; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown | FileCheck %s --check-prefix=SSE --check-prefix=SSE2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse4.1 | FileCheck %s --check-prefix=SSE --check-prefix=SSE41
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx2 | FileCheck %s --check-prefix=AVX --check-prefix=AVX2
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512f | FileCheck %s --check-prefix=AVX --check-prefix=AVX512 --check-prefix=AVX512F
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512bw | FileCheck %s --check-prefix=AVX --check-prefix=AVX512 --check-prefix=AVX512BW

define <16 x i8> @mul_v16i8c(<16 x i8> %i) nounwind  {
; SSE2-LABEL: mul_v16i8c:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    movdqa %xmm0, %xmm1
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa {{.*#+}} xmm2 = [117,117,117,117,117,117,117,117]
; SSE2-NEXT:    pmullw %xmm2, %xmm1
; SSE2-NEXT:    movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT:    pand %xmm3, %xmm1
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm2, %xmm0
; SSE2-NEXT:    pand %xmm3, %xmm0
; SSE2-NEXT:    packuswb %xmm1, %xmm0
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v16i8c:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm1 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    movdqa {{.*#+}} xmm2 = [117,117,117,117,117,117,117,117]
; SSE41-NEXT:    pmullw %xmm2, %xmm0
; SSE41-NEXT:    movdqa {{.*#+}} xmm3 = [255,255,255,255,255,255,255,255]
; SSE41-NEXT:    pand %xmm3, %xmm0
; SSE41-NEXT:    pmullw %xmm2, %xmm1
; SSE41-NEXT:    pand %xmm3, %xmm1
; SSE41-NEXT:    packuswb %xmm0, %xmm1
; SSE41-NEXT:    movdqa %xmm1, %xmm0
; SSE41-NEXT:    retq
;
; AVX2-LABEL: mul_v16i8c:
; AVX2:       # %bb.0: # %entry
; AVX2-NEXT:    vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; AVX2-NEXT:    vpmullw {{.*}}(%rip), %ymm0, %ymm0
; AVX2-NEXT:    vpand {{.*}}(%rip), %ymm0, %ymm0
; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT:    vpackuswb %xmm1, %xmm0, %xmm0
; AVX2-NEXT:    vzeroupper
; AVX2-NEXT:    retq
;
; AVX512F-LABEL: mul_v16i8c:
; AVX512F:       # %bb.0: # %entry
; AVX512F-NEXT:    vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; AVX512F-NEXT:    vpmullw {{.*}}(%rip), %ymm0, %ymm0
; AVX512F-NEXT:    vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512F-NEXT:    vpmovdb %zmm0, %xmm0
; AVX512F-NEXT:    vzeroupper
; AVX512F-NEXT:    retq
;
; AVX512BW-LABEL: mul_v16i8c:
; AVX512BW:       # %bb.0: # %entry
; AVX512BW-NEXT:    vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; AVX512BW-NEXT:    vpmullw {{.*}}(%rip), %ymm0, %ymm0
; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
; AVX512BW-NEXT:    # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX512BW-NEXT:    vzeroupper
; AVX512BW-NEXT:    retq
entry:
  %A = mul <16 x i8> %i, < i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117 >
  ret <16 x i8> %A
}

define <8 x i16> @mul_v8i16c(<8 x i16> %i) nounwind  {
; SSE-LABEL: mul_v8i16c:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    pmullw {{.*}}(%rip), %xmm0
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v8i16c:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmullw {{.*}}(%rip), %xmm0, %xmm0
; AVX-NEXT:    retq
entry:
  %A = mul <8 x i16> %i, < i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117 >
  ret <8 x i16> %A
}

define <4 x i32> @mul_v4i32c(<4 x i32> %i) nounwind  {
; SSE2-LABEL: mul_v4i32c:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    movdqa {{.*#+}} xmm1 = [117,117,117,117]
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm1, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT:    pmuludq %xmm1, %xmm2
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[0,2,2,3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v4i32c:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmulld {{.*}}(%rip), %xmm0
; SSE41-NEXT:    retq
;
; AVX-LABEL: mul_v4i32c:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpbroadcastd {{.*#+}} xmm1 = [117,117,117,117]
; AVX-NEXT:    vpmulld %xmm1, %xmm0, %xmm0
; AVX-NEXT:    retq
entry:
  %A = mul <4 x i32> %i, < i32 117, i32 117, i32 117, i32 117 >
  ret <4 x i32> %A
}

define <2 x i64> @mul_v2i64c(<2 x i64> %i) nounwind  {
; SSE-LABEL: mul_v2i64c:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    movdqa {{.*#+}} xmm1 = [117,117]
; SSE-NEXT:    movdqa %xmm0, %xmm2
; SSE-NEXT:    pmuludq %xmm1, %xmm2
; SSE-NEXT:    psrlq $32, %xmm0
; SSE-NEXT:    pmuludq %xmm1, %xmm0
; SSE-NEXT:    psllq $32, %xmm0
; SSE-NEXT:    paddq %xmm2, %xmm0
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v2i64c:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vmovdqa {{.*#+}} xmm1 = [117,117]
; AVX-NEXT:    vpmuludq %xmm1, %xmm0, %xmm2
; AVX-NEXT:    vpsrlq $32, %xmm0, %xmm0
; AVX-NEXT:    vpmuludq %xmm1, %xmm0, %xmm0
; AVX-NEXT:    vpsllq $32, %xmm0, %xmm0
; AVX-NEXT:    vpaddq %xmm0, %xmm2, %xmm0
; AVX-NEXT:    retq
entry:
  %A = mul <2 x i64> %i, < i64 117, i64 117 >
  ret <2 x i64> %A
}

define <16 x i8> @mul_v16i8(<16 x i8> %i, <16 x i8> %j) nounwind  {
; SSE2-LABEL: mul_v16i8:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    movdqa %xmm1, %xmm2
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa %xmm0, %xmm3
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm2, %xmm3
; SSE2-NEXT:    movdqa {{.*#+}} xmm2 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT:    pand %xmm2, %xmm3
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm1, %xmm0
; SSE2-NEXT:    pand %xmm2, %xmm0
; SSE2-NEXT:    packuswb %xmm3, %xmm0
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v16i8:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm3 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm2 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm1, %xmm0
; SSE41-NEXT:    movdqa {{.*#+}} xmm1 = [255,255,255,255,255,255,255,255]
; SSE41-NEXT:    pand %xmm1, %xmm0
; SSE41-NEXT:    pmullw %xmm3, %xmm2
; SSE41-NEXT:    pand %xmm1, %xmm2
; SSE41-NEXT:    packuswb %xmm0, %xmm2
; SSE41-NEXT:    movdqa %xmm2, %xmm0
; SSE41-NEXT:    retq
;
; AVX2-LABEL: mul_v16i8:
; AVX2:       # %bb.0: # %entry
; AVX2-NEXT:    vpmovzxbw {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero,xmm1[8],zero,xmm1[9],zero,xmm1[10],zero,xmm1[11],zero,xmm1[12],zero,xmm1[13],zero,xmm1[14],zero,xmm1[15],zero
; AVX2-NEXT:    vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; AVX2-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
; AVX2-NEXT:    vpand {{.*}}(%rip), %ymm0, %ymm0
; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm1
; AVX2-NEXT:    vpackuswb %xmm1, %xmm0, %xmm0
; AVX2-NEXT:    vzeroupper
; AVX2-NEXT:    retq
;
; AVX512F-LABEL: mul_v16i8:
; AVX512F:       # %bb.0: # %entry
; AVX512F-NEXT:    vpmovzxbw {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero,xmm1[8],zero,xmm1[9],zero,xmm1[10],zero,xmm1[11],zero,xmm1[12],zero,xmm1[13],zero,xmm1[14],zero,xmm1[15],zero
; AVX512F-NEXT:    vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; AVX512F-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
; AVX512F-NEXT:    vpmovzxwd {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero
; AVX512F-NEXT:    vpmovdb %zmm0, %xmm0
; AVX512F-NEXT:    vzeroupper
; AVX512F-NEXT:    retq
;
; AVX512BW-LABEL: mul_v16i8:
; AVX512BW:       # %bb.0: # %entry
; AVX512BW-NEXT:    vpmovzxbw {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero,xmm1[8],zero,xmm1[9],zero,xmm1[10],zero,xmm1[11],zero,xmm1[12],zero,xmm1[13],zero,xmm1[14],zero,xmm1[15],zero
; AVX512BW-NEXT:    vpmovzxbw {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero,xmm0[8],zero,xmm0[9],zero,xmm0[10],zero,xmm0[11],zero,xmm0[12],zero,xmm0[13],zero,xmm0[14],zero,xmm0[15],zero
; AVX512BW-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
; AVX512BW-NEXT:    # kill: def $xmm0 killed $xmm0 killed $ymm0
; AVX512BW-NEXT:    vzeroupper
; AVX512BW-NEXT:    retq
entry:
  %A = mul <16 x i8> %i, %j
  ret <16 x i8> %A
}

