Run DCE after a LoopFlatten test to reduce spurious output [nfc]

[llvm-project.git] / llvm / test / CodeGen / X86 / vector-interleaved-store-i8-stride-5.ll

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-- -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,FALLBACK0
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx  | FileCheck %s --check-prefixes=AVX,AVX1,AVX1-ONLY,FALLBACK1
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx2 | FileCheck %s --check-prefixes=AVX,AVX1,AVX2,AVX2-ONLY,AVX2-SLOW,FALLBACK2
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx2,+fast-variable-crosslane-shuffle,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=AVX,AVX1,AVX2,AVX2-ONLY,AVX2-FAST,FALLBACK3
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx2,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=AVX,AVX1,AVX2,AVX2-ONLY,AVX2-FAST-PERLANE,FALLBACK4
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512vl | FileCheck %s --check-prefixes=AVX,AVX2,AVX512,AVX512F,AVX512-SLOW,AVX512F-SLOW,AVX512F-ONLY-SLOW,FALLBACK5
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512vl,+fast-variable-crosslane-shuffle,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=AVX,AVX2,AVX512,AVX512F,AVX512-FAST,AVX512F-FAST,AVX512F-ONLY-FAST,FALLBACK6
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512vl,+avx512dq | FileCheck %s --check-prefixes=AVX,AVX2,AVX512,AVX512F,AVX512-SLOW,AVX512F-SLOW,AVX512DQ-SLOW,FALLBACK7
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512vl,+avx512dq,+fast-variable-crosslane-shuffle,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=AVX,AVX2,AVX512,AVX512F,AVX512-FAST,AVX512F-FAST,AVX512DQ-FAST,FALLBACK8
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512vl,+avx512bw | FileCheck %s --check-prefixes=AVX,AVX2,AVX512,AVX512BW,AVX512-SLOW,AVX512BW-SLOW,AVX512BW-ONLY-SLOW,FALLBACK9
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512vl,+avx512bw,+fast-variable-crosslane-shuffle,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=AVX,AVX2,AVX512,AVX512BW,AVX512-FAST,AVX512BW-FAST,AVX512BW-ONLY-FAST,FALLBACK10
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512vl,+avx512dq,+avx512bw | FileCheck %s --check-prefixes=AVX,AVX2,AVX512,AVX512BW,AVX512-SLOW,AVX512BW-SLOW,AVX512DQBW-SLOW,FALLBACK11
; RUN: llc < %s -mtriple=x86_64-- -mattr=+avx512vl,+avx512dq,+avx512bw,+fast-variable-crosslane-shuffle,+fast-variable-perlane-shuffle | FileCheck %s --check-prefixes=AVX,AVX2,AVX512,AVX512BW,AVX512-FAST,AVX512BW-FAST,AVX512DQBW-FAST,FALLBACK12

; These patterns are produced by LoopVectorizer for interleaved stores.

define void @store_i8_stride5_vf2(ptr %in.vecptr0, ptr %in.vecptr1, ptr %in.vecptr2, ptr %in.vecptr3, ptr %in.vecptr4, ptr %out.vec) nounwind {
; SSE-LABEL: store_i8_stride5_vf2:
; SSE:       # %bb.0:
; SSE-NEXT:    movdqa (%rdi), %xmm0
; SSE-NEXT:    movdqa (%rdx), %xmm1
; SSE-NEXT:    punpcklwd {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1],xmm0[2],mem[2],xmm0[3],mem[3]
; SSE-NEXT:    punpcklwd {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1],xmm1[2],mem[2],xmm1[3],mem[3]
; SSE-NEXT:    punpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; SSE-NEXT:    pxor %xmm1, %xmm1
; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3],xmm0[4],xmm1[4],xmm0[5],xmm1[5],xmm0[6],xmm1[6],xmm0[7],xmm1[7]
; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm0[0,2,1,3,4,5,6,7]
; SSE-NEXT:    psrldq {{.*#+}} xmm0 = xmm0[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,5,6,4]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,3,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[0,1,3,2,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,6,7,5]
; SSE-NEXT:    packuswb %xmm0, %xmm1
; SSE-NEXT:    movdqa {{.*#+}} xmm0 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,255,255]
; SSE-NEXT:    pand %xmm0, %xmm1
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = mem[0,0,0,0]
; SSE-NEXT:    pandn %xmm2, %xmm0
; SSE-NEXT:    por %xmm1, %xmm0
; SSE-NEXT:    movq %xmm0, (%r9)
; SSE-NEXT:    pextrw $4, %xmm0, %eax
; SSE-NEXT:    movw %ax, 8(%r9)
; SSE-NEXT:    retq
;
; AVX-LABEL: store_i8_stride5_vf2:
; AVX:       # %bb.0:
; AVX-NEXT:    vmovdqa (%rdi), %xmm0
; AVX-NEXT:    vmovdqa (%rdx), %xmm1
; AVX-NEXT:    vpunpcklwd {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1],xmm0[2],mem[2],xmm0[3],mem[3]
; AVX-NEXT:    vpunpcklwd {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1],xmm1[2],mem[2],xmm1[3],mem[3]
; AVX-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1]
; AVX-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1],xmm0[2],mem[2],xmm0[3],mem[3],xmm0[4],mem[4],xmm0[5],mem[5],xmm0[6],mem[6],xmm0[7],mem[7]
; AVX-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,1,2,6,10,14,3,u,u,u,u,u,u]
; AVX-NEXT:    vpextrw $4, %xmm0, 8(%r9)
; AVX-NEXT:    vmovq %xmm0, (%r9)
; AVX-NEXT:    retq
  %in.vec0 = load <2 x i8>, ptr %in.vecptr0, align 64
  %in.vec1 = load <2 x i8>, ptr %in.vecptr1, align 64
  %in.vec2 = load <2 x i8>, ptr %in.vecptr2, align 64
  %in.vec3 = load <2 x i8>, ptr %in.vecptr3, align 64
  %in.vec4 = load <2 x i8>, ptr %in.vecptr4, align 64
