llvm/test/CodeGen/X86/vector-shuffle-combining-avx2.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,X86,AVX2
; RUN: llc < %s -mtriple=i686-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,X86,AVX512
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx2 | FileCheck %s --check-prefixes=CHECK,X64,AVX2
; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+avx512f | FileCheck %s --check-prefixes=CHECK,X64,AVX512

declare <8 x i32> @llvm.x86.avx2.permd(<8 x i32>, <8 x i32>)
declare <8 x float> @llvm.x86.avx2.permps(<8 x float>, <8 x i32>)
declare <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8>, <16 x i8>)
declare <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8>, <32 x i8>)

define <32 x i8> @combine_pshufb_pslldq(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_pslldq:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vxorps %xmm0, %xmm0, %xmm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7>)
  %2 = shufflevector <32 x i8> %1, <32 x i8> zeroinitializer, <32 x i32> <i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23>
  ret <32 x i8> %2
}

define <32 x i8> @combine_pshufb_psrldq(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_psrldq:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vxorps %xmm0, %xmm0, %xmm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128>)
  %2 = shufflevector <32 x i8> %1, <32 x i8> zeroinitializer, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32, i32 32>
  ret <32 x i8> %2
}

define <32 x i8> @combine_pshufb_vpermd(<8 x i32> %a) {
; CHECK-LABEL: combine_pshufb_vpermd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,16,17,18,18]
; CHECK-NEXT:    ret{{[l|q]}}
  %tmp0 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 4>)
  %tmp1 = bitcast <8 x i32> %tmp0 to <32 x i8>
  %tmp2 = shufflevector <32 x i8> %tmp1, <32 x i8> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 30>
  ret <32 x i8> %tmp2
}

define <32 x i8> @combine_pshufb_vpermps(<8 x float> %a) {
; CHECK-LABEL: combine_pshufb_vpermps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,16,17,18,18]
; CHECK-NEXT:    ret{{[l|q]}}
  %tmp0 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 4>)
  %tmp1 = bitcast <8 x float> %tmp0 to <32 x i8>
  %tmp2 = shufflevector <32 x i8> %tmp1, <32 x i8> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 30>
  ret <32 x i8> %tmp2
}

define <32 x i8> @combine_and_pshufb(<32 x i8> %a0) {
; CHECK-LABEL: combine_and_pshufb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpxor %xmm1, %xmm1, %xmm1
; CHECK-NEXT:    vpblendw {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7],ymm0[8],ymm1[9,10,11],ymm0[12],ymm1[13,14,15]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <32 x i8> %a0, <32 x i8> zeroinitializer, <32 x i32> <i32 0, i32 1, i32 32, i32 32, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  %2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %1, <32 x i8> <i8 0, i8 1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 8, i8 9, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 8, i8 9, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  ret <32 x i8> %2
}

define <32 x i8> @combine_pshufb_and(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_and:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpxor %xmm1, %xmm1, %xmm1
; CHECK-NEXT:    vpblendw {{.*#+}} ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7],ymm0[8],ymm1[9,10,11],ymm0[12],ymm1[13,14,15]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 0, i8 1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 8, i8 9, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 8, i8 9, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  %2 = shufflevector <32 x i8> %1, <32 x i8> zeroinitializer, <32 x i32> <i32 0, i32 1, i32 32, i32 32, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 16, i32 17, i32 18, i32 19, i32 20, i32 21, i32 22, i32 23, i32 24, i32 25, i32 26, i32 27, i32 28, i32 29, i32 30, i32 31>
  ret <32 x i8> %2
}

define <4 x i64> @combine_permq_pshufb_as_vextracti128(<4 x i64> %a0) {
; X86-LABEL: combine_permq_pshufb_as_vextracti128:
; X86:       # %bb.0:
; X86-NEXT:    vextracti128 $1, %ymm0, %xmm0
; X86-NEXT:    vpaddq {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0
; X86-NEXT:    retl
;
; X64-LABEL: combine_permq_pshufb_as_vextracti128:
; X64:       # %bb.0:
; X64-NEXT:    vextracti128 $1, %ymm0, %xmm0
; X64-NEXT:    vpaddq {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
; X64-NEXT:    retq
  %1 = shufflevector <4 x i64> %a0, <4 x i64> undef, <4 x i32> <i32 3, i32 2, i32 1, i32 0>
  %2 = bitcast <4 x i64> %1 to <32 x i8>
  %3 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255>)
  %4 = bitcast <32 x i8> %3 to <4 x i64>
  %5 = add <4 x i64> %4, <i64 1, i64 1, i64 3, i64 3>
  ret <4 x i64> %5
}

define <4 x i64> @combine_permq_pshufb_as_vmovdqa(<4 x i64> %a0) {
; X86-LABEL: combine_permq_pshufb_as_vmovdqa:
; X86:       # %bb.0:
; X86-NEXT:    vmovdqa %xmm0, %xmm0
; X86-NEXT:    vpaddq {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0
; X86-NEXT:    retl
;
; X64-LABEL: combine_permq_pshufb_as_vmovdqa:
; X64:       # %bb.0:
; X64-NEXT:    vmovdqa %xmm0, %xmm0
; X64-NEXT:    vpaddq {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
; X64-NEXT:    retq
  %1 = shufflevector <4 x i64> %a0, <4 x i64> undef, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
  %2 = bitcast <4 x i64> %1 to <32 x i8>
  %3 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255>)
  %4 = bitcast <32 x i8> %3 to <4 x i64>
  %5 = add <4 x i64> %4, <i64 1, i64 1, i64 3, i64 3>
  ret <4 x i64> %5
}

define <8 x i32> @combine_as_vpermd(<8 x i32> %a0) {
; CHECK-LABEL: combine_as_vpermd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vmovaps {{.*#+}} ymm1 = [4,5,4,5,6,7,0,7]
; CHECK-NEXT:    vpermps %ymm0, %ymm1, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <8 x i32> %a0, <8 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
  %2 = tail call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> <i32 5, i32 4, i32 3, i32 2, i32 1, i32 0, i32 7, i32 6>)
  %3 = shufflevector <8 x i32> %1, <8 x i32> %2, <8 x i32> <i32 0, i32 8, i32 9, i32 1, i32 15, i32 14, i32 4, i32 3>
  ret <8 x i32> %3
}

define <8 x float> @combine_as_vpermps(<8 x float> %a0) {
; CHECK-LABEL: combine_as_vpermps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vmovaps {{.*#+}} ymm1 = <6,4,7,5,1,u,4,7>
; CHECK-NEXT:    vpermps %ymm0, %ymm1, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <8 x float> %a0, <8 x float> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 4, i32 5, i32 6, i32 7>
  %2 = tail call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a0, <8 x i32> <i32 1, i32 undef, i32 3, i32 2, i32 5, i32 4, i32 7, i32 6>)
  %3 = shufflevector <8 x float> %1, <8 x float> %2, <8 x i32> <i32 15, i32 0, i32 14, i32 1, i32 8, i32 9, i32 4, i32 3>
  ret <8 x float> %3
}