define <8 x i16> @mul_v8i16(<8 x i16> %i, <8 x i16> %j) nounwind  {
; SSE-LABEL: mul_v8i16:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    pmullw %xmm1, %xmm0
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v8i16:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmullw %xmm1, %xmm0, %xmm0
; AVX-NEXT:    retq
entry:
  %A = mul <8 x i16> %i, %j
  ret <8 x i16> %A
}

define <4 x i32> @mul_v4i32(<4 x i32> %i, <4 x i32> %j) nounwind  {
; SSE2-LABEL: mul_v4i32:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm1, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm1
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v4i32:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmulld %xmm1, %xmm0
; SSE41-NEXT:    retq
;
; AVX-LABEL: mul_v4i32:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmulld %xmm1, %xmm0, %xmm0
; AVX-NEXT:    retq
entry:
  %A = mul <4 x i32> %i, %j
  ret <4 x i32> %A
}

define <2 x i64> @mul_v2i64(<2 x i64> %i, <2 x i64> %j) nounwind  {
; SSE-LABEL: mul_v2i64:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    movdqa %xmm0, %xmm2
; SSE-NEXT:    psrlq $32, %xmm2
; SSE-NEXT:    pmuludq %xmm1, %xmm2
; SSE-NEXT:    movdqa %xmm1, %xmm3
; SSE-NEXT:    psrlq $32, %xmm3
; SSE-NEXT:    pmuludq %xmm0, %xmm3
; SSE-NEXT:    paddq %xmm2, %xmm3
; SSE-NEXT:    psllq $32, %xmm3
; SSE-NEXT:    pmuludq %xmm1, %xmm0
; SSE-NEXT:    paddq %xmm3, %xmm0
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v2i64:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpsrlq $32, %xmm0, %xmm2
; AVX-NEXT:    vpmuludq %xmm1, %xmm2, %xmm2
; AVX-NEXT:    vpsrlq $32, %xmm1, %xmm3
; AVX-NEXT:    vpmuludq %xmm3, %xmm0, %xmm3
; AVX-NEXT:    vpaddq %xmm2, %xmm3, %xmm2
; AVX-NEXT:    vpsllq $32, %xmm2, %xmm2
; AVX-NEXT:    vpmuludq %xmm1, %xmm0, %xmm0
; AVX-NEXT:    vpaddq %xmm2, %xmm0, %xmm0
; AVX-NEXT:    retq
entry:
  %A = mul <2 x i64> %i, %j
  ret <2 x i64> %A
}

declare void @foo()

define <4 x i32> @mul_v4i32spill(<4 x i32> %i, <4 x i32> %j) nounwind  {
; SSE2-LABEL: mul_v4i32spill:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    subq $40, %rsp
; SSE2-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE2-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
; SSE2-NEXT:    callq foo
; SSE2-NEXT:    movdqa (%rsp), %xmm0 # 16-byte Reload
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[1,1,3,3]
; SSE2-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 # 16-byte Reload
; SSE2-NEXT:    pmuludq %xmm2, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm1, %xmm2
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[0,2,2,3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSE2-NEXT:    addq $40, %rsp
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v4i32spill:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    subq $40, %rsp
; SSE41-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE41-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
; SSE41-NEXT:    callq foo
; SSE41-NEXT:    movdqa (%rsp), %xmm0 # 16-byte Reload
; SSE41-NEXT:    pmulld {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; SSE41-NEXT:    addq $40, %rsp
; SSE41-NEXT:    retq
;
; AVX-LABEL: mul_v4i32spill:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    subq $40, %rsp
; AVX-NEXT:    vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX-NEXT:    vmovaps %xmm0, (%rsp) # 16-byte Spill
; AVX-NEXT:    callq foo
; AVX-NEXT:    vmovdqa (%rsp), %xmm0 # 16-byte Reload
; AVX-NEXT:    vpmulld {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; AVX-NEXT:    addq $40, %rsp
; AVX-NEXT:    retq
entry:
  ; Use a call to force spills.
  call void @foo()
  %A = mul <4 x i32> %i, %j
  ret <4 x i32> %A
}

define <2 x i64> @mul_v2i64spill(<2 x i64> %i, <2 x i64> %j) nounwind  {
; SSE-LABEL: mul_v2i64spill:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    subq $40, %rsp
; SSE-NEXT:    movaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movaps %xmm0, (%rsp) # 16-byte Spill
; SSE-NEXT:    callq foo
; SSE-NEXT:    movdqa (%rsp), %xmm0 # 16-byte Reload
; SSE-NEXT:    movdqa %xmm0, %xmm2
; SSE-NEXT:    psrlq $32, %xmm2
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Reload
; SSE-NEXT:    pmuludq %xmm3, %xmm2
; SSE-NEXT:    movdqa %xmm3, %xmm1
; SSE-NEXT:    psrlq $32, %xmm1
; SSE-NEXT:    pmuludq %xmm0, %xmm1
; SSE-NEXT:    paddq %xmm2, %xmm1
; SSE-NEXT:    psllq $32, %xmm1
; SSE-NEXT:    pmuludq %xmm3, %xmm0
; SSE-NEXT:    paddq %xmm1, %xmm0
; SSE-NEXT:    addq $40, %rsp
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v2i64spill:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    subq $40, %rsp
; AVX-NEXT:    vmovaps %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX-NEXT:    vmovaps %xmm0, (%rsp) # 16-byte Spill
; AVX-NEXT:    callq foo
; AVX-NEXT:    vmovdqa (%rsp), %xmm3 # 16-byte Reload
; AVX-NEXT:    vpsrlq $32, %xmm3, %xmm0
; AVX-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 # 16-byte Reload
; AVX-NEXT:    vpmuludq %xmm2, %xmm0, %xmm0
; AVX-NEXT:    vpsrlq $32, %xmm2, %xmm1
; AVX-NEXT:    vpmuludq %xmm1, %xmm3, %xmm1
; AVX-NEXT:    vpaddq %xmm0, %xmm1, %xmm0
; AVX-NEXT:    vpsllq $32, %xmm0, %xmm0
; AVX-NEXT:    vpmuludq %xmm2, %xmm3, %xmm1
; AVX-NEXT:    vpaddq %xmm0, %xmm1, %xmm0
; AVX-NEXT:    addq $40, %rsp
; AVX-NEXT:    retq
entry:
  ; Use a call to force spills.
  call void @foo()
  %A = mul <2 x i64> %i, %j
  ret <2 x i64> %A
}

