0x28,0x04,0xd0,0x16]
; CHECK-NEXT:    ## ymm2 = ymm0[2,1,1,0,6,5,5,4]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermilps $22, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x29,0x04,0xc8,0x16]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[2,1,1,0,6,5,5,4]
; CHECK-NEXT:    vpermilps $22, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0xa9,0x04,0xc0,0x16]
; CHECK-NEXT:    ## ymm0 {%k1} {z} = ymm0[2,1,1,0,6,5,5,4]
; CHECK-NEXT:    vaddps %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x74,0x28,0x58,0xc0]
; CHECK-NEXT:    vaddps %ymm2, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x58,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.vpermil.ps.256(<8 x float> %x0, i32 22, <8 x float> %x2, i8 %x3)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.vpermil.ps.256(<8 x float> %x0, i32 22, <8 x float> zeroinitializer, i8 %x3)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.vpermil.ps.256(<8 x float> %x0, i32 22, <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.vpermil.ps.128(<4 x float>, i32, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_vpermil_ps_128(<4 x float> %x0, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_vpermil_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpermilps $22, %xmm0, %xmm2 ## encoding: [0x62,0xf3,0x7d,0x08,0x04,0xd0,0x16]
; CHECK-NEXT:    ## xmm2 = xmm0[2,1,1,0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpermilps $22, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf3,0x7d,0x09,0x04,0xc8,0x16]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[2,1,1,0]
; CHECK-NEXT:    vpermilps $22, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf3,0x7d,0x89,0x04,0xc0,0x16]
; CHECK-NEXT:    ## xmm0 {%k1} {z} = xmm0[2,1,1,0]
; CHECK-NEXT:    vaddps %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x74,0x08,0x58,0xc0]
; 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.vpermil.ps.128(<4 x float> %x0, i32 22, <4 x float> %x2, i8 %x3)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.vpermil.ps.128(<4 x float> %x0, i32 22, <4 x float> zeroinitializer, i8 %x3)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.vpermil.ps.128(<4 x float> %x0, i32 22, <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 <4 x double> @llvm.x86.avx512.mask.perm.df.256(<4 x double>, i32, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_perm_df_256(<4 x double> %x0, i32 %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_perm_df_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpermpd $3, %ymm0, %ymm2 ## encoding: [0x62,0xf3,0xfd,0x28,0x01,0xd0,0x03]
; CHECK-NEXT:    ## ymm2 = ymm0[3,0,0,0]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpermpd $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x01,0xc8,0x03]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[3,0,0,0]
; CHECK-NEXT:    vpermpd $3, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x01,0xc0,0x03]
; CHECK-NEXT:    ## ymm0 {%k1} {z} = ymm0[3,0,0,0]
; CHECK-NEXT:    vaddpd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0x58,0xc0]
; CHECK-NEXT:    vaddpd %ymm2, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x58,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.perm.df.256(<4 x double> %x0, i32 3, <4 x double> %x2, i8 %x3)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.perm.df.256(<4 x double> %x0, i32 3, <4 x double> zeroinitializer, i8 %x3)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.perm.df.256(<4 x double> %x0, i32 3, <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.perm.di.256(<4 x i64>, i32, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_perm_di_256(<4 x i64> %x0, i32 %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_perm_di_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpermq $3, %ymm0, %ymm2 ## encoding: [0x62,0xf3,0xfd,0x28,0x00,0xd0,0x03]
; CHECK-NEXT:    ## ymm2 = ymm0[3,0,0,0]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpermq $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf3,0xfd,0x29,0x00,0xc8,0x03]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[3,0,0,0]
; CHECK-NEXT:    vpermq $3, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf3,0xfd,0xa9,0x00,0xc0,0x03]
; CHECK-NEXT:    ## ymm0 {%k1} {z} = ymm0[3,0,0,0]
; CHECK-NEXT:    vpaddq %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0xf5,0x28,0xd4,0xc0]
; CHECK-NEXT:    vpaddq %ymm2, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xd4,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.perm.di.256(<4 x i64> %x0, i32 3, <4 x i64> %x2, i8 %x3)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.perm.di.256(<4 x i64> %x0, i32 3, <4 x i64> zeroinitializer, i8 %x3)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.perm.di.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 void @llvm.x86.avx512.mask.store.pd.128(i8*, <2 x double>, i8)

define void@test_int_x86_avx512_mask_store_pd_128(i8* %ptr1, i8* %ptr2, <2 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_store_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovapd %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x29,0x07]
; CHECK-NEXT:    vmovapd %xmm0, (%rsi) ## encoding: [0x62,0xf1,0xfd,0x08,0x29,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.store.pd.128(i8* %ptr1, <2 x double> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.store.pd.128(i8* %ptr2, <2 x double> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_store_pd_256(i8* %ptr1, i8* %ptr2, <4 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_store_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovapd %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x29,0x07]
; CHECK-NEXT:    vmovapd %ymm0, (%rsi) ## encoding: [0x62,0xf1,0xfd,0x28,0x29,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.store.pd.256(i8* %ptr1, <4 x double> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.store.pd.256(i8* %ptr2, <4 x double> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_storeu_pd_128(i8* %ptr1, i8* %ptr2, <2 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovupd %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x11,0x07]
; CHECK-NEXT:    vmovupd %xmm0, (%rsi) ## encoding: [0x62,0xf1,0xfd,0x08,0x11,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.storeu.pd.128(i8* %ptr1, <2 x double> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.storeu.pd.128(i8* %ptr2, <2 x double> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_storeu_pd_256(i8* %ptr1, i8* %ptr2, <4 x double> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovupd %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x11,0x07]
; CHECK-NEXT:    vmovupd %ymm0, (%rsi) ## encoding: [0x62,0xf1,0xfd,0x28,0x11,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.storeu.pd.256(i8* %ptr1, <4 x double> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.storeu.pd.256(i8* %ptr2, <4 x double> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_store_ps_128(i8* %ptr1, i8* %ptr2, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_store_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovaps %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x29,0x07]
; CHECK-NEXT:    vmovaps %xmm0, (%rsi) ## encoding: [0x62,0xf1,0x7c,0x08,0x29,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.store.ps.128(i8* %ptr1, <4 x float> %x1, i8 %x2)
    call void @llvm.x86.avx512.mask.store.ps.128(i8* %ptr2, <4 x float> %x1, i8 -1)
    ret void
}

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

define void@test_int_x86_avx512_mask_store_ps_256(i8* %ptr1, i8* %ptr2, <8 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_store_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovaps %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x29,0x07]
; CHECK-NEXT:    vmovaps %ymm0, (%rsi) ## encoding: [0x62,0xf1,0x7c,0x28,0x29,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.store.ps.256(i8* %ptr1, <8 x float> %x1, i8 %x2)
    call void @llvm.x86.avx512.mask.store.ps.256(i8* %ptr2, <8 x float> %x1, i8 -1)
    ret void
}

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

define void@test_int_x86_avx512_mask_storeu_ps_128(i8* %ptr1, i8* %ptr2, <4 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovups %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x11,0x07]
; CHECK-NEXT:    vmovups %xmm0, (%rsi) ## encoding: [0x62,0xf1,0x7c,0x08,0x11,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.storeu.ps.128(i8* %ptr1, <4 x float> %x1, i8 %x2)
    call void @llvm.x86.avx512.mask.storeu.ps.128(i8* %ptr2, <4 x float> %x1, i8 -1)
    ret void
}