define <32 x i8> @combine_permq_pshufb_as_vmovaps(<4 x i64> %a0) {
; CHECK-LABEL: combine_permq_pshufb_as_vmovaps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vmovaps %xmm0, %xmm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <4 x i64> %a0, <4 x i64> undef, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
  %2 = bitcast <4 x i64> %1 to <32 x i8>
  %3 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255>)
  ret <32 x i8> %3
}

define <32 x i8> @combine_permq_pshufb_as_vpblendd(<4 x i64> %a0) {
; CHECK-LABEL: combine_permq_pshufb_as_vpblendd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vxorps %xmm1, %xmm1, %xmm1
; CHECK-NEXT:    vblendps {{.*#+}} ymm0 = ymm1[0,1,2,3],ymm0[4,5,6,7]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <4 x i64> %a0, <4 x i64> undef, <4 x i32> <i32 1, i32 0, i32 3, i32 2>
  %2 = bitcast <4 x i64> %1 to <32 x i8>
  %3 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 255, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7>)
  ret <32 x i8> %3
}

define <16 x i8> @combine_pshufb_as_vpbroadcastb128(<16 x i8> %a) {
; CHECK-LABEL: combine_pshufb_as_vpbroadcastb128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpbroadcastb %xmm0, %xmm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %a, <16 x i8> zeroinitializer)
  ret <16 x i8> %1
}

define <32 x i8> @combine_pshufb_as_vpbroadcastb256(<2 x i64> %a) {
; CHECK-LABEL: combine_pshufb_as_vpbroadcastb256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpbroadcastb %xmm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <2 x i64> %a, <2 x i64> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
  %2 = bitcast <4 x i64> %1 to <32 x i8>
  %3 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> zeroinitializer)
  %4 = bitcast <32 x i8> %3 to <8 x i32>
  %5 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %4, <8 x i32> zeroinitializer)
  %6 = bitcast <8 x i32> %5 to <32 x i8>
  ret <32 x i8> %6
}

define <16 x i8> @combine_pshufb_as_vpbroadcastw128(<16 x i8> %a) {
; CHECK-LABEL: combine_pshufb_as_vpbroadcastw128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpbroadcastw %xmm0, %xmm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %a, <16 x i8> <i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1>)
  ret <16 x i8> %1
}

define <32 x i8> @combine_pshufb_as_vpbroadcastw256(<2 x i64> %a) {
; CHECK-LABEL: combine_pshufb_as_vpbroadcastw256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpbroadcastw %xmm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <2 x i64> %a, <2 x i64> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
  %2 = bitcast <4 x i64> %1 to <32 x i8>
  %3 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1, i8 0, i8 1>)
  %4 = bitcast <32 x i8> %3 to <8 x i32>
  %5 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %4, <8 x i32> zeroinitializer)
  %6 = bitcast <8 x i32> %5 to <32 x i8>
  ret <32 x i8> %6
}

define <16 x i8> @combine_pshufb_as_vpbroadcastd128(<16 x i8> %a) {
; X86-LABEL: combine_pshufb_as_vpbroadcastd128:
; X86:       # %bb.0:
; X86-NEXT:    vpbroadcastd %xmm0, %xmm0
; X86-NEXT:    vpaddb {{\.?LCPI[0-9]+_[0-9]+}}, %xmm0, %xmm0
; X86-NEXT:    retl
;
; X64-LABEL: combine_pshufb_as_vpbroadcastd128:
; X64:       # %bb.0:
; X64-NEXT:    vpbroadcastd %xmm0, %xmm0
; X64-NEXT:    vpaddb {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; X64-NEXT:    retq
  %1 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %a, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3>)
  %2 = add <16 x i8> %1, <i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3>
  ret <16 x i8> %2
}

define <8 x i32> @combine_permd_as_vpbroadcastd256(<4 x i32> %a) {
; X86-LABEL: combine_permd_as_vpbroadcastd256:
; X86:       # %bb.0:
; X86-NEXT:    vpbroadcastd %xmm0, %ymm0
; X86-NEXT:    vpaddd {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0
; X86-NEXT:    retl
;
; X64-LABEL: combine_permd_as_vpbroadcastd256:
; X64:       # %bb.0:
; X64-NEXT:    vpbroadcastd %xmm0, %ymm0
; X64-NEXT:    vpaddd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
; X64-NEXT:    retq
  %1 = shufflevector <4 x i32> %a, <4 x i32> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %2 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %1, <8 x i32> zeroinitializer)
  %3 = add <8 x i32> %2, <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
  ret <8 x i32> %3
}

define <16 x i8> @combine_pshufb_as_vpbroadcastq128(<16 x i8> %a) {
; CHECK-LABEL: combine_pshufb_as_vpbroadcastq128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vmovddup {{.*#+}} xmm0 = xmm0[0,0]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %a, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7>)
  ret <16 x i8> %1
}

define <8 x i32> @combine_permd_as_vpbroadcastq256(<4 x i32> %a) {
; X86-LABEL: combine_permd_as_vpbroadcastq256:
; X86:       # %bb.0:
; X86-NEXT:    vpbroadcastq %xmm0, %ymm0
; X86-NEXT:    vpaddd {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0
; X86-NEXT:    retl
;
; X64-LABEL: combine_permd_as_vpbroadcastq256:
; X64:       # %bb.0:
; X64-NEXT:    vpbroadcastq %xmm0, %ymm0
; X64-NEXT:    vpaddd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
; X64-NEXT:    retq
  %1 = shufflevector <4 x i32> %a, <4 x i32> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %2 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %1, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>)
  %3 = add <8 x i32> %2, <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3>
  ret <8 x i32> %3
}

define <4 x float> @combine_pshufb_as_vpbroadcastss128(<4 x float> %a) {
; CHECK-LABEL: combine_pshufb_as_vpbroadcastss128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vbroadcastss %xmm0, %xmm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = bitcast <4 x float> %a to <16 x i8>
  %2 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %1, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3>)
  %3 = bitcast <16 x i8> %2 to <4 x float>
  ret <4 x float> %3
}

define <8 x float> @combine_permps_as_vpbroadcastss256(<4 x float> %a) {
; CHECK-LABEL: combine_permps_as_vpbroadcastss256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vbroadcastss %xmm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <4 x float> %a, <4 x float> undef, <8 x i32> <i32 0, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %2 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %1, <8 x i32> zeroinitializer)
  ret <8 x float> %2
}