define <32 x i8> @mul_v32i8c(<32 x i8> %i) nounwind  {
; SSE2-LABEL: mul_v32i8c:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    movdqa %xmm0, %xmm2
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa {{.*#+}} xmm3 = [117,117,117,117,117,117,117,117]
; SSE2-NEXT:    pmullw %xmm3, %xmm2
; SSE2-NEXT:    movdqa {{.*#+}} xmm4 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT:    pand %xmm4, %xmm2
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm3, %xmm0
; SSE2-NEXT:    pand %xmm4, %xmm0
; SSE2-NEXT:    packuswb %xmm2, %xmm0
; SSE2-NEXT:    movdqa %xmm1, %xmm2
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm3, %xmm2
; SSE2-NEXT:    pand %xmm4, %xmm2
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm3, %xmm1
; SSE2-NEXT:    pand %xmm4, %xmm1
; SSE2-NEXT:    packuswb %xmm2, %xmm1
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v32i8c:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm2 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    movdqa {{.*#+}} xmm4 = [117,117,117,117,117,117,117,117]
; SSE41-NEXT:    pmullw %xmm4, %xmm0
; SSE41-NEXT:    movdqa {{.*#+}} xmm5 = [255,255,255,255,255,255,255,255]
; SSE41-NEXT:    pand %xmm5, %xmm0
; SSE41-NEXT:    pmullw %xmm4, %xmm2
; SSE41-NEXT:    pand %xmm5, %xmm2
; SSE41-NEXT:    packuswb %xmm0, %xmm2
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm3 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm4, %xmm1
; SSE41-NEXT:    pand %xmm5, %xmm1
; SSE41-NEXT:    pmullw %xmm4, %xmm3
; SSE41-NEXT:    pand %xmm5, %xmm3
; SSE41-NEXT:    packuswb %xmm1, %xmm3
; SSE41-NEXT:    movdqa %xmm2, %xmm0
; SSE41-NEXT:    movdqa %xmm3, %xmm1
; SSE41-NEXT:    retq
;
; AVX2-LABEL: mul_v32i8c:
; AVX2:       # %bb.0: # %entry
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vmovdqa {{.*#+}} ymm2 = [117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117]
; AVX2-NEXT:    vpmullw %ymm2, %ymm1, %ymm1
; AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX2-NEXT:    vpand %ymm3, %ymm1, %ymm1
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpmullw %ymm2, %ymm0, %ymm0
; AVX2-NEXT:    vpand %ymm3, %ymm0, %ymm0
; AVX2-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
; AVX2-NEXT:    retq
;
; AVX512F-LABEL: mul_v32i8c:
; AVX512F:       # %bb.0: # %entry
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm1 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vmovdqa {{.*#+}} ymm2 = [117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117]
; AVX512F-NEXT:    vpmullw %ymm2, %ymm1, %ymm1
; AVX512F-NEXT:    vmovdqa {{.*#+}} ymm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX512F-NEXT:    vpand %ymm3, %ymm1, %ymm1
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpmullw %ymm2, %ymm0, %ymm0
; AVX512F-NEXT:    vpand %ymm3, %ymm0, %ymm0
; AVX512F-NEXT:    vpackuswb %ymm1, %ymm0, %ymm0
; AVX512F-NEXT:    retq
;
; AVX512BW-LABEL: mul_v32i8c:
; AVX512BW:       # %bb.0: # %entry
; AVX512BW-NEXT:    vpmovzxbw {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
; AVX512BW-NEXT:    vpmullw {{.*}}(%rip), %zmm0, %zmm0
; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
; AVX512BW-NEXT:    retq
entry:
  %A = mul <32 x i8> %i, < i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117 >
  ret <32 x i8> %A
}

define <16 x i16> @mul_v16i16c(<16 x i16> %i) nounwind  {
; SSE-LABEL: mul_v16i16c:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [117,117,117,117,117,117,117,117]
; SSE-NEXT:    pmullw %xmm2, %xmm0
; SSE-NEXT:    pmullw %xmm2, %xmm1
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v16i16c:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmullw {{.*}}(%rip), %ymm0, %ymm0
; AVX-NEXT:    retq
entry:
  %A = mul <16 x i16> %i, < i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117, i16 117 >
  ret <16 x i16> %A
}

define <8 x i32> @mul_v8i32c(<8 x i32> %i) nounwind  {
; SSE2-LABEL: mul_v8i32c:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    movdqa {{.*#+}} xmm2 = [117,117,117,117]
; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm3
; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1]
; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm1
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm3
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm3[0,2,2,3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v8i32c:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    movdqa {{.*#+}} xmm2 = [117,117,117,117]
; SSE41-NEXT:    pmulld %xmm2, %xmm0
; SSE41-NEXT:    pmulld %xmm2, %xmm1
; SSE41-NEXT:    retq
;
; AVX-LABEL: mul_v8i32c:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpbroadcastd {{.*#+}} ymm1 = [117,117,117,117,117,117,117,117]
; AVX-NEXT:    vpmulld %ymm1, %ymm0, %ymm0
; AVX-NEXT:    retq
entry:
  %A = mul <8 x i32> %i, < i32 117, i32 117, i32 117, i32 117, i32 117, i32 117, i32 117, i32 117 >
  ret <8 x i32> %A
}

define <4 x i64> @mul_v4i64c(<4 x i64> %i) nounwind  {
; SSE-LABEL: mul_v4i64c:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [117,117]
; SSE-NEXT:    movdqa %xmm0, %xmm3
; SSE-NEXT:    pmuludq %xmm2, %xmm3
; SSE-NEXT:    psrlq $32, %xmm0
; SSE-NEXT:    pmuludq %xmm2, %xmm0
; SSE-NEXT:    psllq $32, %xmm0
; SSE-NEXT:    paddq %xmm3, %xmm0
; SSE-NEXT:    movdqa %xmm1, %xmm3
; SSE-NEXT:    pmuludq %xmm2, %xmm3
; SSE-NEXT:    psrlq $32, %xmm1
; SSE-NEXT:    pmuludq %xmm2, %xmm1
; SSE-NEXT:    psllq $32, %xmm1
; SSE-NEXT:    paddq %xmm3, %xmm1
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v4i64c:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpbroadcastq {{.*#+}} ymm1 = [117,117,117,117]
; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm2
; AVX-NEXT:    vpsrlq $32, %ymm0, %ymm0
; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0
; AVX-NEXT:    vpsllq $32, %ymm0, %ymm0
; AVX-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
; AVX-NEXT:    retq
entry:
  %A = mul <4 x i64> %i, < i64 117, i64 117, i64 117, i64 117 >
  ret <4 x i64> %A
}

define <32 x i8> @mul_v32i8(<32 x i8> %i, <32 x i8> %j) nounwind  {
; SSE2-LABEL: mul_v32i8:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    movdqa %xmm2, %xmm4
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm4 = xmm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa %xmm0, %xmm5
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm5 = xmm5[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm4, %xmm5
; SSE2-NEXT:    movdqa {{.*#+}} xmm4 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT:    pand %xmm4, %xmm5
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm2, %xmm0
; SSE2-NEXT:    pand %xmm4, %xmm0
; SSE2-NEXT:    packuswb %xmm5, %xmm0
; SSE2-NEXT:    movdqa %xmm3, %xmm2
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa %xmm1, %xmm5
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm5 = xmm5[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm2, %xmm5
; SSE2-NEXT:    pand %xmm4, %xmm5
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm3, %xmm1
; SSE2-NEXT:    pand %xmm4, %xmm1
; SSE2-NEXT:    packuswb %xmm5, %xmm1
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v32i8:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm5 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm4 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm2, %xmm0
; SSE41-NEXT:    movdqa {{.*#+}} xmm6 = [255,255,255,255,255,255,255,255]
; SSE41-NEXT:    pand %xmm6, %xmm0
; SSE41-NEXT:    pmullw %xmm5, %xmm4
; SSE41-NEXT:    pand %xmm6, %xmm4
; SSE41-NEXT:    packuswb %xmm0, %xmm4
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm0 = xmm3[0],zero,xmm3[1],zero,xmm3[2],zero,xmm3[3],zero,xmm3[4],zero,xmm3[5],zero,xmm3[6],zero,xmm3[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm2 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero,xmm1[4],zero,xmm1[5],zero,xmm1[6],zero,xmm1[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm3, %xmm1
; SSE41-NEXT:    pand %xmm6, %xmm1
; SSE41-NEXT:    pmullw %xmm0, %xmm2
; SSE41-NEXT:    pand %xmm6, %xmm2
; SSE41-NEXT:    packuswb %xmm1, %xmm2
; SSE41-NEXT:    movdqa %xmm4, %xmm0
; SSE41-NEXT:    movdqa %xmm2, %xmm1
; SSE41-NEXT:    retq
;
; AVX2-LABEL: mul_v32i8:
; AVX2:       # %bb.0: # %entry
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vpmullw %ymm2, %ymm3, %ymm2
; AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX2-NEXT:    vpand %ymm3, %ymm2, %ymm2
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
; AVX2-NEXT:    vpand %ymm3, %ymm0, %ymm0
; AVX2-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
; AVX2-NEXT:    retq
;
; AVX512F-LABEL: mul_v32i8:
; AVX512F:       # %bb.0: # %entry
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vpmullw %ymm2, %ymm3, %ymm2
; AVX512F-NEXT:    vmovdqa {{.*#+}} ymm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX512F-NEXT:    vpand %ymm3, %ymm2, %ymm2
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
; AVX512F-NEXT:    vpand %ymm3, %ymm0, %ymm0
; AVX512F-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
; AVX512F-NEXT:    retq
;
; AVX512BW-LABEL: mul_v32i8:
; AVX512BW:       # %bb.0: # %entry
; AVX512BW-NEXT:    vpmovzxbw {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero,ymm1[8],zero,ymm1[9],zero,ymm1[10],zero,ymm1[11],zero,ymm1[12],zero,ymm1[13],zero,ymm1[14],zero,ymm1[15],zero,ymm1[16],zero,ymm1[17],zero,ymm1[18],zero,ymm1[19],zero,ymm1[20],zero,ymm1[21],zero,ymm1[22],zero,ymm1[23],zero,ymm1[24],zero,ymm1[25],zero,ymm1[26],zero,ymm1[27],zero,ymm1[28],zero,ymm1[29],zero,ymm1[30],zero,ymm1[31],zero
; AVX512BW-NEXT:    vpmovzxbw {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero,ymm0[8],zero,ymm0[9],zero,ymm0[10],zero,ymm0[11],zero,ymm0[12],zero,ymm0[13],zero,ymm0[14],zero,ymm0[15],zero,ymm0[16],zero,ymm0[17],zero,ymm0[18],zero,ymm0[19],zero,ymm0[20],zero,ymm0[21],zero,ymm0[22],zero,ymm0[23],zero,ymm0[24],zero,ymm0[25],zero,ymm0[26],zero,ymm0[27],zero,ymm0[28],zero,ymm0[29],zero,ymm0[30],zero,ymm0[31],zero
; AVX512BW-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
; AVX512BW-NEXT:    vpmovwb %zmm0, %ymm0
; AVX512BW-NEXT:    retq
entry:
  %A = mul <32 x i8> %i, %j
  ret <32 x i8> %A
}