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

define void@test_int_x86_avx512_mask_storeu_ps_256(i8* %ptr1, i8* %ptr2, <8 x float> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovups %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x11,0x07]
; CHECK-NEXT:    vmovups %ymm0, (%rsi) ## encoding: [0x62,0xf1,0x7c,0x28,0x11,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
    call void @llvm.x86.avx512.mask.storeu.ps.256(i8* %ptr1, <8 x float> %x1, i8 %x2)
    call void @llvm.x86.avx512.mask.storeu.ps.256(i8* %ptr2, <8 x float> %x1, i8 -1)
    ret void
}

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

define void@test_int_x86_avx512_mask_storeu_q_128(i8* %ptr1, i8* %ptr2, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqu64 %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xfe,0x09,0x7f,0x07]
; CHECK-NEXT:    vmovdqu64 %xmm0, (%rsi) ## encoding: [0x62,0xf1,0xfe,0x08,0x7f,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.storeu.q.128(i8* %ptr1, <2 x i64> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.storeu.q.128(i8* %ptr2, <2 x i64> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_storeu_q_256(i8* %ptr1, i8* %ptr2, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqu64 %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xfe,0x29,0x7f,0x07]
; CHECK-NEXT:    vmovdqu64 %ymm0, (%rsi) ## encoding: [0x62,0xf1,0xfe,0x28,0x7f,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.storeu.q.256(i8* %ptr1, <4 x i64> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.storeu.q.256(i8* %ptr2, <4 x i64> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_storeu_d_128(i8* %ptr1, i8* %ptr2, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqu32 %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7e,0x09,0x7f,0x07]
; CHECK-NEXT:    vmovdqu32 %xmm0, (%rsi) ## encoding: [0x62,0xf1,0x7e,0x08,0x7f,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.storeu.d.128(i8* %ptr1, <4 x i32> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.storeu.d.128(i8* %ptr2, <4 x i32> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_storeu_d_256(i8* %ptr1, i8* %ptr2, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_storeu_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqu32 %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7e,0x29,0x7f,0x07]
; CHECK-NEXT:    vmovdqu32 %ymm0, (%rsi) ## encoding: [0x62,0xf1,0x7e,0x28,0x7f,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.storeu.d.256(i8* %ptr1, <8 x i32> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.storeu.d.256(i8* %ptr2, <8 x i32> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_store_q_128(i8* %ptr1, i8* %ptr2, <2 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_store_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqa64 %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x7f,0x07]
; CHECK-NEXT:    vmovdqa64 %xmm0, (%rsi) ## encoding: [0x62,0xf1,0xfd,0x08,0x7f,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.store.q.128(i8* %ptr1, <2 x i64> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.store.q.128(i8* %ptr2, <2 x i64> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_store_q_256(i8* %ptr1, i8* %ptr2, <4 x i64> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_store_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqa64 %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x7f,0x07]
; CHECK-NEXT:    vmovdqa64 %ymm0, (%rsi) ## encoding: [0x62,0xf1,0xfd,0x28,0x7f,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.store.q.256(i8* %ptr1, <4 x i64> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.store.q.256(i8* %ptr2, <4 x i64> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_store_d_128(i8* %ptr1, i8* %ptr2, <4 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_store_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqa32 %xmm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x7f,0x07]
; CHECK-NEXT:    vmovdqa32 %xmm0, (%rsi) ## encoding: [0x62,0xf1,0x7d,0x08,0x7f,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.store.d.128(i8* %ptr1, <4 x i32> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.store.d.128(i8* %ptr2, <4 x i32> %x1, i8 -1)
  ret void
}

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

define void@test_int_x86_avx512_mask_store_d_256(i8* %ptr1, i8* %ptr2, <8 x i32> %x1, i8 %x2) {
; CHECK-LABEL: test_int_x86_avx512_mask_store_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqa32 %ymm0, (%rdi) {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x7f,0x07]
; CHECK-NEXT:    vmovdqa32 %ymm0, (%rsi) ## encoding: [0x62,0xf1,0x7d,0x28,0x7f,0x06]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  call void @llvm.x86.avx512.mask.store.d.256(i8* %ptr1, <8 x i32> %x1, i8 %x2)
  call void @llvm.x86.avx512.mask.store.d.256(i8* %ptr2, <8 x i32> %x1, i8 -1)
  ret void
}

define <8 x float> @test_mask_load_aligned_ps_256(<8 x float> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_aligned_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovaps (%rdi), %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x28,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovaps (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x28,0x07]
; CHECK-NEXT:    vmovaps (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x28,0x0f]
; 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.load.ps.256(i8* %ptr, <8 x float> zeroinitializer, i8 -1)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.load.ps.256(i8* %ptr, <8 x float> %res, i8 %mask)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.load.ps.256(i8* %ptr, <8 x float> zeroinitializer, i8 %mask)
  %res4 = fadd <8 x float> %res2, %res1
  ret <8 x float> %res4
}

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

define <8 x float> @test_mask_load_unaligned_ps_256(<8 x float> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_unaligned_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovups (%rdi), %ymm0 ## encoding: [0x62,0xf1,0x7c,0x28,0x10,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovups (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x10,0x07]
; CHECK-NEXT:    vmovups (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0xa9,0x10,0x0f]
; 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.loadu.ps.256(i8* %ptr, <8 x float> zeroinitializer, i8 -1)
  %res1 = call <8 x float> @llvm.x86.avx512.mask.loadu.ps.256(i8* %ptr, <8 x float> %res, i8 %mask)
  %res2 = call <8 x float> @llvm.x86.avx512.mask.loadu.ps.256(i8* %ptr, <8 x float> zeroinitializer, i8 %mask)
  %res4 = fadd <8 x float> %res2, %res1
  ret <8 x float> %res4
}

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

define <4 x double> @test_mask_load_aligned_pd_256(<4 x double> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_aligned_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovapd (%rdi), %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x28,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovapd (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x28,0x07]
; CHECK-NEXT:    vmovapd (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0x28,0x0f]
; 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.load.pd.256(i8* %ptr, <4 x double> zeroinitializer, i8 -1)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.load.pd.256(i8* %ptr, <4 x double> %res, i8 %mask)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.load.pd.256(i8* %ptr, <4 x double> zeroinitializer, i8 %mask)
  %res4 = fadd <4 x double> %res2, %res1
  ret <4 x double> %res4
}

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

define <4 x double> @test_mask_load_unaligned_pd_256(<4 x double> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_unaligned_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovupd (%rdi), %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x10,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovupd (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x10,0x07]
; CHECK-NEXT:    vmovupd (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0x10,0x0f]
; 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.loadu.pd.256(i8* %ptr, <4 x double> zeroinitializer, i8 -1)
  %res1 = call <4 x double> @llvm.x86.avx512.mask.loadu.pd.256(i8* %ptr, <4 x double> %res, i8 %mask)
  %res2 = call <4 x double> @llvm.x86.avx512.mask.loadu.pd.256(i8* %ptr, <4 x double> zeroinitializer, i8 %mask)
  %res4 = fadd <4 x double> %res2, %res1
  ret <4 x double> %res4
}