define <4 x double> @combine_permps_as_vpbroadcastsd256(<2 x double> %a) {
; CHECK-LABEL: combine_permps_as_vpbroadcastsd256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vbroadcastsd %xmm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <2 x double> %a, <2 x double> undef, <4 x i32> <i32 0, i32 undef, i32 undef, i32 undef>
  %2 = bitcast <4 x double> %1 to <8 x float>
  %3 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %2, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>)
  %4 = bitcast <8 x float> %3 to <4 x double>
  ret <4 x double> %4
}

define <16 x i8> @combine_vpbroadcast_pshufb_as_vpbroadcastb128(<16 x i8> %a) {
; CHECK-LABEL: combine_vpbroadcast_pshufb_as_vpbroadcastb128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpbroadcastb %xmm0, %xmm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <16 x i8> %a, <16 x i8> undef, <16 x i32> zeroinitializer
  %2 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %1, <16 x i8> zeroinitializer)
  ret <16 x i8> %2
}

define <32 x i8> @combine_vpbroadcast_pshufb_as_vpbroadcastb256(<32 x i8> %a) {
; CHECK-LABEL: combine_vpbroadcast_pshufb_as_vpbroadcastb256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpbroadcastb %xmm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <32 x i8> %a, <32 x i8> undef, <32 x i32> zeroinitializer
  %2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %1, <32 x i8> zeroinitializer)
  ret <32 x i8> %2
}

define <4 x float> @combine_vpbroadcast_pshufb_as_vpbroadcastss128(<4 x float> %a) {
; CHECK-LABEL: combine_vpbroadcast_pshufb_as_vpbroadcastss128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vbroadcastss %xmm0, %xmm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <4 x float> %a, <4 x float> undef, <4 x i32> zeroinitializer
  %2 = bitcast <4 x float> %1 to <16 x i8>
  %3 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %2, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3>)
  %4 = bitcast <16 x i8> %3 to <4 x float>
  ret <4 x float> %4
}

define <8 x float> @combine_vpbroadcast_permd_as_vpbroadcastss256(<4 x float> %a) {
; CHECK-LABEL: combine_vpbroadcast_permd_as_vpbroadcastss256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vbroadcastss %xmm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <4 x float> %a, <4 x float> undef, <8 x i32> zeroinitializer
  %2 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %1, <8 x i32> zeroinitializer)
  ret <8 x float> %2
}

define <4 x double> @combine_vpbroadcast_permd_as_vpbroadcastsd256(<2 x double> %a) {
; CHECK-LABEL: combine_vpbroadcast_permd_as_vpbroadcastsd256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vbroadcastsd %xmm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <2 x double> %a, <2 x double> undef, <4 x i32> zeroinitializer
  %2 = bitcast <4 x double> %1 to <8 x float>
  %3 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %2, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>)
  %4 = bitcast <8 x float> %3 to <4 x double>
  ret <4 x double> %4
}

define <8 x i32> @combine_permd_as_permq(<8 x i32> %a) {
; CHECK-LABEL: combine_permd_as_permq:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,2,2,1]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a, <8 x i32> <i32 0, i32 1, i32 4, i32 5, i32 4, i32 5, i32 2, i32 3>)
  ret <8 x i32> %1
}

define <8 x float> @combine_permps_as_permpd(<8 x float> %a) {
; CHECK-LABEL: combine_permps_as_permpd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[3,2,0,1]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a, <8 x i32> <i32 6, i32 7, i32 4, i32 5, i32 0, i32 1, i32 2, i32 3>)
  ret <8 x float> %1
}

define <8 x float> @combine_permps_as_vpermilps(<8 x float> %a, i32 %a1) {
; CHECK-LABEL: combine_permps_as_vpermilps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[2,2,1,0,7,6,5,4]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = insertelement <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>, i32 %a1, i32 0
  %2 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a, <8 x i32> %1)
  %3 = shufflevector <8 x float> %2, <8 x float> undef, <8 x i32> <i32 1, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  ret <8 x float> %3
}

define <4 x i64> @combine_pshufb_as_zext(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_zext:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpmovzxwq {{.*#+}} ymm0 = xmm0[0],zero,zero,zero,xmm0[1],zero,zero,zero,xmm0[2],zero,zero,zero,xmm0[3],zero,zero,zero
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <32 x i8> %a0, <32 x i8> undef, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
  %2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %1, <32 x i8> <i8 8, i8 9, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 10, i8 11, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 4, i8 5, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 6, i8 7, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  %3 = bitcast <32 x i8> %2 to <4 x i64>
  ret <4 x i64> %3
}

define <4 x i64> @combine_pshufb_as_zext128(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_zext128:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[u,u,u,u,u,u,u,u,7,6,5,4,3,2,1,0]
; CHECK-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,0,1]
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[15,14],zero,zero,zero,zero,zero,zero,ymm0[13,12],zero,zero,zero,zero,zero,zero,ymm0[31,30],zero,zero,zero,zero,zero,zero,ymm0[29,28],zero,zero,zero,zero,zero,zero
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <32 x i8> %a0, <32 x i8> undef, <32 x i32> <i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0, i32 15, i32 14, i32 13, i32 12, i32 11, i32 10, i32 9, i32 8, i32 7, i32 6, i32 5, i32 4, i32 3, i32 2, i32 1, i32 0>
  %2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %1, <32 x i8> <i8 15, i8 14, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 13, i8 12, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 15, i8 14, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 13, i8 12, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  %3 = bitcast <32 x i8> %2 to <4 x i64>
  ret <4 x i64> %3
}

define <4 x double> @combine_pshufb_as_vzmovl_64(<4 x double> %a0) {
; CHECK-LABEL: combine_pshufb_as_vzmovl_64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vmovq {{.*#+}} xmm0 = xmm0[0],zero
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = bitcast <4 x double> %a0 to <32 x i8>
  %2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %1, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  %3 = bitcast <32 x i8> %2 to <4 x double>
  ret <4 x double> %3
}

define <8 x float> @combine_pshufb_as_vzmovl_32(<8 x float> %a0) {
; CHECK-LABEL: combine_pshufb_as_vzmovl_32:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vxorps %xmm1, %xmm1, %xmm1
; CHECK-NEXT:    vblendps {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = bitcast <8 x float> %a0 to <32 x i8>
  %2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %1, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  %3 = bitcast <32 x i8> %2 to <8 x float>
  ret <8 x float> %3
}

define <32 x i8> @combine_pshufb_as_pslldq(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_pslldq:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpslldq {{.*#+}} ymm0 = zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,ymm0[0,1,2,3,4,5],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,ymm0[16,17,18,19,20,21]
; CHECK-NEXT:    ret{{[l|q]}}
  %res0 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5>)
  ret <32 x i8> %res0
}