  %1 = shufflevector <2 x i8> %in.vec0, <2 x i8> %in.vec1, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %2 = shufflevector <2 x i8> %in.vec2, <2 x i8> %in.vec3, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  %3 = shufflevector <4 x i8> %1, <4 x i8> %2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %4 = shufflevector <2 x i8> %in.vec4, <2 x i8> poison, <8 x i32> <i32 0, i32 1, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5 = shufflevector <8 x i8> %3, <8 x i8> %4, <10 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7, i32 8, i32 9>
  %interleaved.vec = shufflevector <10 x i8> %5, <10 x i8> poison, <10 x i32> <i32 0, i32 2, i32 4, i32 6, i32 8, i32 1, i32 3, i32 5, i32 7, i32 9>
  store <10 x i8> %interleaved.vec, ptr %out.vec, align 64
  ret void
}

define void @store_i8_stride5_vf4(ptr %in.vecptr0, ptr %in.vecptr1, ptr %in.vecptr2, ptr %in.vecptr3, ptr %in.vecptr4, ptr %out.vec) nounwind {
; SSE-LABEL: store_i8_stride5_vf4:
; SSE:       # %bb.0:
; SSE-NEXT:    movdqa (%rdi), %xmm1
; SSE-NEXT:    movdqa (%rdx), %xmm2
; SSE-NEXT:    movdqa (%r8), %xmm0
; SSE-NEXT:    punpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1]
; SSE-NEXT:    punpckldq {{.*#+}} xmm2 = xmm2[0],mem[0],xmm2[1],mem[1]
; SSE-NEXT:    pxor %xmm3, %xmm3
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm1[3,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[0,1,2,0,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[1,3,2,3]
; SSE-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3],xmm2[4],xmm3[4],xmm2[5],xmm3[5],xmm2[6],xmm3[6],xmm2[7],xmm3[7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm2[2,1,1,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm3 = xmm3[1,1,1,1,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,4,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,2,2,3]
; SSE-NEXT:    punpckldq {{.*#+}} xmm3 = xmm3[0],xmm4[0],xmm3[1],xmm4[1]
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm2[0,2,0,0]
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[0,1,0,2,4,5,6,7]
; SSE-NEXT:    movdqa {{.*#+}} xmm5 = [65535,65535,0,0,65535,65535,65535,0]
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[0,2,2,0]
; SSE-NEXT:    pshuflw {{.*#+}} xmm6 = xmm6[0,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm6 = xmm6[0,1,2,3,4,7,5,7]
; SSE-NEXT:    pand %xmm5, %xmm6
; SSE-NEXT:    pandn %xmm4, %xmm5
; SSE-NEXT:    por %xmm6, %xmm5
; SSE-NEXT:    packuswb %xmm3, %xmm5
; SSE-NEXT:    movdqa {{.*#+}} xmm3 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; SSE-NEXT:    pand %xmm3, %xmm5
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,0,0,0]
; SSE-NEXT:    pandn %xmm4, %xmm3
; SSE-NEXT:    por %xmm5, %xmm3
; SSE-NEXT:    psrldq {{.*#+}} xmm1 = xmm1[14,15],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[3,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm2[0,3,1,3,4,5,6,7]
; SSE-NEXT:    pand {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm2
; SSE-NEXT:    por %xmm1, %xmm2
; SSE-NEXT:    packuswb %xmm2, %xmm2
; SSE-NEXT:    movdqa {{.*#+}} xmm1 = [255,255,255,0,255,255,255,255,255,255,255,255,255,255,255,255]
; SSE-NEXT:    pand %xmm1, %xmm2
; SSE-NEXT:    pandn %xmm0, %xmm1
; SSE-NEXT:    por %xmm2, %xmm1
; SSE-NEXT:    movd %xmm1, 16(%r9)
; SSE-NEXT:    movdqa %xmm3, (%r9)
; SSE-NEXT:    retq
;
; AVX1-ONLY-LABEL: store_i8_stride5_vf4:
; AVX1-ONLY:       # %bb.0:
; AVX1-ONLY-NEXT:    vmovdqa (%rdi), %xmm0
; AVX1-ONLY-NEXT:    vmovdqa (%rdx), %xmm1
; AVX1-ONLY-NEXT:    vmovdqa (%r8), %xmm2
; AVX1-ONLY-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
; AVX1-ONLY-NEXT:    vpunpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1]
; AVX1-ONLY-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm1 = xmm0[0,4,8,12],zero,xmm0[1,5,9,13],zero,xmm0[2,6,10,14],zero,xmm0[3]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = zero,zero,zero,zero,xmm2[0],zero,zero,zero,zero,xmm2[1],zero,zero,zero,zero,xmm2[2],zero
; AVX1-ONLY-NEXT:    vpor %xmm3, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vpalignr {{.*#+}} xmm0 = xmm0[7,8,9,10,11,12,13,14,15],xmm2[0,1,2,3,4,5,6]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,4,8,12,u,u,u,u,u,u,u,u,u,u,u,u]
; AVX1-ONLY-NEXT:    vmovd %xmm0, 16(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm1, (%r9)
; AVX1-ONLY-NEXT:    retq
;
; AVX2-ONLY-LABEL: store_i8_stride5_vf4:
; AVX2-ONLY:       # %bb.0:
; AVX2-ONLY-NEXT:    vmovdqa (%rdi), %xmm0
; AVX2-ONLY-NEXT:    vmovdqa (%rdx), %xmm1
; AVX2-ONLY-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
; AVX2-ONLY-NEXT:    vpunpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1]
; AVX2-ONLY-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX2-ONLY-NEXT:    vinserti128 $1, (%r8), %ymm0, %ymm0
; AVX2-ONLY-NEXT:    vpshufb {{.*#+}} ymm1 = ymm0[0,4,8,12],zero,ymm0[1,5,9,13],zero,ymm0[2,6,10,14],zero,ymm0[3],zero,zero,zero,ymm0[19],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; AVX2-ONLY-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,3,0,1]
; AVX2-ONLY-NEXT:    vpshufb {{.*#+}} ymm0 = zero,zero,zero,zero,ymm0[0],zero,zero,zero,zero,ymm0[1],zero,zero,zero,zero,ymm0[2],zero,ymm0[23,27,31],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; AVX2-ONLY-NEXT:    vpor %ymm0, %ymm1, %ymm0