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

define <4 x float> @test_mask_load_aligned_ps_128(<4 x float> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_aligned_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovaps (%rdi), %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x28,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovaps (%rdi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x28,0x07]
; CHECK-NEXT:    vmovaps (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x28,0x0f]
; 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.load.ps.128(i8* %ptr, <4 x float> zeroinitializer, i8 -1)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.load.ps.128(i8* %ptr, <4 x float> %res, i8 %mask)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.load.ps.128(i8* %ptr, <4 x float> zeroinitializer, i8 %mask)
  %res4 = fadd <4 x float> %res2, %res1
  ret <4 x float> %res4
}

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

define <4 x float> @test_mask_load_unaligned_ps_128(<4 x float> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_unaligned_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovups (%rdi), %xmm0 ## encoding: [0x62,0xf1,0x7c,0x08,0x10,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovups (%rdi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x10,0x07]
; CHECK-NEXT:    vmovups (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7c,0x89,0x10,0x0f]
; 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.loadu.ps.128(i8* %ptr, <4 x float> zeroinitializer, i8 -1)
  %res1 = call <4 x float> @llvm.x86.avx512.mask.loadu.ps.128(i8* %ptr, <4 x float> %res, i8 %mask)
  %res2 = call <4 x float> @llvm.x86.avx512.mask.loadu.ps.128(i8* %ptr, <4 x float> zeroinitializer, i8 %mask)
  %res4 = fadd <4 x float> %res2, %res1
  ret <4 x float> %res4
}

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

define <2 x double> @test_mask_load_aligned_pd_128(<2 x double> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_aligned_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovapd (%rdi), %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x28,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovapd (%rdi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x28,0x07]
; CHECK-NEXT:    vmovapd (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0x28,0x0f]
; 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.load.pd.128(i8* %ptr, <2 x double> zeroinitializer, i8 -1)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.load.pd.128(i8* %ptr, <2 x double> %res, i8 %mask)
  %res2 = call <2 x double> @llvm.x86.avx512.mask.load.pd.128(i8* %ptr, <2 x double> zeroinitializer, i8 %mask)
  %res4 = fadd <2 x double> %res2, %res1
  ret <2 x double> %res4
}

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

define <2 x double> @test_mask_load_unaligned_pd_128(<2 x double> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_unaligned_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovupd (%rdi), %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x10,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovupd (%rdi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x10,0x07]
; CHECK-NEXT:    vmovupd (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0x10,0x0f]
; 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.loadu.pd.128(i8* %ptr, <2 x double> zeroinitializer, i8 -1)
  %res1 = call <2 x double> @llvm.x86.avx512.mask.loadu.pd.128(i8* %ptr, <2 x double> %res, i8 %mask)
  %res2 = call <2 x double> @llvm.x86.avx512.mask.loadu.pd.128(i8* %ptr, <2 x double> zeroinitializer, i8 %mask)
  %res4 = fadd <2 x double> %res2, %res1
  ret <2 x double> %res4
}

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

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

define <4 x i32> @test_mask_load_unaligned_d_128(i8* %ptr, i8* %ptr2, <4 x i32> %data, i8 %mask) {
; CHECK-LABEL: test_mask_load_unaligned_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqu32 (%rdi), %xmm0 ## encoding: [0x62,0xf1,0x7e,0x08,0x6f,0x07]
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqu32 (%rsi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0x7e,0x09,0x6f,0x06]
; CHECK-NEXT:    vmovdqu32 (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7e,0x89,0x6f,0x0f]
; 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.loadu.d.128(i8* %ptr, <4 x i32> zeroinitializer, i8 -1)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.loadu.d.128(i8* %ptr2, <4 x i32> %res, i8 %mask)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.loadu.d.128(i8* %ptr, <4 x i32> zeroinitializer, i8 %mask)
  %res4 = add <4 x i32> %res2, %res1
  ret <4 x i32> %res4
}

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

define <8 x i32> @test_mask_load_unaligned_d_256(i8* %ptr, i8* %ptr2, <8 x i32> %data, i8 %mask) {
; CHECK-LABEL: test_mask_load_unaligned_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqu32 (%rdi), %ymm0 ## encoding: [0x62,0xf1,0x7e,0x28,0x6f,0x07]
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqu32 (%rsi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0x7e,0x29,0x6f,0x06]
; CHECK-NEXT:    vmovdqu32 (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7e,0xa9,0x6f,0x0f]
; 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.loadu.d.256(i8* %ptr, <8 x i32> zeroinitializer, i8 -1)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.loadu.d.256(i8* %ptr2, <8 x i32> %res, i8 %mask)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.loadu.d.256(i8* %ptr, <8 x i32> zeroinitializer, i8 %mask)
  %res4 = add <8 x i32> %res2, %res1
  ret <8 x i32> %res4
}

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

define <2 x i64> @test_mask_load_unaligned_q_128(i8* %ptr, i8* %ptr2, <2 x i64> %data, i8 %mask) {
; CHECK-LABEL: test_mask_load_unaligned_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqu64 (%rdi), %xmm0 ## encoding: [0x62,0xf1,0xfe,0x08,0x6f,0x07]
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqu64 (%rsi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0xfe,0x09,0x6f,0x06]
; CHECK-NEXT:    vmovdqu64 (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0xfe,0x89,0x6f,0x0f]
; 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.loadu.q.128(i8* %ptr, <2 x i64> zeroinitializer, i8 -1)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.loadu.q.128(i8* %ptr2, <2 x i64> %res, i8 %mask)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.loadu.q.128(i8* %ptr, <2 x i64> zeroinitializer, i8 %mask)
  %res4 = add <2 x i64> %res2, %res1
  ret <2 x i64> %res4
}

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

define <4 x i64> @test_mask_load_unaligned_q_256(i8* %ptr, i8* %ptr2, <4 x i64> %data, i8 %mask) {
; CHECK-LABEL: test_mask_load_unaligned_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqu64 (%rdi), %ymm0 ## encoding: [0x62,0xf1,0xfe,0x28,0x6f,0x07]
; CHECK-NEXT:    kmovw %edx, %k1 ## encoding: [0xc5,0xf8,0x92,0xca]
; CHECK-NEXT:    vmovdqu64 (%rsi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0xfe,0x29,0x6f,0x06]
; CHECK-NEXT:    vmovdqu64 (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0xfe,0xa9,0x6f,0x0f]
; 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.loadu.q.256(i8* %ptr, <4 x i64> zeroinitializer, i8 -1)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.loadu.q.256(i8* %ptr2, <4 x i64> %res, i8 %mask)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.loadu.q.256(i8* %ptr, <4 x i64> zeroinitializer, i8 %mask)
  %res4 = add <4 x i64> %res2, %res1
  ret <4 x i64> %res4
}

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

define <4 x i32> @test_mask_load_aligned_d_128(<4 x i32> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_aligned_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqa32 (%rdi), %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0x6f,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovdqa32 (%rdi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x6f,0x07]
; CHECK-NEXT:    vmovdqa32 (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x6f,0x0f]
; 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.load.d.128(i8* %ptr, <4 x i32> zeroinitializer, i8 -1)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.load.d.128(i8* %ptr, <4 x i32> %res, i8 %mask)
  %res2 = call <4 x i32> @llvm.x86.avx512.mask.load.d.128(i8* %ptr, <4 x i32> zeroinitializer, i8 %mask)
  %res4 = add <4 x i32> %res2, %res1
  ret <4 x i32> %res4
}