define <32 x i8> @combine_pshufb_as_psrldq(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_psrldq:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpsrldq {{.*#+}} ymm0 = ymm0[15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,ymm0[31],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; CHECK-NEXT:    ret{{[l|q]}}
  %res0 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 15, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 15, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128, i8 128>)
  ret <32 x i8> %res0
}

define <32 x i8> @combine_pshufb_as_psrlw(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_psrlw:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpsrlw $8, %ymm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %res0 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 1, i8 128, i8 3, i8 128, i8 5, i8 128, i8 7, i8 128, i8 9, i8 128, i8 11, i8 128, i8 13, i8 128, i8 15, i8 128, i8 17, i8 128, i8 19, i8 128, i8 21, i8 128, i8 23, i8 128, i8 25, i8 128, i8 27, i8 128, i8 29, i8 128, i8 31, i8 128>)
  ret <32 x i8> %res0
}

define <32 x i8> @combine_pshufb_as_pslld(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_pslld:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpslld $24, %ymm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %res0 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 128, i8 128, i8 128, i8 0, i8 128, i8 128, i8 128, i8 4, i8 128, i8 128, i8 128, i8 8, i8 128, i8 128, i8 128, i8 12, i8 128, i8 128, i8 128, i8 16, i8 128, i8 128, i8 128, i8 20, i8 128, i8 128, i8 128, i8 24, i8 128, i8 128, i8 128, i8 28>)
  ret <32 x i8> %res0
}

define <32 x i8> @combine_pshufb_as_psrlq(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_psrlq:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpsrlq $40, %ymm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %res0 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 5, i8 6, i8 7, i8 128, i8 128, i8 128, i8 128, i8 128, i8 13, i8 14, i8 15, i8 128, i8 128, i8 128, i8 128, i8 128, i8 21, i8 22, i8 23, i8 128, i8 128, i8 128, i8 128, i8 128, i8 29, i8 30, i8 31, i8 128, i8 128, i8 128, i8 128, i8 128>)
  ret <32 x i8> %res0
}

define <32 x i8> @combine_pshufb_as_pshuflw(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_pshuflw:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshuflw {{.*#+}} ymm0 = ymm0[1,0,3,2,4,5,6,7,9,8,11,10,12,13,14,15]
; CHECK-NEXT:    ret{{[l|q]}}
  %res0 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 2, i8 3, i8 0, i8 1, i8 6, i8 7, i8 4, i8 5, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 2, i8 3, i8 0, i8 1, i8 6, i8 7, i8 4, i8 5, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15>)
  ret <32 x i8> %res0
}

define <32 x i8> @combine_pshufb_as_pshufhw(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_pshufhw:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufhw {{.*#+}} ymm0 = ymm0[0,1,2,3,5,4,7,6,8,9,10,11,13,12,15,14]
; CHECK-NEXT:    ret{{[l|q]}}
  %res0 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 10, i8 11, i8 8, i8 9, i8 14, i8 15, i8 12, i8 13, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 10, i8 11, i8 8, i8 9, i8 14, i8 15, i8 12, i8 13>)
  ret <32 x i8> %res0
}

define <32 x i8> @combine_pshufb_not_as_pshufw(<32 x i8> %a0) {
; AVX2-LABEL: combine_pshufb_not_as_pshufw:
; AVX2:       # %bb.0:
; AVX2-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[2,3,0,1,6,7,4,5,10,11,8,9,14,15,12,13,18,19,16,17,22,23,20,21,26,27,24,25,30,31,28,29]
; AVX2-NEXT:    ret{{[l|q]}}
;
; AVX512-LABEL: combine_pshufb_not_as_pshufw:
; AVX512:       # %bb.0:
; AVX512-NEXT:    # kill: def $ymm0 killed $ymm0 def $zmm0
; AVX512-NEXT:    vprold $16, %zmm0, %zmm0
; AVX512-NEXT:    # kill: def $ymm0 killed $ymm0 killed $zmm0
; AVX512-NEXT:    ret{{[l|q]}}
  %res0 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 2, i8 3, i8 0, i8 1, i8 6, i8 7, i8 4, i8 5, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 2, i8 3, i8 0, i8 1, i8 6, i8 7, i8 4, i8 5, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15>)
  %res1 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %res0, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 10, i8 11, i8 8, i8 9, i8 14, i8 15, i8 12, i8 13, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 10, i8 11, i8 8, i8 9, i8 14, i8 15, i8 12, i8 13>)
  ret <32 x i8> %res1
}

define <32 x i8> @combine_pshufb_as_unpacklo_undef(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_unpacklo_undef:
; CHECK:       # %bb.0:
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 undef, i8 0, i8 undef, i8 1, i8 undef, i8 2, i8 undef, i8 3, i8 undef, i8 4, i8 undef, i8 5, i8 undef, i8 6, i8 undef, i8 7, i8 undef, i8 16, i8 undef, i8 17, i8 undef, i8 18, i8 undef, i8 19, i8 undef, i8 20, i8 undef, i8 21, i8 undef, i8 22, i8 undef, i8 23>)
  %2 = shufflevector <32 x i8> %1, <32 x i8> undef, <32 x i32> <i32 0, i32 0, i32 2, i32 2, i32 4, i32 4, i32 6, i32 6, i32 8, i32 8, i32 10, i32 10, i32 12, i32 12, i32 14, i32 14, i32 16, i32 16, i32 18, i32 18, i32 20, i32 20, i32 22, i32 22, i32 24, i32 24, i32 26, i32 26, i32 28, i32 28, i32 30, i32 30>
  ret <32 x i8> %2
}

define <32 x i8> @combine_pshufb_as_unpacklo_zero(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_unpacklo_zero:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpxor %xmm1, %xmm1, %xmm1
; CHECK-NEXT:    vpunpcklwd {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 0, i8 1, i8 -1, i8 -1, i8 2, i8 3, i8 -1, i8 -1, i8 4, i8 5, i8 -1, i8 -1, i8 6, i8 7, i8 -1, i8 -1, i8 16, i8 17, i8 -1, i8 -1, i8 18, i8 19, i8 -1, i8 -1, i8 20, i8 21, i8 -1, i8 -1, i8 22, i8 23, i8 -1, i8 -1>)
  ret <32 x i8> %1
}