define <16 x i16> @mul_v16i16(<16 x i16> %i, <16 x i16> %j) nounwind  {
; SSE-LABEL: mul_v16i16:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    pmullw %xmm2, %xmm0
; SSE-NEXT:    pmullw %xmm3, %xmm1
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v16i16:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
; AVX-NEXT:    retq
entry:
  %A = mul <16 x i16> %i, %j
  ret <16 x i16> %A
}

define <8 x i32> @mul_v8i32(<8 x i32> %i, <8 x i32> %j) nounwind  {
; SSE2-LABEL: mul_v8i32:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,2,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm4, %xmm2
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,2,3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1]
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm3, %xmm1
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,2,2,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[1,1,3,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm3
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm3[0,2,2,3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v8i32:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmulld %xmm2, %xmm0
; SSE41-NEXT:    pmulld %xmm3, %xmm1
; SSE41-NEXT:    retq
;
; AVX-LABEL: mul_v8i32:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmulld %ymm1, %ymm0, %ymm0
; AVX-NEXT:    retq
entry:
  %A = mul <8 x i32> %i, %j
  ret <8 x i32> %A
}

define <4 x i64> @mul_v4i64(<4 x i64> %i, <4 x i64> %j) nounwind  {
; SSE-LABEL: mul_v4i64:
; SSE:       # %bb.0: # %entry
; SSE-NEXT:    movdqa %xmm0, %xmm4
; SSE-NEXT:    psrlq $32, %xmm4
; SSE-NEXT:    pmuludq %xmm2, %xmm4
; SSE-NEXT:    movdqa %xmm2, %xmm5
; SSE-NEXT:    psrlq $32, %xmm5
; SSE-NEXT:    pmuludq %xmm0, %xmm5
; SSE-NEXT:    paddq %xmm4, %xmm5
; SSE-NEXT:    psllq $32, %xmm5
; SSE-NEXT:    pmuludq %xmm2, %xmm0
; SSE-NEXT:    paddq %xmm5, %xmm0
; SSE-NEXT:    movdqa %xmm1, %xmm2
; SSE-NEXT:    psrlq $32, %xmm2
; SSE-NEXT:    pmuludq %xmm3, %xmm2
; SSE-NEXT:    movdqa %xmm3, %xmm4
; SSE-NEXT:    psrlq $32, %xmm4
; SSE-NEXT:    pmuludq %xmm1, %xmm4
; SSE-NEXT:    paddq %xmm2, %xmm4
; SSE-NEXT:    psllq $32, %xmm4
; SSE-NEXT:    pmuludq %xmm3, %xmm1
; SSE-NEXT:    paddq %xmm4, %xmm1
; SSE-NEXT:    retq
;
; AVX-LABEL: mul_v4i64:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpsrlq $32, %ymm0, %ymm2
; AVX-NEXT:    vpmuludq %ymm1, %ymm2, %ymm2
; AVX-NEXT:    vpsrlq $32, %ymm1, %ymm3
; AVX-NEXT:    vpmuludq %ymm3, %ymm0, %ymm3
; AVX-NEXT:    vpaddq %ymm2, %ymm3, %ymm2
; AVX-NEXT:    vpsllq $32, %ymm2, %ymm2
; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0
; AVX-NEXT:    vpaddq %ymm2, %ymm0, %ymm0
; AVX-NEXT:    retq
entry:
  %A = mul <4 x i64> %i, %j
  ret <4 x i64> %A
}

define <64 x i8> @mul_v64i8c(<64 x i8> %i) nounwind  {
; SSE2-LABEL: mul_v64i8c:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    movdqa %xmm0, %xmm6
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm6 = xmm6[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa {{.*#+}} xmm4 = [117,117,117,117,117,117,117,117]
; SSE2-NEXT:    pmullw %xmm4, %xmm6
; SSE2-NEXT:    movdqa {{.*#+}} xmm5 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT:    pand %xmm5, %xmm6
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm4, %xmm0
; SSE2-NEXT:    pand %xmm5, %xmm0
; SSE2-NEXT:    packuswb %xmm6, %xmm0
; SSE2-NEXT:    movdqa %xmm1, %xmm6
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm6 = xmm6[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm4, %xmm6
; SSE2-NEXT:    pand %xmm5, %xmm6
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm4, %xmm1
; SSE2-NEXT:    pand %xmm5, %xmm1
; SSE2-NEXT:    packuswb %xmm6, %xmm1
; SSE2-NEXT:    movdqa %xmm2, %xmm6
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm6 = xmm6[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm4, %xmm6
; SSE2-NEXT:    pand %xmm5, %xmm6
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm4, %xmm2
; SSE2-NEXT:    pand %xmm5, %xmm2
; SSE2-NEXT:    packuswb %xmm6, %xmm2
; SSE2-NEXT:    movdqa %xmm3, %xmm6
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm6 = xmm6[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm4, %xmm6
; SSE2-NEXT:    pand %xmm5, %xmm6
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm4, %xmm3
; SSE2-NEXT:    pand %xmm5, %xmm3
; SSE2-NEXT:    packuswb %xmm6, %xmm3
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v64i8c:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    movdqa %xmm1, %xmm4
; SSE41-NEXT:    movdqa %xmm0, %xmm1
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    movdqa {{.*#+}} xmm6 = [117,117,117,117,117,117,117,117]
; SSE41-NEXT:    pmullw %xmm6, %xmm1
; SSE41-NEXT:    movdqa {{.*#+}} xmm7 = [255,255,255,255,255,255,255,255]
; SSE41-NEXT:    pand %xmm7, %xmm1
; SSE41-NEXT:    pmullw %xmm6, %xmm0
; SSE41-NEXT:    pand %xmm7, %xmm0
; SSE41-NEXT:    packuswb %xmm1, %xmm0
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm1 = xmm4[0],zero,xmm4[1],zero,xmm4[2],zero,xmm4[3],zero,xmm4[4],zero,xmm4[5],zero,xmm4[6],zero,xmm4[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm4 = xmm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm6, %xmm4
; SSE41-NEXT:    pand %xmm7, %xmm4
; SSE41-NEXT:    pmullw %xmm6, %xmm1
; SSE41-NEXT:    pand %xmm7, %xmm1
; SSE41-NEXT:    packuswb %xmm4, %xmm1
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm4 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm6, %xmm2
; SSE41-NEXT:    pand %xmm7, %xmm2
; SSE41-NEXT:    pmullw %xmm6, %xmm4
; SSE41-NEXT:    pand %xmm7, %xmm4
; SSE41-NEXT:    packuswb %xmm2, %xmm4
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm5 = xmm3[0],zero,xmm3[1],zero,xmm3[2],zero,xmm3[3],zero,xmm3[4],zero,xmm3[5],zero,xmm3[6],zero,xmm3[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm6, %xmm3
; SSE41-NEXT:    pand %xmm7, %xmm3
; SSE41-NEXT:    pmullw %xmm6, %xmm5
; SSE41-NEXT:    pand %xmm7, %xmm5
; SSE41-NEXT:    packuswb %xmm3, %xmm5
; SSE41-NEXT:    movdqa %xmm4, %xmm2
; SSE41-NEXT:    movdqa %xmm5, %xmm3
; SSE41-NEXT:    retq
;
; AVX2-LABEL: mul_v64i8c:
; AVX2:       # %bb.0: # %entry
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vmovdqa {{.*#+}} ymm3 = [117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117]
; AVX2-NEXT:    vpmullw %ymm3, %ymm2, %ymm2
; AVX2-NEXT:    vmovdqa {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX2-NEXT:    vpand %ymm4, %ymm2, %ymm2
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpmullw %ymm3, %ymm0, %ymm0
; AVX2-NEXT:    vpand %ymm4, %ymm0, %ymm0
; AVX2-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vpmullw %ymm3, %ymm2, %ymm2
; AVX2-NEXT:    vpand %ymm4, %ymm2, %ymm2
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpmullw %ymm3, %ymm1, %ymm1
; AVX2-NEXT:    vpand %ymm4, %ymm1, %ymm1
; AVX2-NEXT:    vpackuswb %ymm2, %ymm1, %ymm1
; AVX2-NEXT:    retq
;
; AVX512F-LABEL: mul_v64i8c:
; AVX512F:       # %bb.0: # %entry
; AVX512F-NEXT:    vextracti64x4 $1, %zmm0, %ymm1
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vmovdqa {{.*#+}} ymm3 = [117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117]
; AVX512F-NEXT:    vpmullw %ymm3, %ymm2, %ymm2
; AVX512F-NEXT:    vmovdqa {{.*#+}} ymm4 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX512F-NEXT:    vpand %ymm4, %ymm2, %ymm2
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpmullw %ymm3, %ymm1, %ymm1
; AVX512F-NEXT:    vpand %ymm4, %ymm1, %ymm1
; AVX512F-NEXT:    vpackuswb %ymm2, %ymm1, %ymm1
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vpmullw %ymm3, %ymm2, %ymm2
; AVX512F-NEXT:    vpand %ymm4, %ymm2, %ymm2
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpmullw %ymm3, %ymm0, %ymm0
; AVX512F-NEXT:    vpand %ymm4, %ymm0, %ymm0
; AVX512F-NEXT:    vpackuswb %ymm2, %ymm0, %ymm0
; AVX512F-NEXT:    vinserti64x4 $1, %ymm1, %zmm0, %zmm0
; AVX512F-NEXT:    retq
;
; AVX512BW-LABEL: mul_v64i8c:
; AVX512BW:       # %bb.0: # %entry
; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm1 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
; AVX512BW-NEXT:    vmovdqa64 {{.*#+}} zmm2 = [117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117,117]
; AVX512BW-NEXT:    vpmullw %zmm2, %zmm1, %zmm1
; AVX512BW-NEXT:    vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX512BW-NEXT:    vpandq %zmm3, %zmm1, %zmm1
; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
; AVX512BW-NEXT:    vpmullw %zmm2, %zmm0, %zmm0
; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
; AVX512BW-NEXT:    vpackuswb %zmm1, %zmm0, %zmm0
; AVX512BW-NEXT:    retq
entry:
  %A = mul <64 x i8> %i, < i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117, i8 117 >
  ret <64 x i8> %A
}