; AVX2-ONLY-NEXT:    vextracti128 $1, %ymm0, %xmm1
; AVX2-ONLY-NEXT:    vmovd %xmm1, 16(%r9)
; AVX2-ONLY-NEXT:    vmovdqa %xmm0, (%r9)
; AVX2-ONLY-NEXT:    vzeroupper
; AVX2-ONLY-NEXT:    retq
;
; AVX512F-LABEL: store_i8_stride5_vf4:
; AVX512F:       # %bb.0:
; AVX512F-NEXT:    vmovdqa (%rdi), %xmm0
; AVX512F-NEXT:    vmovdqa (%rdx), %xmm1
; AVX512F-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
; AVX512F-NEXT:    vpunpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1]
; AVX512F-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512F-NEXT:    vinserti128 $1, (%r8), %ymm0, %ymm0
; AVX512F-NEXT:    vpshufb {{.*#+}} ymm1 = ymm0[0,4,8,12],zero,ymm0[1,5,9,13],zero,ymm0[2,6,10,14],zero,ymm0[3],zero,zero,zero,ymm0[19,u,u,u,u,u,u,u,u,u,u,u,u]
; AVX512F-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,3,0,1]
; AVX512F-NEXT:    vpshufb {{.*#+}} ymm0 = zero,zero,zero,zero,ymm0[0],zero,zero,zero,zero,ymm0[1],zero,zero,zero,zero,ymm0[2],zero,ymm0[23,27,31],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; AVX512F-NEXT:    vpor %ymm0, %ymm1, %ymm0
; AVX512F-NEXT:    vextracti128 $1, %ymm0, %xmm1
; AVX512F-NEXT:    vmovd %xmm1, 16(%r9)
; AVX512F-NEXT:    vmovdqa %xmm0, (%r9)
; AVX512F-NEXT:    vzeroupper
; AVX512F-NEXT:    retq
;
; AVX512BW-LABEL: store_i8_stride5_vf4:
; AVX512BW:       # %bb.0:
; AVX512BW-NEXT:    vmovdqa (%rdi), %xmm0
; AVX512BW-NEXT:    vmovdqa (%rdx), %xmm1
; AVX512BW-NEXT:    vpunpckldq {{.*#+}} xmm0 = xmm0[0],mem[0],xmm0[1],mem[1]
; AVX512BW-NEXT:    vpunpckldq {{.*#+}} xmm1 = xmm1[0],mem[0],xmm1[1],mem[1]
; AVX512BW-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0]
; AVX512BW-NEXT:    vinserti128 $1, (%r8), %ymm0, %ymm0
; AVX512BW-NEXT:    vpshufb {{.*#+}} ymm1 = ymm0[0,4,8,12],zero,ymm0[1,5,9,13],zero,ymm0[2,6,10,14],zero,ymm0[3],zero,zero,zero,ymm0[19],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; AVX512BW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,3,0,1]
; AVX512BW-NEXT:    vpshufb {{.*#+}} ymm0 = zero,zero,zero,zero,ymm0[0],zero,zero,zero,zero,ymm0[1],zero,zero,zero,zero,ymm0[2],zero,ymm0[23,27,31],zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero,zero
; AVX512BW-NEXT:    vpor %ymm1, %ymm0, %ymm0
; AVX512BW-NEXT:    vextracti128 $1, %ymm0, %xmm1
; AVX512BW-NEXT:    vmovd %xmm1, 16(%r9)
; AVX512BW-NEXT:    vmovdqa %xmm0, (%r9)
; AVX512BW-NEXT:    vzeroupper
; AVX512BW-NEXT:    retq
  %in.vec0 = load <4 x i8>, ptr %in.vecptr0, align 64
  %in.vec1 = load <4 x i8>, ptr %in.vecptr1, align 64
  %in.vec2 = load <4 x i8>, ptr %in.vecptr2, align 64
  %in.vec3 = load <4 x i8>, ptr %in.vecptr3, align 64
  %in.vec4 = load <4 x i8>, ptr %in.vecptr4, align 64
  %1 = shufflevector <4 x i8> %in.vec0, <4 x i8> %in.vec1, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %2 = shufflevector <4 x i8> %in.vec2, <4 x i8> %in.vec3, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
  %3 = shufflevector <8 x i8> %1, <8 x i8> %2, <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>
  %4 = shufflevector <4 x i8> %in.vec4, <4 x i8> poison, <16 x i32> <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5 = shufflevector <16 x i8> %3, <16 x i8> %4, <20 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>
  %interleaved.vec = shufflevector <20 x i8> %5, <20 x i8> poison, <20 x i32> <i32 0, i32 4, i32 8, i32 12, i32 16, i32 1, i32 5, i32 9, i32 13, i32 17, i32 2, i32 6, i32 10, i32 14, i32 18, i32 3, i32 7, i32 11, i32 15, i32 19>
  store <20 x i8> %interleaved.vec, ptr %out.vec, align 64
  ret void
}

define void @store_i8_stride5_vf8(ptr %in.vecptr0, ptr %in.vecptr1, ptr %in.vecptr2, ptr %in.vecptr3, ptr %in.vecptr4, ptr %out.vec) nounwind {
; SSE-LABEL: store_i8_stride5_vf8:
; SSE:       # %bb.0:
; SSE-NEXT:    movq {{.*#+}} xmm1 = mem[0],zero
; SSE-NEXT:    movq {{.*#+}} xmm2 = mem[0],zero
; SSE-NEXT:    movq {{.*#+}} xmm4 = mem[0],zero
; SSE-NEXT:    movq {{.*#+}} xmm3 = mem[0],zero
; SSE-NEXT:    movq {{.*#+}} xmm0 = mem[0],zero
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[1,1,1,1]
; SSE-NEXT:    movdqa {{.*#+}} xmm8 = [65535,65535,0,65535,0,65535,65535,0]
; SSE-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm3[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm6 = xmm6[0,3,2,0,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm6 = xmm6[0,1,2,3,5,5,5,5]
; SSE-NEXT:    pand %xmm8, %xmm6
; SSE-NEXT:    pandn %xmm5, %xmm8
; SSE-NEXT:    por %xmm6, %xmm8
; SSE-NEXT:    movdqa {{.*#+}} xmm9 = [255,255,0,255,0,255,255,0,255,0,255,255,0,255,0,255]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm4[1,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,0,0,0]
; SSE-NEXT:    movdqa {{.*#+}} xmm6 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; SSE-NEXT:    movdqa %xmm6, %xmm10
; SSE-NEXT:    pandn %xmm5, %xmm10
; SSE-NEXT:    movdqa %xmm2, %xmm7
; SSE-NEXT:    punpcklbw {{.*#+}} xmm7 = xmm7[0],xmm2[0],xmm7[1],xmm2[1],xmm7[2],xmm2[2],xmm7[3],xmm2[3],xmm7[4],xmm2[4],xmm7[5],xmm2[5],xmm7[6],xmm2[6],xmm7[7],xmm2[7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm7[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,4,5,6,6]
; SSE-NEXT:    pand %xmm6, %xmm5
; SSE-NEXT:    por %xmm10, %xmm5
; SSE-NEXT:    pand %xmm9, %xmm5