define <32 x i8> @combine_pshufb_as_unpackhi_zero(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_as_unpackhi_zero:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpxor %xmm1, %xmm1, %xmm1
; CHECK-NEXT:    vpunpckhbw {{.*#+}} ymm0 = ymm1[8],ymm0[8],ymm1[9],ymm0[9],ymm1[10],ymm0[10],ymm1[11],ymm0[11],ymm1[12],ymm0[12],ymm1[13],ymm0[13],ymm1[14],ymm0[14],ymm1[15],ymm0[15],ymm1[24],ymm0[24],ymm1[25],ymm0[25],ymm1[26],ymm0[26],ymm1[27],ymm0[27],ymm1[28],ymm0[28],ymm1[29],ymm0[29],ymm1[30],ymm0[30],ymm1[31],ymm0[31]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 -1, i8 8, i8 -1, i8 9, i8 -1, i8 10, i8 -1, i8 11, i8 -1, i8 12, i8 -1, i8 13, i8 -1, i8 14, i8 -1, i8 15, i8 -1, i8 24, i8 -1, i8 25, i8 -1, i8 26, i8 -1, i8 27, i8 -1, i8 28, i8 -1, i8 29, i8 -1, i8 30, i8 -1, i8 31>)
  ret <32 x i8> %1
}

define <32 x i8> @combine_psrlw_pshufb(<16 x i16> %a0) {
; X86-LABEL: combine_psrlw_pshufb:
; X86:       # %bb.0:
; X86-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0
; X86-NEXT:    retl
;
; X64-LABEL: combine_psrlw_pshufb:
; X64:       # %bb.0:
; X64-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
; X64-NEXT:    retq
  %1 = lshr <16 x i16> %a0, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
  %2 = bitcast <16 x i16> %1 to <32 x i8>
  %3 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 1, i8 0, i8 3, i8 2, i8 5, i8 4, i8 7, i8 6, i8 9, i8 8, i8 11, i8 10, i8 13, i8 12, i8 15, i8 14, i8 17, i8 16, i8 19, i8 18, i8 21, i8 20, i8 23, i8 22, i8 25, i8 24, i8 27, i8 26, i8 29, i8 28, i8 31, i8 30>)
  ret <32 x i8> %3
}

define <32 x i8> @combine_pslld_pshufb(<8 x i32> %a0) {
; X86-LABEL: combine_pslld_pshufb:
; X86:       # %bb.0:
; X86-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}, %ymm0, %ymm0
; X86-NEXT:    retl
;
; X64-LABEL: combine_pslld_pshufb:
; X64:       # %bb.0:
; X64-NEXT:    vandps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
; X64-NEXT:    retq
  %1 = shl <8 x i32> %a0, <i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24, i32 24>
  %2 = bitcast <8 x i32> %1 to <32 x i8>
  %3 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 3, i8 2, i8 1, i8 0, i8 7, i8 6, i8 5, i8 4, i8 11, i8 10, i8 9, i8 8, i8 15, i8 14, i8 13, i8 12, i8 19, i8 18, i8 17, i8 16, i8 23, i8 22, i8 21, i8 20, i8 27, i8 26, i8 25, i8 24, i8 31, i8 30, i8 29, i8 28>)
  ret <32 x i8> %3
}

define <32 x i8> @combine_psrlq_pshufb(<4 x i64> %a0) {
; CHECK-LABEL: combine_psrlq_pshufb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = zero,zero,zero,zero,ymm0[7,6,5,4],zero,zero,zero,zero,ymm0[15,14,13,12],zero,zero,zero,zero,ymm0[23,22,21],zero,zero,zero,zero,ymm0[31,30,29,28],zero
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = lshr <4 x i64> %a0, <i64 32, i64 32, i64 32, i64 32>
  %2 = bitcast <4 x i64> %1 to <32 x i8>
  %3 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %2, <32 x i8> <i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 15, i8 14, i8 13, i8 12, i8 11, i8 10, i8 9, i8 8, i8 23, i8 22, i8 21, i8 20, i8 19, i8 18, i8 17, i8 31, i8 30, i8 29, i8 28, i8 27, i8 26, i8 25, i8 24, i8 23>)
  ret <32 x i8> %3
}

define <32 x i8> @combine_unpack_unpack_pshufb(<32 x i8> %a0) {
; CHECK-LABEL: combine_unpack_unpack_pshufb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[0,0,4,8,1,1,5,9,2,2,6,10,3,3,7,11,16,16,20,24,17,17,21,25,18,18,22,26,19,19,23,27]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <32 x i8> %a0, <32 x i8> undef, <32 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3, i32 16, i32 17, i32 18, i32 19, i32 16, i32 17, i32 18, i32 19, i32 16, i32 17, i32 18, i32 19, i32 16, i32 17, i32 18, i32 19>
  %2 = shufflevector <32 x i8> %a0, <32 x i8> undef, <32 x i32> <i32 4, i32 5, i32 6, i32 7, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 20, i32 21, i32 22, i32 23, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %3 = shufflevector <32 x i8> %a0, <32 x i8> undef, <32 x i32> <i32 8, i32 9, i32 10, i32 11, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 24, i32 25, i32 26, i32 27, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %4 = shufflevector <32 x i8> %1, <32 x i8> %2, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5 = shufflevector <32 x i8> %1, <32 x i8> %3, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %6 = shufflevector <32 x i8> %4, <32 x i8> %5, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 4, i32 36, i32 5, i32 37, i32 6, i32 38, i32 7, i32 39, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 20, i32 52, i32 21, i32 53, i32 22, i32 54, i32 23, i32 55>
  ret <32 x i8> %6
}

define <16 x i16> @shuffle_combine_packssdw_pshufb(<8 x i32> %a0) {
; CHECK-LABEL: shuffle_combine_packssdw_pshufb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpsrad $31, %ymm0, %ymm0
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[12,13,8,9,4,5,0,1,12,13,8,9,4,5,0,1,16,17,20,21,24,25,28,29,28,29,24,25,20,21,16,17]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = ashr <8 x i32> %a0, <i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31, i32 31>
  %2 = tail call <16 x i16> @llvm.x86.avx2.packssdw(<8 x i32> %1, <8 x i32> %1)
  %3 = shufflevector <16 x i16> %2, <16 x i16> undef, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0, i32 8, i32 9, i32 10, i32 11, i32 11, i32 10, i32 9, i32 8>
  ret <16 x i16> %3
}
declare <16 x i16> @llvm.x86.avx2.packssdw(<8 x i32>, <8 x i32>) nounwind readnone

define <32 x i8> @shuffle_combine_packsswb_pshufb(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: shuffle_combine_packsswb_pshufb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpsraw $15, %ymm0, %ymm0
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[14,12,10,8,6,4,2,0,14,12,10,8,6,4,2,0,30,28,26,24,22,20,18,16,30,28,26,24,22,20,18,16]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = ashr <16 x i16> %a0, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
  %2 = ashr <16 x i16> %a1, <i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15, i16 15>
  %3 = tail call <32 x i8> @llvm.x86.avx2.packsswb(<16 x i16> %1, <16 x i16> %2)
  %4 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %3, <32 x i8> <i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
  ret <32 x i8> %4
}
declare <32 x i8> @llvm.x86.avx2.packsswb(<16 x i16>, <16 x i16>) nounwind readnone