define <64 x i8> @mul_v64i8(<64 x i8> %i, <64 x i8> %j) nounwind  {
; SSE2-LABEL: mul_v64i8:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    movdqa %xmm4, %xmm8
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm8 = xmm8[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa %xmm0, %xmm9
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm9 = xmm9[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm8, %xmm9
; SSE2-NEXT:    movdqa {{.*#+}} xmm8 = [255,255,255,255,255,255,255,255]
; SSE2-NEXT:    pand %xmm8, %xmm9
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm4, %xmm0
; SSE2-NEXT:    pand %xmm8, %xmm0
; SSE2-NEXT:    packuswb %xmm9, %xmm0
; SSE2-NEXT:    movdqa %xmm5, %xmm9
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm9 = xmm9[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa %xmm1, %xmm4
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm4 = xmm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm9, %xmm4
; SSE2-NEXT:    pand %xmm8, %xmm4
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm5 = xmm5[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm5, %xmm1
; SSE2-NEXT:    pand %xmm8, %xmm1
; SSE2-NEXT:    packuswb %xmm4, %xmm1
; SSE2-NEXT:    movdqa %xmm6, %xmm4
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm4 = xmm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa %xmm2, %xmm5
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm5 = xmm5[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm4, %xmm5
; SSE2-NEXT:    pand %xmm8, %xmm5
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm6 = xmm6[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm6, %xmm2
; SSE2-NEXT:    pand %xmm8, %xmm2
; SSE2-NEXT:    packuswb %xmm5, %xmm2
; SSE2-NEXT:    movdqa %xmm7, %xmm4
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm4 = xmm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    movdqa %xmm3, %xmm5
; SSE2-NEXT:    punpckhbw {{.*#+}} xmm5 = xmm5[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE2-NEXT:    pmullw %xmm4, %xmm5
; SSE2-NEXT:    pand %xmm8, %xmm5
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE2-NEXT:    pmullw %xmm7, %xmm3
; SSE2-NEXT:    pand %xmm8, %xmm3
; SSE2-NEXT:    packuswb %xmm5, %xmm3
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v64i8:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    movdqa %xmm1, %xmm8
; SSE41-NEXT:    movdqa %xmm0, %xmm1
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm10 = xmm4[0],zero,xmm4[1],zero,xmm4[2],zero,xmm4[3],zero,xmm4[4],zero,xmm4[5],zero,xmm4[6],zero,xmm4[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm4 = xmm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero,xmm0[4],zero,xmm0[5],zero,xmm0[6],zero,xmm0[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm4, %xmm1
; SSE41-NEXT:    movdqa {{.*#+}} xmm9 = [255,255,255,255,255,255,255,255]
; SSE41-NEXT:    pand %xmm9, %xmm1
; SSE41-NEXT:    pmullw %xmm10, %xmm0
; SSE41-NEXT:    pand %xmm9, %xmm0
; SSE41-NEXT:    packuswb %xmm1, %xmm0
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm4 = xmm5[0],zero,xmm5[1],zero,xmm5[2],zero,xmm5[3],zero,xmm5[4],zero,xmm5[5],zero,xmm5[6],zero,xmm5[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm5 = xmm5[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm1 = xmm8[0],zero,xmm8[1],zero,xmm8[2],zero,xmm8[3],zero,xmm8[4],zero,xmm8[5],zero,xmm8[6],zero,xmm8[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm8 = xmm8[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm5, %xmm8
; SSE41-NEXT:    pand %xmm9, %xmm8
; SSE41-NEXT:    pmullw %xmm4, %xmm1
; SSE41-NEXT:    pand %xmm9, %xmm1
; SSE41-NEXT:    packuswb %xmm8, %xmm1
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm5 = xmm6[0],zero,xmm6[1],zero,xmm6[2],zero,xmm6[3],zero,xmm6[4],zero,xmm6[5],zero,xmm6[6],zero,xmm6[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm6 = xmm6[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm4 = xmm2[0],zero,xmm2[1],zero,xmm2[2],zero,xmm2[3],zero,xmm2[4],zero,xmm2[5],zero,xmm2[6],zero,xmm2[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm2 = xmm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm6, %xmm2
; SSE41-NEXT:    pand %xmm9, %xmm2
; SSE41-NEXT:    pmullw %xmm5, %xmm4
; SSE41-NEXT:    pand %xmm9, %xmm4
; SSE41-NEXT:    packuswb %xmm2, %xmm4
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm2 = xmm7[0],zero,xmm7[1],zero,xmm7[2],zero,xmm7[3],zero,xmm7[4],zero,xmm7[5],zero,xmm7[6],zero,xmm7[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm7 = xmm7[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmovzxbw {{.*#+}} xmm5 = xmm3[0],zero,xmm3[1],zero,xmm3[2],zero,xmm3[3],zero,xmm3[4],zero,xmm3[5],zero,xmm3[6],zero,xmm3[7],zero
; SSE41-NEXT:    punpckhbw {{.*#+}} xmm3 = xmm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE41-NEXT:    pmullw %xmm7, %xmm3
; SSE41-NEXT:    pand %xmm9, %xmm3
; SSE41-NEXT:    pmullw %xmm2, %xmm5
; SSE41-NEXT:    pand %xmm9, %xmm5
; SSE41-NEXT:    packuswb %xmm3, %xmm5
; SSE41-NEXT:    movdqa %xmm4, %xmm2
; SSE41-NEXT:    movdqa %xmm5, %xmm3
; SSE41-NEXT:    retq
;
; AVX2-LABEL: mul_v64i8:
; AVX2:       # %bb.0: # %entry
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm4 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm5 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vpmullw %ymm4, %ymm5, %ymm4
; AVX2-NEXT:    vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX2-NEXT:    vpand %ymm5, %ymm4, %ymm4
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpmullw %ymm2, %ymm0, %ymm0
; AVX2-NEXT:    vpand %ymm5, %ymm0, %ymm0
; AVX2-NEXT:    vpackuswb %ymm4, %ymm0, %ymm0
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm2 = ymm3[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vpunpckhbw {{.*#+}} ymm4 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX2-NEXT:    vpmullw %ymm2, %ymm4, %ymm2
; AVX2-NEXT:    vpand %ymm5, %ymm2, %ymm2
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm3 = ymm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX2-NEXT:    vpmullw %ymm3, %ymm1, %ymm1
; AVX2-NEXT:    vpand %ymm5, %ymm1, %ymm1
; AVX2-NEXT:    vpackuswb %ymm2, %ymm1, %ymm1
; AVX2-NEXT:    retq
;
; AVX512F-LABEL: mul_v64i8:
; AVX512F:       # %bb.0: # %entry
; AVX512F-NEXT:    vextracti64x4 $1, %zmm1, %ymm2
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm2[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vextracti64x4 $1, %zmm0, %ymm4
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm5 = ymm4[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vpmullw %ymm3, %ymm5, %ymm3
; AVX512F-NEXT:    vmovdqa {{.*#+}} ymm5 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX512F-NEXT:    vpand %ymm5, %ymm3, %ymm3
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm2 = ymm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm4 = ymm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpmullw %ymm2, %ymm4, %ymm2
; AVX512F-NEXT:    vpand %ymm5, %ymm2, %ymm2
; AVX512F-NEXT:    vpackuswb %ymm3, %ymm2, %ymm2
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm3 = ymm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vpunpckhbw {{.*#+}} ymm4 = ymm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31]
; AVX512F-NEXT:    vpmullw %ymm3, %ymm4, %ymm3
; AVX512F-NEXT:    vpand %ymm5, %ymm3, %ymm3
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm1 = ymm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpunpcklbw {{.*#+}} ymm0 = ymm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23]
; AVX512F-NEXT:    vpmullw %ymm1, %ymm0, %ymm0
; AVX512F-NEXT:    vpand %ymm5, %ymm0, %ymm0
; AVX512F-NEXT:    vpackuswb %ymm3, %ymm0, %ymm0
; AVX512F-NEXT:    vinserti64x4 $1, %ymm2, %zmm0, %zmm0
; AVX512F-NEXT:    retq
;
; AVX512BW-LABEL: mul_v64i8:
; AVX512BW:       # %bb.0: # %entry
; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm2 = zmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
; AVX512BW-NEXT:    vpunpckhbw {{.*#+}} zmm3 = zmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15,24,24,25,25,26,26,27,27,28,28,29,29,30,30,31,31,40,40,41,41,42,42,43,43,44,44,45,45,46,46,47,47,56,56,57,57,58,58,59,59,60,60,61,61,62,62,63,63]
; AVX512BW-NEXT:    vpmullw %zmm2, %zmm3, %zmm2
; AVX512BW-NEXT:    vmovdqa64 {{.*#+}} zmm3 = [255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255]
; AVX512BW-NEXT:    vpandq %zmm3, %zmm2, %zmm2
; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm1 = zmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
; AVX512BW-NEXT:    vpunpcklbw {{.*#+}} zmm0 = zmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7,16,16,17,17,18,18,19,19,20,20,21,21,22,22,23,23,32,32,33,33,34,34,35,35,36,36,37,37,38,38,39,39,48,48,49,49,50,50,51,51,52,52,53,53,54,54,55,55]
; AVX512BW-NEXT:    vpmullw %zmm1, %zmm0, %zmm0
; AVX512BW-NEXT:    vpandq %zmm3, %zmm0, %zmm0
; AVX512BW-NEXT:    vpackuswb %zmm2, %zmm0, %zmm0
; AVX512BW-NEXT:    retq
entry:
  %A = mul <64 x i8> %i, %j
  ret <64 x i8> %A
}