; SSE-NEXT:    pandn %xmm8, %xmm9
; SSE-NEXT:    por %xmm5, %xmm9
; SSE-NEXT:    movdqa {{.*#+}} xmm8 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; SSE-NEXT:    pand %xmm8, %xmm9
; SSE-NEXT:    pshufd {{.*#+}} xmm10 = xmm0[0,0,1,1]
; SSE-NEXT:    movdqa %xmm8, %xmm5
; SSE-NEXT:    pandn %xmm10, %xmm5
; SSE-NEXT:    por %xmm9, %xmm5
; SSE-NEXT:    pshuflw {{.*#+}} xmm9 = xmm4[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm9 = xmm9[0,0,0,0]
; SSE-NEXT:    pand %xmm8, %xmm9
; SSE-NEXT:    pshuflw {{.*#+}} xmm10 = xmm3[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm10 = xmm10[0,1,0,1]
; SSE-NEXT:    pandn %xmm10, %xmm8
; SSE-NEXT:    por %xmm9, %xmm8
; SSE-NEXT:    movdqa {{.*#+}} xmm9 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; SSE-NEXT:    pshufd {{.*#+}} xmm10 = xmm1[0,0,0,0]
; SSE-NEXT:    pand %xmm6, %xmm10
; SSE-NEXT:    pshuflw {{.*#+}} xmm7 = xmm7[0,1,2,2,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm7[0,0,1,3]
; SSE-NEXT:    pandn %xmm7, %xmm6
; SSE-NEXT:    por %xmm10, %xmm6
; SSE-NEXT:    pand %xmm9, %xmm6
; SSE-NEXT:    pandn %xmm8, %xmm9
; SSE-NEXT:    por %xmm6, %xmm9
; SSE-NEXT:    movdqa {{.*#+}} xmm6 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; SSE-NEXT:    pand %xmm6, %xmm9
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm0[0,0,0,0]
; SSE-NEXT:    pandn %xmm7, %xmm6
; SSE-NEXT:    por %xmm9, %xmm6
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[3,3,3,3,4,5,6,7]
; SSE-NEXT:    movdqa {{.*#+}} xmm7 = [255,0,255,255,255,255,0,255,255,255,255,255,255,255,255,255]
; SSE-NEXT:    pand %xmm7, %xmm4
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[3,3,3,3]
; SSE-NEXT:    pandn %xmm3, %xmm7
; SSE-NEXT:    por %xmm4, %xmm7
; SSE-NEXT:    movdqa {{.*#+}} xmm3 = [255,255,255,0,0,255,255,255,255,255,255,255,255,255,255,255]
; SSE-NEXT:    pand %xmm3, %xmm7
; SSE-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm2[0,1,2,3,7,7,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[2,2,2,2]
; SSE-NEXT:    pandn %xmm1, %xmm3
; SSE-NEXT:    por %xmm7, %xmm3
; SSE-NEXT:    movdqa {{.*#+}} xmm1 = [255,255,0,255,255,255,255,0,255,255,255,255,255,255,255,255]
; SSE-NEXT:    pand %xmm1, %xmm3
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,1,1]
; SSE-NEXT:    pandn %xmm0, %xmm1
; SSE-NEXT:    por %xmm3, %xmm1
; SSE-NEXT:    movq %xmm1, 32(%r9)
; SSE-NEXT:    movdqa %xmm6, (%r9)
; SSE-NEXT:    movdqa %xmm5, 16(%r9)
; SSE-NEXT:    retq
;
; AVX1-ONLY-LABEL: store_i8_stride5_vf8:
; AVX1-ONLY:       # %bb.0:
; AVX1-ONLY-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
; AVX1-ONLY-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX1-ONLY-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
; AVX1-ONLY-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX1-ONLY-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX1-ONLY-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
; AVX1-ONLY-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = zero,zero,xmm0[u,7,15],zero,zero,xmm0[u,u,u,u,u,u,u,u,u]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm1[6,14,u],zero,zero,xmm1[7,15,u,u,u,u,u,u,u,u,u]
; AVX1-ONLY-NEXT:    vpor %xmm3, %xmm4, %xmm3
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = xmm3[0,2,13,6,8,10,12,15,u,u,u,u,u,u,u,u]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = zero,zero,xmm1[0,8,u],zero,zero,xmm1[1,9,u],zero,zero,xmm1[2,10,u],zero
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = xmm0[0,8],zero,zero,xmm0[u,1,9],zero,zero,xmm0[u,2,10],zero,zero,xmm0[u,3]
; AVX1-ONLY-NEXT:    vpor %xmm4, %xmm5, %xmm4
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm4[0,1,2,3],zero,xmm4[5,6,7,8],zero,xmm4[10,11,12,13],zero,xmm4[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = zero,zero,zero,zero,xmm2[0],zero,zero,zero,zero,xmm2[1],zero,zero,zero,zero,xmm2[2],zero
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm4, %xmm4
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm1 = zero,xmm1[3,11,u],zero,zero,xmm1[4,12,u],zero,zero,xmm1[5,13,u],zero,zero
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[11],zero,zero,xmm0[u,4,12],zero,zero,xmm0[u,5,13],zero,zero,xmm0[u,6,14]
; AVX1-ONLY-NEXT:    vpor %xmm1, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[0,1,2],zero,xmm0[4,5,6,7],zero,xmm0[9,10,11,12],zero,xmm0[14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm1 = zero,zero,zero,xmm2[3],zero,zero,zero,zero,xmm2[4],zero,zero,zero,zero,xmm2[5],zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm1, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vmovdqa %xmm0, 16(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm4, (%r9)
; AVX1-ONLY-NEXT:    vmovq %xmm3, 32(%r9)
; AVX1-ONLY-NEXT:    retq
;
; AVX2-SLOW-LABEL: store_i8_stride5_vf8:
; AVX2-SLOW:       # %bb.0:
; AVX2-SLOW-NEXT:    movq (%r8), %rax
; AVX2-SLOW-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
; AVX2-SLOW-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX2-SLOW-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
; AVX2-SLOW-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX2-SLOW-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX2-SLOW-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