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

define <8 x i32> @test_mask_load_aligned_d_256(<8 x i32> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_aligned_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqa32 (%rdi), %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0x6f,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovdqa32 (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x6f,0x07]
; CHECK-NEXT:    vmovdqa32 (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x6f,0x0f]
; 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.load.d.256(i8* %ptr, <8 x i32> zeroinitializer, i8 -1)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.load.d.256(i8* %ptr, <8 x i32> %res, i8 %mask)
  %res2 = call <8 x i32> @llvm.x86.avx512.mask.load.d.256(i8* %ptr, <8 x i32> zeroinitializer, i8 %mask)
  %res4 = add <8 x i32> %res2, %res1
  ret <8 x i32> %res4
}

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

define <2 x i64> @test_mask_load_aligned_q_128(<2 x i64> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_aligned_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqa64 (%rdi), %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0x6f,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovdqa64 (%rdi), %xmm0 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x6f,0x07]
; CHECK-NEXT:    vmovdqa64 (%rdi), %xmm1 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0x6f,0x0f]
; 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.load.q.128(i8* %ptr, <2 x i64> zeroinitializer, i8 -1)
  %res1 = call <2 x i64> @llvm.x86.avx512.mask.load.q.128(i8* %ptr, <2 x i64> %res, i8 %mask)
  %res2 = call <2 x i64> @llvm.x86.avx512.mask.load.q.128(i8* %ptr, <2 x i64> zeroinitializer, i8 %mask)
  %res4 = add <2 x i64> %res2, %res1
  ret <2 x i64> %res4
}

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

define <4 x i64> @test_mask_load_aligned_q_256(<4 x i64> %data, i8* %ptr, i8 %mask) {
; CHECK-LABEL: test_mask_load_aligned_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vmovdqa64 (%rdi), %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0x6f,0x07]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vmovdqa64 (%rdi), %ymm0 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x6f,0x07]
; CHECK-NEXT:    vmovdqa64 (%rdi), %ymm1 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0x6f,0x0f]
; 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.load.q.256(i8* %ptr, <4 x i64> zeroinitializer, i8 -1)
  %res1 = call <4 x i64> @llvm.x86.avx512.mask.load.q.256(i8* %ptr, <4 x i64> %res, i8 %mask)
  %res2 = call <4 x i64> @llvm.x86.avx512.mask.load.q.256(i8* %ptr, <4 x i64> zeroinitializer, i8 %mask)
  %res4 = add <4 x i64> %res2, %res1
  ret <4 x i64> %res4
}

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

define <4 x i32>@test_int_x86_avx512_mask_pshuf_d_128(<4 x i32> %x0, i32 %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pshuf_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpshufd $3, %xmm0, %xmm2 ## encoding: [0x62,0xf1,0x7d,0x08,0x70,0xd0,0x03]
; CHECK-NEXT:    ## xmm2 = xmm0[3,0,0,0]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpshufd $3, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x70,0xc8,0x03]
; CHECK-NEXT:    ## xmm1 {%k1} = xmm0[3,0,0,0]
; CHECK-NEXT:    vpshufd $3, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0x70,0xc0,0x03]
; CHECK-NEXT:    ## xmm0 {%k1} {z} = xmm0[3,0,0,0]
; CHECK-NEXT:    vpaddd %xmm0, %xmm1, %xmm0 ## encoding: [0x62,0xf1,0x75,0x08,0xfe,0xc0]
; CHECK-NEXT:    vpaddd %xmm2, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xfe,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
	%res = call <4 x i32> @llvm.x86.avx512.mask.pshuf.d.128(<4 x i32> %x0, i32 3, <4 x i32> %x2, i8 %x3)
	%res1 = call <4 x i32> @llvm.x86.avx512.mask.pshuf.d.128(<4 x i32> %x0, i32 3, <4 x i32> zeroinitializer, i8 %x3)
	%res2 = call <4 x i32> @llvm.x86.avx512.mask.pshuf.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.pshuf.d.256(<8 x i32>, i32, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_pshuf_d_256(<8 x i32> %x0, i32 %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_pshuf_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpshufd $3, %ymm0, %ymm2 ## encoding: [0x62,0xf1,0x7d,0x28,0x70,0xd0,0x03]
; CHECK-NEXT:    ## ymm2 = ymm0[3,0,0,0,7,4,4,4]
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpshufd $3, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x70,0xc8,0x03]
; CHECK-NEXT:    ## ymm1 {%k1} = ymm0[3,0,0,0,7,4,4,4]
; CHECK-NEXT:    vpshufd $3, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0x70,0xc0,0x03]
; CHECK-NEXT:    ## ymm0 {%k1} {z} = ymm0[3,0,0,0,7,4,4,4]
; CHECK-NEXT:    vpaddd %ymm0, %ymm1, %ymm0 ## encoding: [0x62,0xf1,0x75,0x28,0xfe,0xc0]
; CHECK-NEXT:    vpaddd %ymm2, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xfe,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
	%res = call <8 x i32> @llvm.x86.avx512.mask.pshuf.d.256(<8 x i32> %x0, i32 3, <8 x i32> %x2, i8 %x3)
	%res1 = call <8 x i32> @llvm.x86.avx512.mask.pshuf.d.256(<8 x i32> %x0, i32 3, <8 x i32> zeroinitializer, i8 %x3)
	%res2 = call <8 x i32> @llvm.x86.avx512.mask.pshuf.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
}

define i8 @test_pcmpeq_d_256(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: test_pcmpeq_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpeqd %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x28,0x76,0xc1]
; 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.pcmpeq.d.256(<8 x i32> %a, <8 x i32> %b, i8 -1)
  ret i8 %res
}

define i8 @test_mask_pcmpeq_d_256(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_pcmpeq_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpeqd %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x76,0xc1]
; 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.pcmpeq.d.256(<8 x i32> %a, <8 x i32> %b, i8 %mask)
  ret i8 %res
}

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

define i8 @test_pcmpeq_q_256(<4 x i64> %a, <4 x i64> %b) {
; CHECK-LABEL: test_pcmpeq_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpeqq %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf2,0xfd,0x28,0x29,0xc1]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpeq.q.256(<4 x i64> %a, <4 x i64> %b, i8 -1)
  ret i8 %res
}

define i8 @test_mask_pcmpeq_q_256(<4 x i64> %a, <4 x i64> %b, i8 %mask) {
; CHECK-LABEL: test_mask_pcmpeq_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpeqq %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x29,0xc1]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpeq.q.256(<4 x i64> %a, <4 x i64> %b, i8 %mask)
  ret i8 %res
}

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

define i8 @test_pcmpgt_d_256(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: test_pcmpgt_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpgtd %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x28,0x66,0xc1]
; 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.pcmpgt.d.256(<8 x i32> %a, <8 x i32> %b, i8 -1)
  ret i8 %res
}

define i8 @test_mask_pcmpgt_d_256(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_pcmpgt_d_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpgtd %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x66,0xc1]
; 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.pcmpgt.d.256(<8 x i32> %a, <8 x i32> %b, i8 %mask)
  ret i8 %res
}