; PR30845
define <4 x i32> @mul_v4i64_zero_upper(<4 x i32> %val1, <4 x i32> %val2) {
; SSE2-LABEL: mul_v4i64_zero_upper:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[0,1,1,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[2,1,3,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[2,1,3,3]
; SSE2-NEXT:    pmuludq %xmm3, %xmm1
; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v4i64_zero_upper:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm2 = xmm0[0],zero,xmm0[1],zero
; SSE41-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[2,1,3,3]
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm0 = xmm1[0],zero,xmm1[1],zero
; SSE41-NEXT:    pmuludq %xmm2, %xmm0
; SSE41-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[2,1,3,3]
; SSE41-NEXT:    pmuludq %xmm3, %xmm1
; SSE41-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; SSE41-NEXT:    retq
;
; AVX-LABEL: mul_v4i64_zero_upper:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmovzxdq {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; AVX-NEXT:    vpmovzxdq {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0
; AVX-NEXT:    vextracti128 $1, %ymm0, %xmm1
; AVX-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; AVX-NEXT:    vzeroupper
; AVX-NEXT:    retq
entry:
  %val1a = zext <4 x i32> %val1 to <4 x i64>
  %val2a = zext <4 x i32> %val2 to <4 x i64>
  %res64 = mul <4 x i64> %val1a, %val2a
  %rescast = bitcast <4 x i64> %res64 to <8 x i32>
  %res = shufflevector <8 x i32> %rescast, <8 x i32> undef, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
  ret <4 x i32> %res
}

define <4 x i32> @mul_v4i64_zero_upper_left(<4 x i32> %val1, <4 x i64> %val2) {
; SSE2-LABEL: mul_v4i64_zero_upper_left:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    pxor %xmm3, %xmm3
; SSE2-NEXT:    movdqa %xmm0, %xmm4
; SSE2-NEXT:    punpckldq {{.*#+}} xmm4 = xmm4[0],xmm3[0],xmm4[1],xmm3[1]
; SSE2-NEXT:    punpckhdq {{.*#+}} xmm0 = xmm0[2],xmm3[2],xmm0[3],xmm3[3]
; SSE2-NEXT:    movdqa %xmm0, %xmm3
; SSE2-NEXT:    pmuludq %xmm2, %xmm3
; SSE2-NEXT:    psrlq $32, %xmm2
; SSE2-NEXT:    pmuludq %xmm0, %xmm2
; SSE2-NEXT:    psllq $32, %xmm2
; SSE2-NEXT:    paddq %xmm3, %xmm2
; SSE2-NEXT:    movdqa %xmm4, %xmm0
; SSE2-NEXT:    pmuludq %xmm1, %xmm0
; SSE2-NEXT:    psrlq $32, %xmm1
; SSE2-NEXT:    pmuludq %xmm4, %xmm1
; SSE2-NEXT:    psllq $32, %xmm1
; SSE2-NEXT:    paddq %xmm1, %xmm0
; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,3],xmm2[1,3]
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v4i64_zero_upper_left:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pxor %xmm3, %xmm3
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm4 = xmm0[0],zero,xmm0[1],zero
; SSE41-NEXT:    punpckhdq {{.*#+}} xmm0 = xmm0[2],xmm3[2],xmm0[3],xmm3[3]
; SSE41-NEXT:    movdqa %xmm0, %xmm3
; SSE41-NEXT:    pmuludq %xmm2, %xmm3
; SSE41-NEXT:    psrlq $32, %xmm2
; SSE41-NEXT:    pmuludq %xmm0, %xmm2
; SSE41-NEXT:    psllq $32, %xmm2
; SSE41-NEXT:    paddq %xmm3, %xmm2
; SSE41-NEXT:    movdqa %xmm4, %xmm0
; SSE41-NEXT:    pmuludq %xmm1, %xmm0
; SSE41-NEXT:    psrlq $32, %xmm1
; SSE41-NEXT:    pmuludq %xmm4, %xmm1
; SSE41-NEXT:    psllq $32, %xmm1
; SSE41-NEXT:    paddq %xmm1, %xmm0
; SSE41-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,3],xmm2[1,3]
; SSE41-NEXT:    retq
;
; AVX-LABEL: mul_v4i64_zero_upper_left:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmovzxdq {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm2
; AVX-NEXT:    vpsrlq $32, %ymm1, %ymm1
; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0
; AVX-NEXT:    vpsllq $32, %ymm0, %ymm0
; AVX-NEXT:    vpaddq %ymm0, %ymm2, %ymm0
; AVX-NEXT:    vextracti128 $1, %ymm0, %xmm1
; AVX-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; AVX-NEXT:    vzeroupper
; AVX-NEXT:    retq
entry:
  %val1a = zext <4 x i32> %val1 to <4 x i64>
  %res64 = mul <4 x i64> %val1a, %val2
  %rescast = bitcast <4 x i64> %res64 to <8 x i32>
  %res = shufflevector <8 x i32> %rescast, <8 x i32> undef, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
  ret <4 x i32> %res
}