; AVX2-SLOW-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm2
; AVX2-SLOW-NEXT:    vmovq %rax, %xmm3
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm4 = ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,zero,ymm2[3],zero,ymm2[19,27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,3,0,1]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,ymm2[27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero,ymm2[22,30]
; AVX2-SLOW-NEXT:    vpor %ymm2, %ymm4, %ymm2
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm3 = xmm3[0,0,1,1]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,0,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm4 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm4, %ymm2, %ymm3, %ymm2
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm0 = zero,zero,xmm0[u,7,15],zero,zero,xmm0[u,u,u,u,u,u,u,u,u]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[6,14,u],zero,zero,xmm1[7,15,u,u,u,u,u,u,u,u,u]
; AVX2-SLOW-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX2-SLOW-NEXT:    shrq $48, %rax
; AVX2-SLOW-NEXT:    vmovd %eax, %xmm1
; AVX2-SLOW-NEXT:    vpbroadcastw %xmm1, %xmm1
; AVX2-SLOW-NEXT:    vpbroadcastq {{.*#+}} xmm3 = [255,255,0,255,255,255,255,0,255,255,0,255,255,255,255,0]
; AVX2-SLOW-NEXT:    vpblendvb %xmm3, %xmm0, %xmm1, %xmm0
; AVX2-SLOW-NEXT:    vmovq %xmm0, 32(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm2, (%r9)
; AVX2-SLOW-NEXT:    vzeroupper
; AVX2-SLOW-NEXT:    retq
;
; AVX2-FAST-LABEL: store_i8_stride5_vf8:
; AVX2-FAST:       # %bb.0:
; AVX2-FAST-NEXT:    movq (%r8), %rax
; AVX2-FAST-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
; AVX2-FAST-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX2-FAST-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
; AVX2-FAST-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX2-FAST-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX2-FAST-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
; AVX2-FAST-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm2
; AVX2-FAST-NEXT:    vmovq %rax, %xmm3
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm4 = ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,zero,ymm2[3],zero,ymm2[19,27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,3,0,1]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,ymm2[27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero,ymm2[22,30]
; AVX2-FAST-NEXT:    vpor %ymm2, %ymm4, %ymm2
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = [0,0,0,0,0,0,1,1]
; AVX2-FAST-NEXT:    vpermd %ymm3, %ymm4, %ymm3
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm4, %ymm2, %ymm3, %ymm2
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm0 = zero,zero,xmm0[u,7,15],zero,zero,xmm0[u,u,u,u,u,u,u,u,u]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[6,14,u],zero,zero,xmm1[7,15,u,u,u,u,u,u,u,u,u]
; AVX2-FAST-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX2-FAST-NEXT:    shrq $48, %rax
; AVX2-FAST-NEXT:    vmovd %eax, %xmm1
; AVX2-FAST-NEXT:    vpbroadcastw %xmm1, %xmm1
; AVX2-FAST-NEXT:    vpbroadcastq {{.*#+}} xmm3 = [255,255,0,255,255,255,255,0,255,255,0,255,255,255,255,0]
; AVX2-FAST-NEXT:    vpblendvb %xmm3, %xmm0, %xmm1, %xmm0
; AVX2-FAST-NEXT:    vmovq %xmm0, 32(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm2, (%r9)
; AVX2-FAST-NEXT:    vzeroupper
; AVX2-FAST-NEXT:    retq
;
; AVX2-FAST-PERLANE-LABEL: store_i8_stride5_vf8:
; AVX2-FAST-PERLANE:       # %bb.0:
; AVX2-FAST-PERLANE-NEXT:    movq (%r8), %rax
; AVX2-FAST-PERLANE-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
; AVX2-FAST-PERLANE-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX2-FAST-PERLANE-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
; AVX2-FAST-PERLANE-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX2-FAST-PERLANE-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX2-FAST-PERLANE-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
; AVX2-FAST-PERLANE-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vmovq %rax, %xmm3
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm4 = ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,zero,ymm2[3],zero,ymm2[19,27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,3,0,1]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,ymm2[27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero,ymm2[22,30]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm2, %ymm4, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} xmm3 = xmm3[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,0,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm4 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm4, %ymm2, %ymm3, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm0 = zero,zero,xmm0[u,7,15],zero,zero,xmm0[u,u,u,u,u,u,u,u,u]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[6,14,u],zero,zero,xmm1[7,15,u,u,u,u,u,u,u,u,u]
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX2-FAST-PERLANE-NEXT:    shrq $48, %rax
; AVX2-FAST-PERLANE-NEXT:    vmovd %eax, %xmm1
; AVX2-FAST-PERLANE-NEXT:    vpbroadcastw %xmm1, %xmm1
; AVX2-FAST-PERLANE-NEXT:    vpbroadcastq {{.*#+}} xmm3 = [255,255,0,255,255,255,255,0,255,255,0,255,255,255,255,0]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %xmm3, %xmm0, %xmm1, %xmm0