define <16 x i16> @shuffle_combine_packusdw_pshufb(<8 x i32> %a0, <8 x i32> %a1) {
; CHECK-LABEL: shuffle_combine_packusdw_pshufb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[14,15,10,11,6,7,2,3,14,15,10,11,6,7,2,3,18,19,22,23,26,27,30,31,30,31,26,27,22,23,18,19]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = lshr <8 x i32> %a0, <i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16, i32 16>
  %2 = tail call <16 x i16> @llvm.x86.avx2.packusdw(<8 x i32> %1, <8 x i32> %1)
  %3 = shufflevector <16 x i16> %2, <16 x i16> undef, <16 x i32> <i32 3, i32 2, i32 1, i32 0, i32 3, i32 2, i32 1, i32 0, i32 8, i32 9, i32 10, i32 11, i32 11, i32 10, i32 9, i32 8>
  ret <16 x i16> %3
}
declare <16 x i16> @llvm.x86.avx2.packusdw(<8 x i32>, <8 x i32>) nounwind readnone

define <32 x i8> @shuffle_combine_packuswb_pshufb(<16 x i16> %a0, <16 x i16> %a1) {
; CHECK-LABEL: shuffle_combine_packuswb_pshufb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[15,13,11,9,7,5,3,1,15,13,11,9,7,5,3,1,31,29,27,25,23,21,19,17,31,29,27,25,23,21,19,17]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = lshr <16 x i16> %a0, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
  %2 = lshr <16 x i16> %a1, <i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8, i16 8>
  %3 = tail call <32 x i8> @llvm.x86.avx2.packuswb(<16 x i16> %1, <16 x i16> %2)
  %4 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %3, <32 x i8> <i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>)
  ret <32 x i8> %4
}
declare <32 x i8> @llvm.x86.avx2.packuswb(<16 x i16>, <16 x i16>) nounwind readnone

define <32 x i8> @combine_pshufb_as_packsswb(<16 x i16> %a0, <16 x i16> %a1) nounwind {
; CHECK-LABEL: combine_pshufb_as_packsswb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpsraw $11, %ymm0, %ymm0
; CHECK-NEXT:    vpsraw $11, %ymm1, %ymm1
; CHECK-NEXT:    vpacksswb %ymm1, %ymm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = ashr <16 x i16> %a0, <i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11>
  %2 = ashr <16 x i16> %a1, <i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11>
  %3 = bitcast <16 x i16> %1 to <32 x i8>
  %4 = bitcast <16 x i16> %2 to <32 x i8>
  %5 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %3, <32 x i8> <i8 0, i8 2, i8 4, i8 6, i8 8, i8 10, i8 12, i8 14, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 2, i8 4, i8 6, i8 8, i8 10, i8 12, i8 14, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  %6 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %4, <32 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 2, i8 4, i8 6, i8 8, i8 10, i8 12, i8 14, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 2, i8 4, i8 6, i8 8, i8 10, i8 12, i8 14>)
  %7 = or <32 x i8> %5, %6
  ret <32 x i8> %7
}

define <32 x i8> @combine_pshufb_as_packuswb(<16 x i16> %a0, <16 x i16> %a1) nounwind {
; CHECK-LABEL: combine_pshufb_as_packuswb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpsrlw $11, %ymm0, %ymm0
; CHECK-NEXT:    vpsrlw $11, %ymm1, %ymm1
; CHECK-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = lshr <16 x i16> %a0, <i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11>
  %2 = lshr <16 x i16> %a1, <i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11, i16 11>
  %3 = bitcast <16 x i16> %1 to <32 x i8>
  %4 = bitcast <16 x i16> %2 to <32 x i8>
  %5 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %3, <32 x i8> <i8 0, i8 2, i8 4, i8 6, i8 8, i8 10, i8 12, i8 14, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 2, i8 4, i8 6, i8 8, i8 10, i8 12, i8 14, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  %6 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %4, <32 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 2, i8 4, i8 6, i8 8, i8 10, i8 12, i8 14, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 2, i8 4, i8 6, i8 8, i8 10, i8 12, i8 14>)
  %7 = or <32 x i8> %5, %6
  ret <32 x i8> %7
}

define <16 x i8> @combine_pshufb_insertion_as_broadcast_v2i64(i64 %a0) {
; X86-LABEL: combine_pshufb_insertion_as_broadcast_v2i64:
; X86:       # %bb.0:
; X86-NEXT:    vmovddup {{.*#+}} xmm0 = mem[0,0]
; X86-NEXT:    retl
;
; X64-LABEL: combine_pshufb_insertion_as_broadcast_v2i64:
; X64:       # %bb.0:
; X64-NEXT:    vmovq %rdi, %xmm0
; X64-NEXT:    vpbroadcastq %xmm0, %xmm0
; X64-NEXT:    retq
  %1 = insertelement <2 x i64> undef, i64 %a0, i32 0
  %2 = bitcast <2 x i64> %1 to <16 x i8>
  %3 = call <16 x i8> @llvm.x86.ssse3.pshuf.b.128(<16 x i8> %2, <16 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7>)
  ret <16 x i8> %3
}

define <8 x i32> @combine_permd_insertion_as_broadcast_v4i64(i64 %a0) {
; X86-LABEL: combine_permd_insertion_as_broadcast_v4i64:
; X86:       # %bb.0:
; X86-NEXT:    vbroadcastsd {{[0-9]+}}(%esp), %ymm0
; X86-NEXT:    retl
;
; X64-LABEL: combine_permd_insertion_as_broadcast_v4i64:
; X64:       # %bb.0:
; X64-NEXT:    vmovq %rdi, %xmm0
; X64-NEXT:    vpbroadcastq %xmm0, %ymm0
; X64-NEXT:    retq
  %1 = insertelement <4 x i64> undef, i64 %a0, i32 0
  %2 = bitcast <4 x i64> %1 to <8 x i32>
  %3 = tail call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %2, <8 x i32> <i32 0, i32 1, i32 0, i32 1, i32 0, i32 1, i32 0, i32 1>)
  ret <8 x i32> %3
}

define <32 x i8> @combine_pshufb_pshufb_or_as_blend(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: combine_pshufb_pshufb_or_as_blend:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vblendps {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5],ymm1[6,7]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1>)
  %2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a1, <32 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 8, i8 9, i8 10, i8 11, i8 12, i8 13, i8 14, i8 15>)
  %3 = or <32 x i8> %1, %2
  ret <32 x i8> %3
}