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

define i8 @test_pcmpgt_q_256(<4 x i64> %a, <4 x i64> %b) {
; CHECK-LABEL: test_pcmpgt_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpgtq %ymm1, %ymm0, %k0 ## encoding: [0x62,0xf2,0xfd,0x28,0x37,0xc1]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpgt.q.256(<4 x i64> %a, <4 x i64> %b, i8 -1)
  ret i8 %res
}

define i8 @test_mask_pcmpgt_q_256(<4 x i64> %a, <4 x i64> %b, i8 %mask) {
; CHECK-LABEL: test_mask_pcmpgt_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpgtq %ymm1, %ymm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x29,0x37,0xc1]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpgt.q.256(<4 x i64> %a, <4 x i64> %b, i8 %mask)
  ret i8 %res
}

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

define i8 @test_pcmpeq_d_128(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_pcmpeq_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpeqd %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x08,0x76,0xc1]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpeq.d.128(<4 x i32> %a, <4 x i32> %b, i8 -1)
  ret i8 %res
}

define i8 @test_mask_pcmpeq_d_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_pcmpeq_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpeqd %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x76,0xc1]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpeq.d.128(<4 x i32> %a, <4 x i32> %b, i8 %mask)
  ret i8 %res
}

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

define i8 @test_pcmpeq_q_128(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: test_pcmpeq_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpeqq %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf2,0xfd,0x08,0x29,0xc1]
; CHECK-NEXT:    kshiftlw $14, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0e]
; CHECK-NEXT:    kshiftrw $14, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0e]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpeq.q.128(<2 x i64> %a, <2 x i64> %b, i8 -1)
  ret i8 %res
}

define i8 @test_mask_pcmpeq_q_128(<2 x i64> %a, <2 x i64> %b, i8 %mask) {
; CHECK-LABEL: test_mask_pcmpeq_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpeqq %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x29,0xc1]
; CHECK-NEXT:    kshiftlw $14, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0e]
; CHECK-NEXT:    kshiftrw $14, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0e]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpeq.q.128(<2 x i64> %a, <2 x i64> %b, i8 %mask)
  ret i8 %res
}

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

define i8 @test_pcmpgt_d_128(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_pcmpgt_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpgtd %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf1,0x7d,0x08,0x66,0xc1]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpgt.d.128(<4 x i32> %a, <4 x i32> %b, i8 -1)
  ret i8 %res
}

define i8 @test_mask_pcmpgt_d_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_pcmpgt_d_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpgtd %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x66,0xc1]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpgt.d.128(<4 x i32> %a, <4 x i32> %b, i8 %mask)
  ret i8 %res
}

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

define i8 @test_pcmpgt_q_128(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: test_pcmpgt_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpcmpgtq %xmm1, %xmm0, %k0 ## encoding: [0x62,0xf2,0xfd,0x08,0x37,0xc1]
; CHECK-NEXT:    kshiftlw $14, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0e]
; CHECK-NEXT:    kshiftrw $14, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0e]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpgt.q.128(<2 x i64> %a, <2 x i64> %b, i8 -1)
  ret i8 %res
}

define i8 @test_mask_pcmpgt_q_128(<2 x i64> %a, <2 x i64> %b, i8 %mask) {
; CHECK-LABEL: test_mask_pcmpgt_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpcmpgtq %xmm1, %xmm0, %k0 {%k1} ## encoding: [0x62,0xf2,0xfd,0x09,0x37,0xc1]
; CHECK-NEXT:    kshiftlw $14, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0e]
; CHECK-NEXT:    kshiftrw $14, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0e]
; CHECK-NEXT:    kshiftlw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x32,0xc0,0x0c]
; CHECK-NEXT:    kshiftrw $12, %k0, %k0 ## encoding: [0xc4,0xe3,0xf9,0x30,0xc0,0x0c]
; 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.pcmpgt.q.128(<2 x i64> %a, <2 x i64> %b, i8 %mask)
  ret i8 %res
}

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

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

define <2 x double>@test_int_x86_avx512_mask_unpckh_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_unpckh_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vunpckhpd %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x15,0xd9]
; CHECK-NEXT:    ## xmm3 = xmm0[1],xmm1[1]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vunpckhpd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x15,0xd1]
; CHECK-NEXT:    ## xmm2 {%k1} = xmm0[1],xmm1[1]
; CHECK-NEXT:    vaddpd %xmm3, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0x58,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.mask.unpckh.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.unpckh.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.unpckh.pd.256(<4 x double>, <4 x double>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_unpckh_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_unpckh_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vunpckhpd %ymm1, %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x15,0xd9]
; CHECK-NEXT:    ## ymm3 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vunpckhpd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x15,0xd1]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
; CHECK-NEXT:    vaddpd %ymm3, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0x58,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.unpckh.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.unpckh.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.unpckh.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_unpckh_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_unpckh_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vunpckhps %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x15,0xd9]
; CHECK-NEXT:    ## xmm3 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vunpckhps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x15,0xd1]
; CHECK-NEXT:    ## xmm2 {%k1} = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
; CHECK-NEXT:    vaddps %xmm3, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6c,0x08,0x58,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.unpckh.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.unpckh.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.unpckh.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_unpckh_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_unpckh_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vunpckhps %ymm1, %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x15,0xd9]
; CHECK-NEXT:    ## ymm3 = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vunpckhps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x15,0xd1]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
; CHECK-NEXT:    vaddps %ymm3, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.unpckh.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.unpckh.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 <2 x double> @llvm.x86.avx512.mask.unpckl.pd.128(<2 x double>, <2 x double>, <2 x double>, i8)

define <2 x double>@test_int_x86_avx512_mask_unpckl_pd_128(<2 x double> %x0, <2 x double> %x1, <2 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_unpckl_pd_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vunpcklpd %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x14,0xd9]
; CHECK-NEXT:    ## xmm3 = xmm0[0],xmm1[0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vunpcklpd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x14,0xd1]
; CHECK-NEXT:    ## xmm2 {%k1} = xmm0[0],xmm1[0]
; CHECK-NEXT:    vaddpd %xmm3, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0x58,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.avx512.mask.unpckl.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.unpckl.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.unpckl.pd.256(<4 x double>, <4 x double>, <4 x double>, i8)

define <4 x double>@test_int_x86_avx512_mask_unpckl_pd_256(<4 x double> %x0, <4 x double> %x1, <4 x double> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_unpckl_pd_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vunpcklpd %ymm1, %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x14,0xd9]
; CHECK-NEXT:    ## ymm3 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vunpcklpd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x14,0xd1]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
; CHECK-NEXT:    vaddpd %ymm3, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0x58,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.avx512.mask.unpckl.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.unpckl.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.unpckl.ps.128(<4 x float>, <4 x float>, <4 x float>, i8)

define <4 x float>@test_int_x86_avx512_mask_unpckl_ps_128(<4 x float> %x0, <4 x float> %x1, <4 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_unpckl_ps_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vunpcklps %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7c,0x08,0x14,0xd9]
; CHECK-NEXT:    ## xmm3 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vunpcklps %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x09,0x14,0xd1]
; CHECK-NEXT:    ## xmm2 {%k1} = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; CHECK-NEXT:    vaddps %xmm3, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6c,0x08,0x58,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.avx512.mask.unpckl.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.unpckl.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.unpckl.ps.256(<8 x float>, <8 x float>, <8 x float>, i8)