; AVX2-FAST-PERLANE-NEXT:    vmovq %xmm0, 32(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm2, (%r9)
; AVX2-FAST-PERLANE-NEXT:    vzeroupper
; AVX2-FAST-PERLANE-NEXT:    retq
;
; AVX512F-SLOW-LABEL: store_i8_stride5_vf8:
; AVX512F-SLOW:       # %bb.0:
; AVX512F-SLOW-NEXT:    movq (%r8), %rax
; AVX512F-SLOW-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
; AVX512F-SLOW-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX512F-SLOW-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
; AVX512F-SLOW-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX512F-SLOW-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX512F-SLOW-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
; AVX512F-SLOW-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm2
; AVX512F-SLOW-NEXT:    vmovq %rax, %xmm3
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm4 = ymm2[0,8],zero,zero,ymm2[u,1,9],zero,zero,ymm2[u,2,10],zero,zero,ymm2[u,3],zero,ymm2[19,27,u],zero,zero,ymm2[20,28,u],zero,zero,ymm2[21,29,u],zero,zero
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,3,0,1]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,ymm2[0,8,u],zero,zero,ymm2[1,9,u],zero,zero,ymm2[2,10,u],zero,ymm2[27],zero,zero,ymm2[u,20,28],zero,zero,ymm2[u,21,29],zero,zero,ymm2[u,22,30]
; AVX512F-SLOW-NEXT:    vpor %ymm2, %ymm4, %ymm2
; AVX512F-SLOW-NEXT:    vpshufd {{.*#+}} xmm3 = xmm3[0,0,1,1]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,0,1]
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm3
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm0 = zero,zero,xmm0[u,7,15],zero,zero,xmm0[u,u,u,u,u,u,u,u,u]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[6,14,u],zero,zero,xmm1[7,15,u,u,u,u,u,u,u,u,u]
; AVX512F-SLOW-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX512F-SLOW-NEXT:    shrq $48, %rax
; AVX512F-SLOW-NEXT:    vmovd %eax, %xmm1
; AVX512F-SLOW-NEXT:    vpbroadcastw %xmm1, %xmm1
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm1
; AVX512F-SLOW-NEXT:    vinserti32x4 $2, %xmm1, %zmm3, %zmm0
; AVX512F-SLOW-NEXT:    vmovq %xmm1, 32(%r9)
; AVX512F-SLOW-NEXT:    vmovdqa %ymm0, (%r9)
; AVX512F-SLOW-NEXT:    vzeroupper
; AVX512F-SLOW-NEXT:    retq
;
; AVX512F-FAST-LABEL: store_i8_stride5_vf8:
; AVX512F-FAST:       # %bb.0:
; AVX512F-FAST-NEXT:    movq (%r8), %rax
; AVX512F-FAST-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
; AVX512F-FAST-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX512F-FAST-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
; AVX512F-FAST-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX512F-FAST-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX512F-FAST-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
; AVX512F-FAST-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm2
; AVX512F-FAST-NEXT:    vmovq %rax, %xmm3
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm4 = ymm2[0,8],zero,zero,ymm2[u,1,9],zero,zero,ymm2[u,2,10],zero,zero,ymm2[u,3],zero,ymm2[19,27,u],zero,zero,ymm2[20,28,u],zero,zero,ymm2[21,29,u],zero,zero
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,3,0,1]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,ymm2[0,8,u],zero,zero,ymm2[1,9,u],zero,zero,ymm2[2,10,u],zero,ymm2[27],zero,zero,ymm2[u,20,28],zero,zero,ymm2[u,21,29],zero,zero,ymm2[u,22,30]
; AVX512F-FAST-NEXT:    vpor %ymm2, %ymm4, %ymm2
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = <u,0,0,0,0,u,1,1>
; AVX512F-FAST-NEXT:    vpermd %ymm3, %ymm4, %ymm3
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm3
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm0 = zero,zero,xmm0[u,7,15],zero,zero,xmm0[u,u,u,u,u,u,u,u,u]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[6,14,u],zero,zero,xmm1[7,15,u,u,u,u,u,u,u,u,u]
; AVX512F-FAST-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX512F-FAST-NEXT:    shrq $48, %rax
; AVX512F-FAST-NEXT:    vmovd %eax, %xmm1
; AVX512F-FAST-NEXT:    vpbroadcastw %xmm1, %xmm1
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm1
; AVX512F-FAST-NEXT:    vinserti32x4 $2, %xmm1, %zmm3, %zmm0
; AVX512F-FAST-NEXT:    vmovq %xmm1, 32(%r9)
; AVX512F-FAST-NEXT:    vmovdqa %ymm0, (%r9)
; AVX512F-FAST-NEXT:    vzeroupper
; AVX512F-FAST-NEXT:    retq
;
; AVX512BW-SLOW-LABEL: store_i8_stride5_vf8:
; AVX512BW-SLOW:       # %bb.0:
; AVX512BW-SLOW-NEXT:    movq (%r8), %rax
; AVX512BW-SLOW-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
; AVX512BW-SLOW-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX512BW-SLOW-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
; AVX512BW-SLOW-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX512BW-SLOW-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX512BW-SLOW-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
; AVX512BW-SLOW-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm2
; AVX512BW-SLOW-NEXT:    vmovq %rax, %xmm3
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm4 = ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,zero,ymm2[3],zero,ymm2[19,27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,3,0,1]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,ymm2[27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero,ymm2[22,30]