define <32 x i8> @combine_pshufb_pshufb_or_as_unpcklbw(<32 x i8> %a0, <32 x i8> %a1) {
; CHECK-LABEL: combine_pshufb_pshufb_or_as_unpcklbw:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0],ymm1[0],ymm0[1],ymm1[1],ymm0[2],ymm1[2],ymm0[3],ymm1[3],ymm0[4],ymm1[4],ymm0[5],ymm1[5],ymm0[6],ymm1[6],ymm0[7],ymm1[7],ymm0[16],ymm1[16],ymm0[17],ymm1[17],ymm0[18],ymm1[18],ymm0[19],ymm1[19],ymm0[20],ymm1[20],ymm0[21],ymm1[21],ymm0[22],ymm1[22],ymm0[23],ymm1[23]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 0, i8 -1, i8 1, i8 -1, i8 2, i8 -1, i8 3, i8 -1, i8 4, i8 -1, i8 5, i8 -1, i8 6, i8 -1, i8 7, i8 -1, i8 0, i8 -1, i8 1, i8 -1, i8 2, i8 -1, i8 3, i8 -1, i8 4, i8 -1, i8 5, i8 -1, i8 6, i8 -1, i8 7, i8 -1>)
  %2 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a1, <32 x i8> <i8 -1, i8 0, i8 -1, i8 1, i8 -1, i8 2, i8 -1, i8 3, i8 -1, i8 4, i8 -1, i8 5, i8 -1, i8 6, i8 -1, i8 7, i8 -1, i8 0, i8 -1, i8 1, i8 -1, i8 2, i8 -1, i8 3, i8 -1, i8 4, i8 -1, i8 5, i8 -1, i8 6, i8 -1, i8 7>)
  %3 = or <32 x i8> %1, %2
  ret <32 x i8> %3
}

define <32 x i8> @combine_pshufb_pshufb_or_pshufb(<32 x i8> %a0) {
; CHECK-LABEL: combine_pshufb_pshufb_or_pshufb:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpermilps {{.*#+}} ymm0 = ymm0[0,0,0,0,4,4,4,4]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 2, i8 3, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 2, i8 3, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 2, i8 3, i8 -1, i8 -1, i8 -1, i8 -1>)
  %2 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %a0, <32 x i8> <i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 2, i8 3, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 2, i8 3, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 2, i8 3, i8 -1, i8 -1, i8 -1, i8 -1, i8 0, i8 1, i8 2, i8 3>)
  %3 = or <32 x i8> %1, %2
  %4 = call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> %3, <32 x i8> <i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7>)
  ret <32 x i8> %4
}

define <8 x i32> @constant_fold_permd() {
; CHECK-LABEL: constant_fold_permd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vmovaps {{.*#+}} ymm0 = [5,7,3,2,8,2,6,1]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> <i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8>, <8 x i32> <i32 4, i32 6, i32 2, i32 1, i32 7, i32 1, i32 5, i32 0>)
  ret <8 x i32> %1
}

define <8 x float> @constant_fold_permps() {
; CHECK-LABEL: constant_fold_permps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vmovaps {{.*#+}} ymm0 = [5.0E+0,7.0E+0,3.0E+0,2.0E+0,8.0E+0,2.0E+0,6.0E+0,1.0E+0]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> <float 1.0, float 2.0, float 3.0, float 4.0, float 5.0, float 6.0, float 7.0, float 8.0>, <8 x i32> <i32 4, i32 6, i32 2, i32 1, i32 7, i32 1, i32 5, i32 0>)
  ret <8 x float> %1
}

define <32 x i8> @constant_fold_pshufb_256() {
; CHECK-LABEL: constant_fold_pshufb_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vmovaps {{.*#+}} ymm0 = <14,0,0,0,u,u,0,0,0,0,0,0,0,0,8,9,255,0,0,0,u,u,0,0,241,0,0,0,0,0,249,250>
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = tail call <32 x i8> @llvm.x86.avx2.pshuf.b(<32 x i8> <i8 15, i8 14, i8 13, i8 12, i8 11, i8 10, i8 9, i8 8, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 -1, i8 -2, i8 -3, i8 -4, i8 -5, i8 -6, i8 -7, i8 -8, i8 -9, i8 -10, i8 -11, i8 -12, i8 -13, i8 -14, i8 -15>, <32 x i8> <i8 1, i8 -1, i8 -1, i8 -1, i8 undef, i8 undef, i8 -1, i8 -1, i8 15, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 7, i8 6, i8 1, i8 -1, i8 -1, i8 -1, i8 undef, i8 undef, i8 -1, i8 -1, i8 15, i8 -1, i8 -1, i8 -1, i8 -1, i8 -1, i8 7, i8 6>)
  ret <32 x i8> %1
}

define i32 @broadcast_v2i64_multiuse(i64* %p0) {
; X86-LABEL: broadcast_v2i64_multiuse:
; X86:       # %bb.0: # %entry
; X86-NEXT:    movl {{[0-9]+}}(%esp), %eax
; X86-NEXT:    movl (%eax), %eax
; X86-NEXT:    addl %eax, %eax
; X86-NEXT:    retl
;
; X64-LABEL: broadcast_v2i64_multiuse:
; X64:       # %bb.0: # %entry
; X64-NEXT:    movl (%rdi), %eax
; X64-NEXT:    addl %eax, %eax
; X64-NEXT:    retq
entry:
  %tmp = load i64, i64* %p0, align 8
  %tmp1 = trunc i64 %tmp to i32
  %tmp2 = insertelement <2 x i64> undef, i64 %tmp, i32 0
  %tmp3 = shufflevector <2 x i64> %tmp2, <2 x i64> undef, <2 x i32> zeroinitializer
  %tmp4 = trunc <2 x i64> %tmp3 to <2 x i32>
  %tmp5 = extractelement <2 x i32> %tmp4, i32 1
  %tmp6 = add i32 %tmp1, %tmp5
  ret i32 %tmp6
}

define <32 x i8> @PR27320(<8 x i32> %a0) {
; CHECK-LABEL: PR27320:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,1,2]
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[0,1,1,2,3,4,4,5,6,7,7,8,9,10,10,11,20,21,21,22,23,24,24,25,26,27,27,28,29,30,30,31]
; CHECK-NEXT:    ret{{[l|q]}}
  %1 = shufflevector <8 x i32> %a0, <8 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 undef, i32 3, i32 4, i32 5, i32 undef>
  %2 = bitcast <8 x i32> %1 to <32 x i8>
  %3 = shufflevector <32 x i8> %2, <32 x i8> undef, <32 x i32> <i32 0, i32 1, i32 1, i32 2, i32 3, i32 4, i32 4, i32 5, i32 6, i32 7, i32 7, i32 8, i32 9, i32 10, i32 10, i32 11, i32 16, i32 17, i32 17, i32 18, i32 19, i32 20, i32 20, i32 21, i32 22, i32 23, i32 23, i32 24, i32 25, i32 26, i32 26, i32 27>
  ret <32 x i8> %3
}