define <8 x float>@test_int_x86_avx512_mask_unpckl_ps_256(<8 x float> %x0, <8 x float> %x1, <8 x float> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_unpckl_ps_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vunpcklps %ymm1, %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7c,0x28,0x14,0xd9]
; CHECK-NEXT:    ## ymm3 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vunpcklps %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7c,0x29,0x14,0xd1]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
; CHECK-NEXT:    vaddps %ymm3, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6c,0x28,0x58,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.avx512.mask.unpckl.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.unpckl.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 <4 x i32> @llvm.x86.avx512.mask.punpckhd.q.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_punpckhd_q_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckhd_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpunpckhdq %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7d,0x08,0x6a,0xd9]
; CHECK-NEXT:    ## xmm3 = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpunpckhdq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x6a,0xd1]
; CHECK-NEXT:    ## xmm2 {%k1} = xmm0[2],xmm1[2],xmm0[3],xmm1[3]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.punpckhd.q.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.punpckhd.q.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.mask.punpckld.q.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <4 x i32>@test_int_x86_avx512_mask_punpckld_q_128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckld_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpunpckldq %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf1,0x7d,0x08,0x62,0xd9]
; CHECK-NEXT:    ## xmm3 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpunpckldq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0x62,0xd1]
; CHECK-NEXT:    ## xmm2 {%k1} = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; CHECK-NEXT:    vpaddd %xmm3, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0x6d,0x08,0xfe,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.punpckld.q.128(<4 x i32> %x0, <4 x i32> %x1, <4 x i32> %x2, i8 %x3)
  %res1 = call <4 x i32> @llvm.x86.avx512.mask.punpckld.q.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.punpckhd.q.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_punpckhd_q_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckhd_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpunpckhdq %ymm1, %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7d,0x28,0x6a,0xd9]
; CHECK-NEXT:    ## ymm3 = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpunpckhdq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x6a,0xd1]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[6],ymm1[6],ymm0[7],ymm1[7]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.punpckhd.q.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.punpckhd.q.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.mask.punpckld.q.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <8 x i32>@test_int_x86_avx512_mask_punpckld_q_256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckld_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpunpckldq %ymm1, %ymm0, %ymm3 ## encoding: [0x62,0xf1,0x7d,0x28,0x62,0xd9]
; CHECK-NEXT:    ## ymm3 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpunpckldq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0x62,0xd1]
; CHECK-NEXT:    ## ymm2 {%k1} = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[4],ymm1[4],ymm0[5],ymm1[5]
; CHECK-NEXT:    vpaddd %ymm3, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0x6d,0x28,0xfe,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.punpckld.q.256(<8 x i32> %x0, <8 x i32> %x1, <8 x i32> %x2, i8 %x3)
  %res1 = call <8 x i32> @llvm.x86.avx512.mask.punpckld.q.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.punpckhqd.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_punpckhqd_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckhqd_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpunpckhqdq %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x6d,0xd9]
; CHECK-NEXT:    ## xmm3 = xmm0[1],xmm1[1]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpunpckhqdq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x6d,0xd1]
; CHECK-NEXT:    ## xmm2 = xmm0[1],xmm1[1]
; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.punpckhqd.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.punpckhqd.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 <2 x i64> @llvm.x86.avx512.mask.punpcklqd.q.128(<2 x i64>, <2 x i64>, <2 x i64>, i8)

define <2 x i64>@test_int_x86_avx512_mask_punpcklqd_q_128(<2 x i64> %x0, <2 x i64> %x1, <2 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpcklqd_q_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpunpcklqdq %xmm1, %xmm0, %xmm3 ## encoding: [0x62,0xf1,0xfd,0x08,0x6c,0xd9]
; CHECK-NEXT:    ## xmm3 = xmm0[0],xmm1[0]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpunpcklqdq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0x6c,0xd1]
; CHECK-NEXT:    ## xmm2 = xmm0[0],xmm1[0]
; CHECK-NEXT:    vpaddq %xmm3, %xmm2, %xmm0 ## encoding: [0x62,0xf1,0xed,0x08,0xd4,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.punpcklqd.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.punpcklqd.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.punpcklqd.q.256(<4 x i64>, <4 x i64>, <4 x i64>, i8)

define <4 x i64>@test_int_x86_avx512_mask_punpcklqd_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpcklqd_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpunpcklqdq %ymm1, %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x6c,0xd9]
; CHECK-NEXT:    ## ymm3 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpunpcklqdq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x6c,0xd1]
; CHECK-NEXT:    ## ymm2 = ymm0[0],ymm1[0],ymm0[2],ymm1[2]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.punpcklqd.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.punpcklqd.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

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

define <4 x i64>@test_int_x86_avx512_mask_punpckhqd_q_256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 %x3) {
; CHECK-LABEL: test_int_x86_avx512_mask_punpckhqd_q_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpunpckhqdq %ymm1, %ymm0, %ymm3 ## encoding: [0x62,0xf1,0xfd,0x28,0x6d,0xd9]
; CHECK-NEXT:    ## ymm3 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpunpckhqdq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0x6d,0xd1]
; CHECK-NEXT:    ## ymm2 = ymm0[1],ymm1[1],ymm0[3],ymm1[3]
; CHECK-NEXT:    vpaddq %ymm3, %ymm2, %ymm0 ## encoding: [0x62,0xf1,0xed,0x28,0xd4,0xc3]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.punpckhqd.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.punpckhqd.q.256(<4 x i64> %x0, <4 x i64> %x1, <4 x i64> %x2, i8 -1)
  %res2 = add <4 x i64> %res, %res1
  ret <4 x i64> %res2
}