; AVX512BW-SLOW-NEXT:    vpor %ymm4, %ymm2, %ymm2
; AVX512BW-SLOW-NEXT:    vpshufd {{.*#+}} xmm3 = xmm3[0,0,1,1]
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,0,1]
; AVX512BW-SLOW-NEXT:    movl $554189328, %ecx # imm = 0x21084210
; AVX512BW-SLOW-NEXT:    kmovd %ecx, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %ymm3, %ymm2 {%k1}
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm0 = zero,zero,xmm0[u,7,15],zero,zero,xmm0[u,u,u,u,u,u,u,u,u]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[6,14,u],zero,zero,xmm1[7,15,u,u,u,u,u,u,u,u,u]
; AVX512BW-SLOW-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX512BW-SLOW-NEXT:    shrq $48, %rax
; AVX512BW-SLOW-NEXT:    vpbroadcastw %eax, %xmm1
; AVX512BW-SLOW-NEXT:    movw $132, %ax
; AVX512BW-SLOW-NEXT:    kmovd %eax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %xmm1, %xmm0 {%k1}
; AVX512BW-SLOW-NEXT:    vinserti32x4 $2, %xmm0, %zmm2, %zmm1
; AVX512BW-SLOW-NEXT:    vmovq %xmm0, 32(%r9)
; AVX512BW-SLOW-NEXT:    vmovdqa %ymm1, (%r9)
; AVX512BW-SLOW-NEXT:    vzeroupper
; AVX512BW-SLOW-NEXT:    retq
;
; AVX512BW-FAST-LABEL: store_i8_stride5_vf8:
; AVX512BW-FAST:       # %bb.0:
; AVX512BW-FAST-NEXT:    movq (%r8), %rax
; AVX512BW-FAST-NEXT:    vmovq {{.*#+}} xmm0 = mem[0],zero
; AVX512BW-FAST-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX512BW-FAST-NEXT:    vpunpcklqdq {{.*#+}} xmm0 = xmm1[0],xmm0[0]
; AVX512BW-FAST-NEXT:    vmovq {{.*#+}} xmm1 = mem[0],zero
; AVX512BW-FAST-NEXT:    vmovq {{.*#+}} xmm2 = mem[0],zero
; AVX512BW-FAST-NEXT:    vpunpcklqdq {{.*#+}} xmm1 = xmm2[0],xmm1[0]
; AVX512BW-FAST-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm2
; AVX512BW-FAST-NEXT:    vmovq %rax, %xmm3
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm4 = ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,zero,ymm2[3],zero,ymm2[19,27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,3,0,1]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,ymm2[0,8],zero,zero,zero,ymm2[1,9],zero,zero,zero,ymm2[2,10],zero,zero,ymm2[27],zero,zero,zero,ymm2[20,28],zero,zero,zero,ymm2[21,29],zero,zero,zero,ymm2[22,30]
; AVX512BW-FAST-NEXT:    vpor %ymm4, %ymm2, %ymm2
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = [0,0,0,0,0,0,1,1]
; AVX512BW-FAST-NEXT:    vpermd %ymm3, %ymm4, %ymm3
; AVX512BW-FAST-NEXT:    movl $554189328, %ecx # imm = 0x21084210
; AVX512BW-FAST-NEXT:    kmovd %ecx, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %ymm3, %ymm2 {%k1}
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm0 = zero,zero,xmm0[u,7,15],zero,zero,xmm0[u,u,u,u,u,u,u,u,u]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[6,14,u],zero,zero,xmm1[7,15,u,u,u,u,u,u,u,u,u]
; AVX512BW-FAST-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX512BW-FAST-NEXT:    shrq $48, %rax
; AVX512BW-FAST-NEXT:    vpbroadcastw %eax, %xmm1
; AVX512BW-FAST-NEXT:    movw $132, %ax
; AVX512BW-FAST-NEXT:    kmovd %eax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %xmm1, %xmm0 {%k1}
; AVX512BW-FAST-NEXT:    vinserti32x4 $2, %xmm0, %zmm2, %zmm1
; AVX512BW-FAST-NEXT:    vmovq %xmm0, 32(%r9)
; AVX512BW-FAST-NEXT:    vmovdqa %ymm1, (%r9)
; AVX512BW-FAST-NEXT:    vzeroupper
; AVX512BW-FAST-NEXT:    retq
  %in.vec0 = load <8 x i8>, ptr %in.vecptr0, align 64
  %in.vec1 = load <8 x i8>, ptr %in.vecptr1, align 64
  %in.vec2 = load <8 x i8>, ptr %in.vecptr2, align 64
  %in.vec3 = load <8 x i8>, ptr %in.vecptr3, align 64
  %in.vec4 = load <8 x i8>, ptr %in.vecptr4, align 64
  %1 = shufflevector <8 x i8> %in.vec0, <8 x i8> %in.vec1, <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>
  %2 = shufflevector <8 x i8> %in.vec2, <8 x i8> %in.vec3, <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>
  %3 = shufflevector <16 x i8> %1, <16 x i8> %2, <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 31>
  %4 = shufflevector <8 x i8> %in.vec4, <8 x i8> poison, <32 x i32> <i32 0, i32 1, i32 2, i32 3, 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 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5 = shufflevector <32 x i8> %3, <32 x i8> %4, <40 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39>
  %interleaved.vec = shufflevector <40 x i8> %5, <40 x i8> poison, <40 x i32> <i32 0, i32 8, i32 16, i32 24, i32 32, i32 1, i32 9, i32 17, i32 25, i32 33, i32 2, i32 10, i32 18, i32 26, i32 34, i32 3, i32 11, i32 19, i32 27, i32 35, i32 4, i32 12, i32 20, i32 28, i32 36, i32 5, i32 13, i32 21, i32 29, i32 37, i32 6, i32 14, i32 22, i32 30, i32 38, i32 7, i32 15, i32 23, i32 31, i32 39>
  store <40 x i8> %interleaved.vec, ptr %out.vec, align 64
  ret void
}

define void @store_i8_stride5_vf16(ptr %in.vecptr0, ptr %in.vecptr1, ptr %in.vecptr2, ptr %in.vecptr3, ptr %in.vecptr4, ptr %out.vec) nounwind {
; SSE-LABEL: store_i8_stride5_vf16:
; SSE:       # %bb.0:
; SSE-NEXT:    movdqa (%rdi), %xmm10
; SSE-NEXT:    movdqa (%rsi), %xmm8
; SSE-NEXT:    movdqa (%rdx), %xmm2
; SSE-NEXT:    movdqa (%rcx), %xmm4
; SSE-NEXT:    movdqa (%r8), %xmm0
; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm2[3,3,3,3,4,5,6,7]
; SSE-NEXT:    movdqa %xmm2, %xmm9