define internal fastcc <8 x float> @PR34577(<8 x float> %inp0, <8 x float> %inp1, <8 x float> %inp2) {
; AVX2-LABEL: PR34577:
; AVX2:       # %bb.0: # %entry
; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[1,1,1,1]
; AVX2-NEXT:    vxorps %xmm2, %xmm2, %xmm2
; AVX2-NEXT:    vblendps {{.*#+}} ymm0 = ymm2[0,1,2,3],ymm0[4,5,6,7]
; AVX2-NEXT:    vmovaps {{.*#+}} ymm2 = <u,u,7,2,u,u,3,2>
; AVX2-NEXT:    vpermps %ymm1, %ymm2, %ymm1
; AVX2-NEXT:    vblendps {{.*#+}} ymm0 = ymm0[0,1],ymm1[2,3],ymm0[4,5],ymm1[6,7]
; AVX2-NEXT:    ret{{[l|q]}}
;
; AVX512-LABEL: PR34577:
; AVX512:       # %bb.0: # %entry
; AVX512-NEXT:    # kill: def $ymm1 killed $ymm1 def $zmm1
; AVX512-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[1,1,1,1]
; AVX512-NEXT:    vxorps %xmm2, %xmm2, %xmm2
; AVX512-NEXT:    vblendps {{.*#+}} ymm2 = ymm0[0,1],ymm2[2,3],ymm0[4,5],ymm2[6,7]
; AVX512-NEXT:    vmovaps {{.*#+}} ymm0 = <23,18,7,2,20,u,3,2>
; AVX512-NEXT:    vpermi2ps %zmm2, %zmm1, %zmm0
; AVX512-NEXT:    # kill: def $ymm0 killed $ymm0 killed $zmm0
; AVX512-NEXT:    ret{{[l|q]}}
entry:
  %shuf0 = shufflevector <8 x float> %inp0, <8 x float> %inp2, <8 x i32> <i32 1, i32 10, i32 11, i32 13, i32 2, i32 13, i32 5, i32 0>
  %sel = select <8 x i1> <i1 false, i1 true, i1 true, i1 false, i1 true, i1 false, i1 true, i1 false>, <8 x float> %shuf0, <8 x float> zeroinitializer
  %shuf1 = shufflevector <8 x float> zeroinitializer, <8 x float> %sel, <8 x i32> <i32 6, i32 11, i32 6, i32 15, i32 12, i32 11, i32 1, i32 3>
  %shuf2 = shufflevector <8 x float> %inp1, <8 x float> %shuf1, <8 x i32> <i32 15, i32 10, i32 7, i32 2, i32 12, i32 undef, i32 3, i32 2>
  ret <8 x float> %shuf2
}

define <32 x i8> @PR52122(<32 x i8> %0, <32 x i8> %1) {
; CHECK-LABEL: PR52122:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vpshufb {{.*#+}} ymm1 = zero,zero,ymm1[4],zero,zero,zero,ymm1[5],zero,zero,zero,ymm1[6],zero,zero,zero,ymm1[7],zero,zero,zero,ymm1[20],zero,zero,zero,ymm1[21],zero,zero,zero,ymm1[22],zero,zero,zero,ymm1[23],zero
; CHECK-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[2,4],zero,zero,zero,ymm0[5],zero,zero,ymm0[3,6],zero,zero,zero,ymm0[7],zero,zero,ymm0[18,20],zero,zero,zero,ymm0[21],zero,zero,ymm0[19,22],zero,zero,zero,ymm0[23],zero,zero
; CHECK-NEXT:    vpor %ymm1, %ymm0, %ymm0
; CHECK-NEXT:    ret{{[l|q]}}
  %3 = shufflevector <32 x i8> %0, <32 x i8> <i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 poison, i8 poison, i8 poison, i8 poison, i8 poison, i8 poison, i8 poison, i8 poison, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 0, i8 poison, i8 poison, i8 poison, i8 poison, i8 poison, i8 poison, i8 poison, i8 poison>, <32 x i32> <i32 0, i32 32, i32 1, i32 33, i32 2, i32 34, i32 3, i32 35, i32 4, i32 36, i32 5, i32 37, i32 6, i32 38, i32 7, i32 39, i32 16, i32 48, i32 17, i32 49, i32 18, i32 50, i32 19, i32 51, i32 20, i32 52, i32 21, i32 53, i32 22, i32 54, i32 23, i32 55>
  %4 = shufflevector <32 x i8> %3, <32 x i8> %1, <32 x i32> <i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 4, i32 36, i32 5, i32 37, i32 6, i32 38, i32 7, i32 39, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 20, i32 52, i32 21, i32 53, i32 22, i32 54, i32 23, i32 55>
  %5 = shufflevector <32 x i8> %4, <32 x i8> %3, <32 x i32> <i32 8, i32 40, i32 9, i32 41, i32 10, i32 42, i32 11, i32 43, i32 12, i32 44, i32 13, i32 45, i32 14, i32 46, i32 15, i32 47, i32 24, i32 56, i32 25, i32 57, i32 26, i32 58, i32 27, i32 59, i32 28, i32 60, i32 29, i32 61, i32 30, i32 62, i32 31, i32 63>
  ret <32 x i8> %5
}

define void @packss_zext_v8i1() {
; X86-LABEL: packss_zext_v8i1:
; X86:       # %bb.0:
; X86-NEXT:    vxorps %xmm0, %xmm0, %xmm0
; X86-NEXT:    vmovups %ymm0, (%eax)
; X86-NEXT:    vzeroupper
; X86-NEXT:    retl
;
; X64-LABEL: packss_zext_v8i1:
; X64:       # %bb.0:
; X64-NEXT:    vxorps %xmm0, %xmm0, %xmm0
; X64-NEXT:    vmovups %ymm0, (%rax)
; X64-NEXT:    vzeroupper
; X64-NEXT:    retq
  %tmp0 = icmp sgt <8 x i32> undef, undef
  %tmp1 = zext <8 x i1> %tmp0 to <8 x i32>
  %tmp2 = shufflevector <8 x i32> %tmp1, <8 x i32> zeroinitializer, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9, i32 10, i32 11, i32 12, i32 13, i32 14, i32 15>
  %tmp3 = trunc <16 x i32> %tmp2 to <16 x i16>
  %tmp4 = add <16 x i16> zeroinitializer, %tmp3
  %tmp6 = sext <16 x i16> %tmp4 to <16 x i32>
  %tmp10 = shufflevector <16 x i32> %tmp6, <16 x i32> undef, <8 x i32> <i32 4, i32 5, i32 6, i32 7, i32 12, i32 13, i32 14, i32 15>
  %tmp11 = tail call <16 x i16> @llvm.x86.avx2.packssdw(<8 x i32> undef, <8 x i32> %tmp10)
  store <16 x i16> %tmp11, <16 x i16>* undef, align 2
  ret void
}