define <4 x i32> @mul_v4i64_zero_lower(<4 x i32> %val1, <4 x i64> %val2) {
; SSE2-LABEL: mul_v4i64_zero_lower:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[0,1,1,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,1,3,3]
; SSE2-NEXT:    psrlq $32, %xmm2
; SSE2-NEXT:    pmuludq %xmm0, %xmm2
; SSE2-NEXT:    psrlq $32, %xmm1
; SSE2-NEXT:    pmuludq %xmm1, %xmm3
; SSE2-NEXT:    shufps {{.*#+}} xmm3 = xmm3[0,2],xmm2[0,2]
; SSE2-NEXT:    movaps %xmm3, %xmm0
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v4i64_zero_lower:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[2,1,3,3]
; SSE41-NEXT:    psrlq $32, %xmm2
; SSE41-NEXT:    pmuludq %xmm3, %xmm2
; SSE41-NEXT:    psrlq $32, %xmm1
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero
; SSE41-NEXT:    pmuludq %xmm1, %xmm0
; SSE41-NEXT:    shufps {{.*#+}} xmm0 = xmm0[0,2],xmm2[0,2]
; SSE41-NEXT:    retq
;
; AVX-LABEL: mul_v4i64_zero_lower:
; AVX:       # %bb.0: # %entry
; AVX-NEXT:    vpmovzxdq {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; AVX-NEXT:    vpsrlq $32, %ymm1, %ymm1
; AVX-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0
; AVX-NEXT:    vpsllq $32, %ymm0, %ymm0
; AVX-NEXT:    vextracti128 $1, %ymm0, %xmm1
; AVX-NEXT:    vshufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; AVX-NEXT:    vzeroupper
; AVX-NEXT:    retq
entry:
  %val1a = zext <4 x i32> %val1 to <4 x i64>
  %val2a = and <4 x i64> %val2, <i64 -4294967296, i64 -4294967296, i64 -4294967296, i64 -4294967296>
  %res64 = mul <4 x i64> %val1a, %val2a
  %rescast = bitcast <4 x i64> %res64 to <8 x i32>
  %res = shufflevector <8 x i32> %rescast, <8 x i32> undef, <4 x i32> <i32 1, i32 3, i32 5, i32 7>
  ret <4 x i32> %res
}

define <8 x i32> @mul_v8i64_zero_upper(<8 x i32> %val1, <8 x i32> %val2) {
; SSE2-LABEL: mul_v8i64_zero_upper:
; SSE2:       # %bb.0: # %entry
; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,1,1,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[2,1,3,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[0,1,1,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm7 = xmm1[2,1,3,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm0 = xmm2[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm4, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[2,1,3,3]
; SSE2-NEXT:    pmuludq %xmm5, %xmm1
; SSE2-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm3[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm6, %xmm1
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm3[2,1,3,3]
; SSE2-NEXT:    pmuludq %xmm7, %xmm2
; SSE2-NEXT:    shufps {{.*#+}} xmm1 = xmm1[1,3],xmm2[1,3]
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v8i64_zero_upper:
; SSE41:       # %bb.0: # %entry
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm4 = xmm0[0],zero,xmm0[1],zero
; SSE41-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[2,1,3,3]
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm6 = xmm1[0],zero,xmm1[1],zero
; SSE41-NEXT:    pshufd {{.*#+}} xmm7 = xmm1[2,1,3,3]
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm0 = xmm2[0],zero,xmm2[1],zero
; SSE41-NEXT:    pmuludq %xmm4, %xmm0
; SSE41-NEXT:    pshufd {{.*#+}} xmm1 = xmm2[2,1,3,3]
; SSE41-NEXT:    pmuludq %xmm5, %xmm1
; SSE41-NEXT:    shufps {{.*#+}} xmm0 = xmm0[1,3],xmm1[1,3]
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm1 = xmm3[0],zero,xmm3[1],zero
; SSE41-NEXT:    pmuludq %xmm6, %xmm1
; SSE41-NEXT:    pshufd {{.*#+}} xmm2 = xmm3[2,1,3,3]
; SSE41-NEXT:    pmuludq %xmm7, %xmm2
; SSE41-NEXT:    shufps {{.*#+}} xmm1 = xmm1[1,3],xmm2[1,3]
; SSE41-NEXT:    retq
;
; AVX2-LABEL: mul_v8i64_zero_upper:
; AVX2:       # %bb.0: # %entry
; AVX2-NEXT:    vpmovzxdq {{.*#+}} ymm2 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; AVX2-NEXT:    vextracti128 $1, %ymm0, %xmm0
; AVX2-NEXT:    vpmovzxdq {{.*#+}} ymm0 = xmm0[0],zero,xmm0[1],zero,xmm0[2],zero,xmm0[3],zero
; AVX2-NEXT:    vpmovzxdq {{.*#+}} ymm3 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
; AVX2-NEXT:    vpmuludq %ymm3, %ymm2, %ymm2
; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm1
; AVX2-NEXT:    vpmovzxdq {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
; AVX2-NEXT:    vpmuludq %ymm1, %ymm0, %ymm0
; AVX2-NEXT:    vshufps {{.*#+}} ymm0 = ymm2[1,3],ymm0[1,3],ymm2[5,7],ymm0[5,7]
; AVX2-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,2,1,3]
; AVX2-NEXT:    retq
;
; AVX512-LABEL: mul_v8i64_zero_upper:
; AVX512:       # %bb.0: # %entry
; AVX512-NEXT:    vpmovzxdq {{.*#+}} zmm0 = ymm0[0],zero,ymm0[1],zero,ymm0[2],zero,ymm0[3],zero,ymm0[4],zero,ymm0[5],zero,ymm0[6],zero,ymm0[7],zero
; AVX512-NEXT:    vpmovzxdq {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero
; AVX512-NEXT:    vpmuludq %zmm1, %zmm0, %zmm0
; AVX512-NEXT:    vextracti64x4 $1, %zmm0, %ymm1
; AVX512-NEXT:    vshufps {{.*#+}} ymm0 = ymm0[1,3],ymm1[1,3],ymm0[5,7],ymm1[5,7]
; AVX512-NEXT:    vpermpd {{.*#+}} ymm0 = ymm0[0,2,1,3]
; AVX512-NEXT:    retq
entry:
  %val1a = zext <8 x i32> %val1 to <8 x i64>
  %val2a = zext <8 x i32> %val2 to <8 x i64>
  %res64 = mul <8 x i64> %val1a, %val2a
  %rescast = bitcast <8 x i64> %res64 to <16 x i32>
  %res = shufflevector <16 x i32> %rescast, <16 x i32> undef, <8 x i32> <i32 1, i32 3, i32 5, i32 7,i32 9, i32 11, i32 13, i32 15 >
  ret <8 x i32> %res
}