define <4 x i32> @test_mask_and_epi32_rr_128(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_mask_and_epi32_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xdb,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rrk_128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xdb,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rrkz_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandd %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xdb,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rm_128(<4 x i32> %a, <4 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_and_epi32_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandd (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xdb,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.pand.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_and_epi32_rmk_128(<4 x i32> %a, <4 x i32>* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandd (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xdb,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rmkz_128(<4 x i32> %a, <4 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandd (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xdb,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.pand.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_and_epi32_rmb_128(<4 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_and_epi32_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandd (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x18,0xdb,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.pand.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_and_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandd (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x19,0xdb,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i32, i32* %ptr_b
  %vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
  %b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
  %res = call <4 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandd (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0xdb,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.pand.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.pand.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <8 x i32> @test_mask_and_epi32_rr_256(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: test_mask_and_epi32_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xdb,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rrk_256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xdb,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rrkz_256(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandd %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xdb,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rm_256(<8 x i32> %a, <8 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_and_epi32_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandd (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xdb,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.pand.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_and_epi32_rmk_256(<8 x i32> %a, <8 x i32>* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandd (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xdb,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rmkz_256(<8 x i32> %a, <8 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandd (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xdb,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.pand.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_and_epi32_rmb_256(<8 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_and_epi32_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandd (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x38,0xdb,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.pand.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_and_epi32_rmbk_256(<8 x i32> %a, i32* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandd (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x39,0xdb,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i32, i32* %ptr_b
  %vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
  %b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
  %res = call <8 x i32> @llvm.x86.avx512.mask.pand.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_and_epi32_rmbkz_256(<8 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_and_epi32_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandd (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xb9,0xdb,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.pand.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.pand.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <4 x i32> @test_mask_or_epi32_rr_128(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_mask_or_epi32_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpord %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xeb,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rrk_128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpord %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xeb,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rrkz_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpord %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xeb,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rm_128(<4 x i32> %a, <4 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_or_epi32_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpord (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xeb,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.por.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_or_epi32_rmk_128(<4 x i32> %a, <4 x i32>* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpord (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xeb,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rmkz_128(<4 x i32> %a, <4 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpord (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xeb,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.por.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_or_epi32_rmb_128(<4 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_or_epi32_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpord (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x18,0xeb,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.por.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_or_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpord (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x19,0xeb,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i32, i32* %ptr_b
  %vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
  %b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
  %res = call <4 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpord (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0xeb,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.por.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.por.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <8 x i32> @test_mask_or_epi32_rr_256(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: test_mask_or_epi32_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpord %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xeb,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rrk_256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpord %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xeb,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rrkz_256(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpord %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xeb,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rm_256(<8 x i32> %a, <8 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_or_epi32_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpord (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xeb,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.por.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_or_epi32_rmk_256(<8 x i32> %a, <8 x i32>* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpord (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xeb,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rmkz_256(<8 x i32> %a, <8 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpord (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xeb,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.por.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_or_epi32_rmb_256(<8 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_or_epi32_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpord (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x38,0xeb,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.por.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_or_epi32_rmbk_256(<8 x i32> %a, i32* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpord (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x39,0xeb,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i32, i32* %ptr_b
  %vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
  %b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
  %res = call <8 x i32> @llvm.x86.avx512.mask.por.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_or_epi32_rmbkz_256(<8 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_or_epi32_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpord (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xb9,0xeb,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.por.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.por.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <4 x i32> @test_mask_xor_epi32_rr_128(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_mask_xor_epi32_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpxord %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xef,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rrk_128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpxord %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xef,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rrkz_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpxord %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xef,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rm_128(<4 x i32> %a, <4 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_xor_epi32_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpxord (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xef,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.pxor.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_xor_epi32_rmk_128(<4 x i32> %a, <4 x i32>* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpxord (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xef,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rmkz_128(<4 x i32> %a, <4 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpxord (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xef,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.pxor.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_xor_epi32_rmb_128(<4 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_xor_epi32_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpxord (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x18,0xef,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.pxor.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_xor_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpxord (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x19,0xef,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i32, i32* %ptr_b
  %vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
  %b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
  %res = call <4 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpxord (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0xef,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.pxor.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.pxor.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <8 x i32> @test_mask_xor_epi32_rr_256(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: test_mask_xor_epi32_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpxord %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xef,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rrk_256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpxord %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xef,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rrkz_256(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpxord %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xef,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rm_256(<8 x i32> %a, <8 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_xor_epi32_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpxord (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xef,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.pxor.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_xor_epi32_rmk_256(<8 x i32> %a, <8 x i32>* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpxord (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xef,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rmkz_256(<8 x i32> %a, <8 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpxord (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xef,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.pxor.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_xor_epi32_rmb_256(<8 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_xor_epi32_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpxord (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x38,0xef,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.pxor.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_xor_epi32_rmbk_256(<8 x i32> %a, i32* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpxord (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x39,0xef,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i32, i32* %ptr_b
  %vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
  %b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
  %res = call <8 x i32> @llvm.x86.avx512.mask.pxor.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_xor_epi32_rmbkz_256(<8 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_xor_epi32_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpxord (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xb9,0xef,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.pxor.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.pxor.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <4 x i32> @test_mask_andnot_epi32_rr_128(<4 x i32> %a, <4 x i32> %b) {
; CHECK-LABEL: test_mask_andnot_epi32_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnd %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xdf,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rrk_128(<4 x i32> %a, <4 x i32> %b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandnd %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xdf,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rrkz_128(<4 x i32> %a, <4 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandnd %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xdf,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rm_128(<4 x i32> %a, <4 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_andnot_epi32_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnd (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x08,0xdf,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.pandn.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_andnot_epi32_rmk_128(<4 x i32> %a, <4 x i32>* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnd (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x09,0xdf,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i32>, <4 x i32>* %ptr_b
  %res = call <4 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rmkz_128(<4 x i32> %a, <4 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnd (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x89,0xdf,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.pandn.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_andnot_epi32_rmb_128(<4 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_andnot_epi32_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnd (%rdi){1to4}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0x7d,0x18,0xdf,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.pandn.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_andnot_epi32_rmbk_128(<4 x i32> %a, i32* %ptr_b, <4 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnd (%rdi){1to4}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x19,0xdf,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i32, i32* %ptr_b
  %vecinit.i = insertelement <4 x i32> undef, i32 %q, i32 0
  %b = shufflevector <4 x i32> %vecinit.i, <4 x i32> undef, <4 x i32> zeroinitializer
  %res = call <4 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rmbkz_128(<4 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnd (%rdi){1to4}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0x99,0xdf,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.pandn.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.pandn.d.128(<4 x i32>, <4 x i32>, <4 x i32>, i8)

define <8 x i32> @test_mask_andnot_epi32_rr_256(<8 x i32> %a, <8 x i32> %b) {
; CHECK-LABEL: test_mask_andnot_epi32_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnd %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xdf,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rrk_256(<8 x i32> %a, <8 x i32> %b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandnd %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xdf,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rrkz_256(<8 x i32> %a, <8 x i32> %b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandnd %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xdf,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <8 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rm_256(<8 x i32> %a, <8 x i32>* %ptr_b) {
; CHECK-LABEL: test_mask_andnot_epi32_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnd (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x28,0xdf,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.pandn.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_andnot_epi32_rmk_256(<8 x i32> %a, <8 x i32>* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnd (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x29,0xdf,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <8 x i32>, <8 x i32>* %ptr_b
  %res = call <8 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rmkz_256(<8 x i32> %a, <8 x i32>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnd (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xa9,0xdf,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.pandn.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_andnot_epi32_rmb_256(<8 x i32> %a, i32* %ptr_b) {
; CHECK-LABEL: test_mask_andnot_epi32_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnd (%rdi){1to8}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0x7d,0x38,0xdf,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.pandn.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_andnot_epi32_rmbk_256(<8 x i32> %a, i32* %ptr_b, <8 x i32> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnd (%rdi){1to8}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0x7d,0x39,0xdf,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i32, i32* %ptr_b
  %vecinit.i = insertelement <8 x i32> undef, i32 %q, i32 0
  %b = shufflevector <8 x i32> %vecinit.i, <8 x i32> undef, <8 x i32> zeroinitializer
  %res = call <8 x i32> @llvm.x86.avx512.mask.pandn.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_andnot_epi32_rmbkz_256(<8 x i32> %a, i32* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi32_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnd (%rdi){1to8}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0x7d,0xb9,0xdf,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.pandn.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.pandn.d.256(<8 x i32>, <8 x i32>, <8 x i32>, i8)

define <2 x i64> @test_mask_andnot_epi64_rr_128(<2 x i64> %a, <2 x i64> %b) {
; CHECK-LABEL: test_mask_andnot_epi64_rr_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnq %xmm1, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xdf,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_mask_andnot_epi64_rrk_128(<2 x i64> %a, <2 x i64> %b, <2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rrk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandnq %xmm1, %xmm0, %xmm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0xdf,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> %passThru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_mask_andnot_epi64_rrkz_128(<2 x i64> %a, <2 x i64> %b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rrkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandnq %xmm1, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0xdf,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_mask_andnot_epi64_rm_128(<2 x i64> %a, <2 x i64>* %ptr_b) {
; CHECK-LABEL: test_mask_andnot_epi64_rm_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnq (%rdi), %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x08,0xdf,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <2 x i64>, <2 x i64>* %ptr_b
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_mask_andnot_epi64_rmk_128(<2 x i64> %a, <2 x i64>* %ptr_b, <2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rmk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnq (%rdi), %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x09,0xdf,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <2 x i64>, <2 x i64>* %ptr_b
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> %passThru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_mask_andnot_epi64_rmkz_128(<2 x i64> %a, <2 x i64>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rmkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnq (%rdi), %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x89,0xdf,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <2 x i64>, <2 x i64>* %ptr_b
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_mask_andnot_epi64_rmb_128(<2 x i64> %a, i64* %ptr_b) {
; CHECK-LABEL: test_mask_andnot_epi64_rmb_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnq (%rdi){1to2}, %xmm0, %xmm0 ## encoding: [0x62,0xf1,0xfd,0x18,0xdf,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement <2 x i64> undef, i64 %q, i32 0
  %b = shufflevector <2 x i64> %vecinit.i, <2 x i64> undef, <2 x i32> zeroinitializer
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 -1)
  ret <2 x i64> %res
}

define <2 x i64> @test_mask_andnot_epi64_rmbk_128(<2 x i64> %a, i64* %ptr_b, <2 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rmbk_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnq (%rdi){1to2}, %xmm0, %xmm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x19,0xdf,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement <2 x i64> undef, i64 %q, i32 0
  %b = shufflevector <2 x i64> %vecinit.i, <2 x i64> undef, <2 x i32> zeroinitializer
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> %passThru, i8 %mask)
  ret <2 x i64> %res
}

define <2 x i64> @test_mask_andnot_epi64_rmbkz_128(<2 x i64> %a, i64* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rmbkz_128:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnq (%rdi){1to2}, %xmm0, %xmm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0x99,0xdf,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement <2 x i64> undef, i64 %q, i32 0
  %b = shufflevector <2 x i64> %vecinit.i, <2 x i64> undef, <2 x i32> zeroinitializer
  %res = call <2 x i64> @llvm.x86.avx512.mask.pandn.q.128(<2 x i64> %a, <2 x i64> %b, <2 x i64> zeroinitializer, i8 %mask)
  ret <2 x i64> %res
}

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

define <4 x i64> @test_mask_andnot_epi64_rr_256(<4 x i64> %a, <4 x i64> %b) {
; CHECK-LABEL: test_mask_andnot_epi64_rr_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnq %ymm1, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xdf,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 -1)
  ret <4 x i64> %res
}

define <4 x i64> @test_mask_andnot_epi64_rrk_256(<4 x i64> %a, <4 x i64> %b, <4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rrk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandnq %ymm1, %ymm0, %ymm2 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xdf,0xd1]
; CHECK-NEXT:    vmovaps %zmm2, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc2]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> %passThru, i8 %mask)
  ret <4 x i64> %res
}

define <4 x i64> @test_mask_andnot_epi64_rrkz_256(<4 x i64> %a, <4 x i64> %b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rrkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %edi, %k1 ## encoding: [0xc5,0xf8,0x92,0xcf]
; CHECK-NEXT:    vpandnq %ymm1, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0xdf,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 %mask)
  ret <4 x i64> %res
}

define <4 x i64> @test_mask_andnot_epi64_rm_256(<4 x i64> %a, <4 x i64>* %ptr_b) {
; CHECK-LABEL: test_mask_andnot_epi64_rm_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnq (%rdi), %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x28,0xdf,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i64>, <4 x i64>* %ptr_b
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 -1)
  ret <4 x i64> %res
}

define <4 x i64> @test_mask_andnot_epi64_rmk_256(<4 x i64> %a, <4 x i64>* %ptr_b, <4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rmk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnq (%rdi), %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x29,0xdf,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i64>, <4 x i64>* %ptr_b
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> %passThru, i8 %mask)
  ret <4 x i64> %res
}

define <4 x i64> @test_mask_andnot_epi64_rmkz_256(<4 x i64> %a, <4 x i64>* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rmkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnq (%rdi), %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xa9,0xdf,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %b = load <4 x i64>, <4 x i64>* %ptr_b
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 %mask)
  ret <4 x i64> %res
}

define <4 x i64> @test_mask_andnot_epi64_rmb_256(<4 x i64> %a, i64* %ptr_b) {
; CHECK-LABEL: test_mask_andnot_epi64_rmb_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    vpandnq (%rdi){1to4}, %ymm0, %ymm0 ## encoding: [0x62,0xf1,0xfd,0x38,0xdf,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement <4 x i64> undef, i64 %q, i32 0
  %b = shufflevector <4 x i64> %vecinit.i, <4 x i64> undef, <4 x i32> zeroinitializer
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 -1)
  ret <4 x i64> %res
}

define <4 x i64> @test_mask_andnot_epi64_rmbk_256(<4 x i64> %a, i64* %ptr_b, <4 x i64> %passThru, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rmbk_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnq (%rdi){1to4}, %ymm0, %ymm1 {%k1} ## encoding: [0x62,0xf1,0xfd,0x39,0xdf,0x0f]
; CHECK-NEXT:    vmovaps %zmm1, %zmm0 ## encoding: [0x62,0xf1,0x7c,0x48,0x28,0xc1]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement <4 x i64> undef, i64 %q, i32 0
  %b = shufflevector <4 x i64> %vecinit.i, <4 x i64> undef, <4 x i32> zeroinitializer
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> %passThru, i8 %mask)
  ret <4 x i64> %res
}

define <4 x i64> @test_mask_andnot_epi64_rmbkz_256(<4 x i64> %a, i64* %ptr_b, i8 %mask) {
; CHECK-LABEL: test_mask_andnot_epi64_rmbkz_256:
; CHECK:       ## BB#0:
; CHECK-NEXT:    kmovw %esi, %k1 ## encoding: [0xc5,0xf8,0x92,0xce]
; CHECK-NEXT:    vpandnq (%rdi){1to4}, %ymm0, %ymm0 {%k1} {z} ## encoding: [0x62,0xf1,0xfd,0xb9,0xdf,0x07]
; CHECK-NEXT:    retq ## encoding: [0xc3]
  %q = load i64, i64* %ptr_b
  %vecinit.i = insertelement <4 x i64> undef, i64 %q, i32 0
  %b = shufflevector <4 x i64> %vecinit.i, <4 x i64> undef, <4 x i32> zeroinitializer
  %res = call <4 x i64> @llvm.x86.avx512.mask.pandn.q.256(<4 x i64> %a, <4 x i64> %b, <4 x i64> zeroinitializer, i8 %mask)
  ret <4 x i64> %res
}

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