define <8 x i64> @mul_v8i64_sext(<8 x i16> %val1, <8 x i32> %val2) {
; SSE2-LABEL: mul_v8i64_sext:
; SSE2:       # %bb.0:
; SSE2-NEXT:    movdqa %xmm1, %xmm15
; SSE2-NEXT:    punpckhwd {{.*#+}} xmm14 = xmm14[4],xmm0[4],xmm14[5],xmm0[5],xmm14[6],xmm0[6],xmm14[7],xmm0[7]
; SSE2-NEXT:    psrad $16, %xmm14
; SSE2-NEXT:    pxor %xmm13, %xmm13
; SSE2-NEXT:    pxor %xmm10, %xmm10
; SSE2-NEXT:    pcmpgtd %xmm14, %xmm10
; SSE2-NEXT:    movdqa %xmm14, %xmm8
; SSE2-NEXT:    punpckhdq {{.*#+}} xmm8 = xmm8[2],xmm10[2],xmm8[3],xmm10[3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm14 = xmm14[0],xmm10[0],xmm14[1],xmm10[1]
; SSE2-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3]
; SSE2-NEXT:    psrad $16, %xmm0
; SSE2-NEXT:    pxor %xmm5, %xmm5
; SSE2-NEXT:    pcmpgtd %xmm0, %xmm5
; SSE2-NEXT:    movdqa %xmm0, %xmm11
; SSE2-NEXT:    punpckhdq {{.*#+}} xmm11 = xmm11[2],xmm5[2],xmm11[3],xmm5[3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm5[0],xmm0[1],xmm5[1]
; SSE2-NEXT:    pshufd {{.*#+}} xmm3 = xmm2[2,3,2,3]
; SSE2-NEXT:    pxor %xmm9, %xmm9
; SSE2-NEXT:    pcmpgtd %xmm3, %xmm9
; SSE2-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm9[0],xmm3[1],xmm9[1]
; SSE2-NEXT:    pxor %xmm12, %xmm12
; SSE2-NEXT:    pcmpgtd %xmm2, %xmm12
; SSE2-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],xmm12[0],xmm2[1],xmm12[1]
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[2,3,2,3]
; SSE2-NEXT:    pxor %xmm7, %xmm7
; SSE2-NEXT:    pcmpgtd %xmm1, %xmm7
; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm7[0],xmm1[1],xmm7[1]
; SSE2-NEXT:    pcmpgtd %xmm15, %xmm13
; SSE2-NEXT:    punpckldq {{.*#+}} xmm15 = xmm15[0],xmm13[0],xmm15[1],xmm13[1]
; SSE2-NEXT:    pshufd {{.*#+}} xmm6 = xmm5[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm15, %xmm6
; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm13[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm0, %xmm4
; SSE2-NEXT:    paddq %xmm6, %xmm4
; SSE2-NEXT:    psllq $32, %xmm4
; SSE2-NEXT:    pmuludq %xmm15, %xmm0
; SSE2-NEXT:    paddq %xmm4, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm5[2,1,3,3]
; SSE2-NEXT:    pmuludq %xmm1, %xmm4
; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm7[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm11, %xmm5
; SSE2-NEXT:    paddq %xmm4, %xmm5
; SSE2-NEXT:    psllq $32, %xmm5
; SSE2-NEXT:    pmuludq %xmm11, %xmm1
; SSE2-NEXT:    paddq %xmm5, %xmm1
; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm10[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm2, %xmm4
; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm12[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm14, %xmm5
; SSE2-NEXT:    paddq %xmm4, %xmm5
; SSE2-NEXT:    psllq $32, %xmm5
; SSE2-NEXT:    pmuludq %xmm14, %xmm2
; SSE2-NEXT:    paddq %xmm5, %xmm2
; SSE2-NEXT:    pshufd {{.*#+}} xmm4 = xmm10[2,1,3,3]
; SSE2-NEXT:    pmuludq %xmm3, %xmm4
; SSE2-NEXT:    pshufd {{.*#+}} xmm5 = xmm9[0,1,1,3]
; SSE2-NEXT:    pmuludq %xmm8, %xmm5
; SSE2-NEXT:    paddq %xmm4, %xmm5
; SSE2-NEXT:    psllq $32, %xmm5
; SSE2-NEXT:    pmuludq %xmm8, %xmm3
; SSE2-NEXT:    paddq %xmm5, %xmm3
; SSE2-NEXT:    retq
;
; SSE41-LABEL: mul_v8i64_sext:
; SSE41:       # %bb.0:
; SSE41-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[3,3,3,3]
; SSE41-NEXT:    pmovsxwq %xmm3, %xmm4
; SSE41-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[2,3,2,3]
; SSE41-NEXT:    pmovsxwq %xmm3, %xmm5
; SSE41-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[1,1,1,1]
; SSE41-NEXT:    pmovsxwq %xmm3, %xmm6
; SSE41-NEXT:    pmovsxwq %xmm0, %xmm7
; SSE41-NEXT:    pshufd {{.*#+}} xmm3 = xmm2[2,1,3,3]
; SSE41-NEXT:    pmuldq %xmm4, %xmm3
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm2 = xmm2[0],zero,xmm2[1],zero
; SSE41-NEXT:    pmuldq %xmm5, %xmm2
; SSE41-NEXT:    pshufd {{.*#+}} xmm4 = xmm1[2,1,3,3]
; SSE41-NEXT:    pmuldq %xmm6, %xmm4
; SSE41-NEXT:    pmovzxdq {{.*#+}} xmm0 = xmm1[0],zero,xmm1[1],zero
; SSE41-NEXT:    pmuldq %xmm7, %xmm0
; SSE41-NEXT:    movdqa %xmm4, %xmm1
; SSE41-NEXT:    retq
;
; AVX2-LABEL: mul_v8i64_sext:
; AVX2:       # %bb.0:
; AVX2-NEXT:    vpshufd {{.*#+}} xmm2 = xmm0[2,3,2,3]
; AVX2-NEXT:    vpmovsxwq %xmm2, %ymm2
; AVX2-NEXT:    vpmovsxwq %xmm0, %ymm0
; AVX2-NEXT:    vextracti128 $1, %ymm1, %xmm3
; AVX2-NEXT:    vpmovzxdq {{.*#+}} ymm3 = xmm3[0],zero,xmm3[1],zero,xmm3[2],zero,xmm3[3],zero
; AVX2-NEXT:    vpmuldq %ymm3, %ymm2, %ymm2
; AVX2-NEXT:    vpmovzxdq {{.*#+}} ymm1 = xmm1[0],zero,xmm1[1],zero,xmm1[2],zero,xmm1[3],zero
; AVX2-NEXT:    vpmuldq %ymm1, %ymm0, %ymm0
; AVX2-NEXT:    vmovdqa %ymm2, %ymm1
; AVX2-NEXT:    retq
;
; AVX512-LABEL: mul_v8i64_sext:
; AVX512:       # %bb.0:
; AVX512-NEXT:    vpmovsxwq %xmm0, %zmm0
; AVX512-NEXT:    vpmovzxdq {{.*#+}} zmm1 = ymm1[0],zero,ymm1[1],zero,ymm1[2],zero,ymm1[3],zero,ymm1[4],zero,ymm1[5],zero,ymm1[6],zero,ymm1[7],zero
; AVX512-NEXT:    vpmuldq %zmm1, %zmm0, %zmm0
; AVX512-NEXT:    retq
  %1 = sext <8 x i16> %val1 to <8 x i64>
  %2 = sext <8 x i32> %val2 to <8 x i64>
  %3 = mul <8 x i64> %1, %2
  ret <8 x i64> %3
}

define <2 x i64> @pmuldq_square(<2 x i64> %x) {
; SSE2-LABEL: pmuldq_square:
; SSE2:       # %bb.0:
; SSE2-NEXT:    pshufd {{.*#+}} xmm1 = xmm0[0,2,2,3]
; SSE2-NEXT:    psllq $32, %xmm0
; SSE2-NEXT:    psrad $31, %xmm0
; SSE2-NEXT:    pshufd {{.*#+}} xmm2 = xmm0[1,3,2,3]
; SSE2-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1]
; SSE2-NEXT:    psrlq $32, %xmm0
; SSE2-NEXT:    pmuludq %xmm1, %xmm0
; SSE2-NEXT:    pmuludq %xmm1, %xmm1
; SSE2-NEXT:    paddq %xmm0, %xmm0
; SSE2-NEXT:    psllq $32, %xmm0
; SSE2-NEXT:    paddq %xmm1, %xmm0
; SSE2-NEXT:    retq
;
; SSE41-LABEL: pmuldq_square:
; SSE41:       # %bb.0:
; SSE41-NEXT:    pmuldq %xmm0, %xmm0
; SSE41-NEXT:    retq
;
; AVX-LABEL: pmuldq_square:
; AVX:       # %bb.0:
; AVX-NEXT:    vpmuldq %xmm0, %xmm0, %xmm0
; AVX-NEXT:    retq
  %1 = shl <2 x i64> %x, <i64 32, i64 32>
  %2 = ashr exact <2 x i64> %1, <i64 32, i64 32>
  %3 = mul nsw <2 x i64> %2, %2
  ret <2 x i64> %3
}

define <2 x i64> @pmuludq_square(<2 x i64> %x) {
; SSE-LABEL: pmuludq_square:
; SSE:       # %bb.0:
; SSE-NEXT:    pmuludq %xmm0, %xmm0
; SSE-NEXT:    retq
;
; AVX-LABEL: pmuludq_square:
; AVX:       # %bb.0:
; AVX-NEXT:    vpmuludq %xmm0, %xmm0, %xmm0
; AVX-NEXT:    retq
  %1 = and <2 x i64> %x, <i64 4294967295, i64 4294967295>
  %2 = mul nuw <2 x i64> %1, %1
  ret <2 x i64> %2
}