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa {{.*#+}} xmm6 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; SSE-NEXT:    pand %xmm6, %xmm1
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm4[2,1,2,3]
; SSE-NEXT:    punpcklbw {{.*#+}} xmm3 = xmm3[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[3,3,0,3]
; SSE-NEXT:    pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa %xmm6, %xmm5
; SSE-NEXT:    pandn %xmm3, %xmm5
; SSE-NEXT:    por %xmm1, %xmm5
; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255]
; SSE-NEXT:    pand %xmm2, %xmm5
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm10[1,1,2,2]
; SSE-NEXT:    movdqa %xmm10, %xmm12
; SSE-NEXT:    movdqa %xmm10, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa {{.*#+}} xmm1 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; SSE-NEXT:    pand %xmm1, %xmm7
; SSE-NEXT:    pshufd {{.*#+}} xmm10 = xmm8[0,1,2,1]
; SSE-NEXT:    pshufhw {{.*#+}} xmm10 = xmm10[0,1,2,3,4,7,6,7]
; SSE-NEXT:    punpckhbw {{.*#+}} xmm10 = xmm10[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    pshuflw {{.*#+}} xmm10 = xmm10[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm10 = xmm10[0,0,1,0]
; SSE-NEXT:    movdqa %xmm1, %xmm11
; SSE-NEXT:    pandn %xmm10, %xmm11
; SSE-NEXT:    por %xmm7, %xmm11
; SSE-NEXT:    movdqa %xmm2, %xmm10
; SSE-NEXT:    pandn %xmm11, %xmm10
; SSE-NEXT:    por %xmm5, %xmm10
; SSE-NEXT:    movdqa {{.*#+}} xmm7 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; SSE-NEXT:    pand %xmm7, %xmm10
; SSE-NEXT:    movdqa %xmm0, %xmm5
; SSE-NEXT:    pshufd {{.*#+}} xmm11 = xmm0[1,1,2,2]
; SSE-NEXT:    movdqa %xmm7, %xmm0
; SSE-NEXT:    pandn %xmm11, %xmm0
; SSE-NEXT:    por %xmm10, %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm11 = xmm12[2,2,3,3]
; SSE-NEXT:    movdqa {{.*#+}} xmm12 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; SSE-NEXT:    pand %xmm12, %xmm11
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm8[8],xmm0[9],xmm8[9],xmm0[10],xmm8[10],xmm0[11],xmm8[11],xmm0[12],xmm8[12],xmm0[13],xmm8[13],xmm0[14],xmm8[14],xmm0[15],xmm8[15]
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm13 = xmm0[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm13 = xmm13[2,2,2,2,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm13 = xmm13[0,1,2,3,7,5,4,7]
; SSE-NEXT:    movdqa %xmm12, %xmm14
; SSE-NEXT:    pandn %xmm13, %xmm14
; SSE-NEXT:    por %xmm11, %xmm14
; SSE-NEXT:    movdqa {{.*#+}} xmm13 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; SSE-NEXT:    movdqa %xmm13, %xmm11
; SSE-NEXT:    pandn %xmm14, %xmm11
; SSE-NEXT:    pshufhw {{.*#+}} xmm14 = xmm9[0,1,2,3,5,6,6,7]
; SSE-NEXT:    movdqa %xmm9, %xmm10
; SSE-NEXT:    pshufd {{.*#+}} xmm14 = xmm14[2,2,2,2]
; SSE-NEXT:    movdqa %xmm1, %xmm15
; SSE-NEXT:    pandn %xmm14, %xmm15
; SSE-NEXT:    movdqa %xmm4, %xmm14
; SSE-NEXT:    punpckhbw {{.*#+}} xmm14 = xmm14[8],xmm4[8],xmm14[9],xmm4[9],xmm14[10],xmm4[10],xmm14[11],xmm4[11],xmm14[12],xmm4[12],xmm14[13],xmm4[13],xmm14[14],xmm4[14],xmm14[15],xmm4[15]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm14[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[1,1,2,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,7,6,4]
; SSE-NEXT:    pand %xmm1, %xmm0
; SSE-NEXT:    por %xmm15, %xmm0
; SSE-NEXT:    pand %xmm13, %xmm0
; SSE-NEXT:    por %xmm11, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm15 = xmm5[2,2,2,2]
; SSE-NEXT:    movdqa %xmm6, %xmm11
; SSE-NEXT:    pandn %xmm15, %xmm11
; SSE-NEXT:    pand %xmm6, %xmm0
; SSE-NEXT:    por %xmm0, %xmm11
; SSE-NEXT:    punpcklbw {{.*#+}} xmm4 = xmm4[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm4[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,3,2,0,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,5,5,5]
; SSE-NEXT:    movdqa %xmm7, %xmm15
; SSE-NEXT:    pandn %xmm0, %xmm15
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm9[1,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE-NEXT:    pand %xmm7, %xmm0
; SSE-NEXT:    por %xmm0, %xmm15
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm9 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm9[1,1,1,1]
; SSE-NEXT:    movdqa %xmm1, %xmm3
; SSE-NEXT:    pandn %xmm0, %xmm3
; SSE-NEXT:    punpcklbw {{.*#+}} xmm8 = xmm8[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm8[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,5,6,6]
; SSE-NEXT:    pand %xmm1, %xmm0
; SSE-NEXT:    por %xmm3, %xmm0
; SSE-NEXT:    movdqa {{.*#+}} xmm3 = [255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; SSE-NEXT:    pand %xmm3, %xmm0
; SSE-NEXT:    pandn %xmm15, %xmm3
; SSE-NEXT:    por %xmm0, %xmm3
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm5[0,0,1,1]
; SSE-NEXT:    movdqa %xmm12, %xmm15
; SSE-NEXT:    pandn %xmm0, %xmm15
; SSE-NEXT:    pand %xmm12, %xmm3
; SSE-NEXT:    por %xmm3, %xmm15
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm10[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE-NEXT:    pand %xmm12, %xmm0
; SSE-NEXT:    pshuflw {{.*#+}} xmm3 = xmm4[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,1,0,1]
; SSE-NEXT:    pandn %xmm3, %xmm12
; SSE-NEXT:    por %xmm0, %xmm12
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm9[0,0,0,0]
; SSE-NEXT:    pand %xmm6, %xmm0
; SSE-NEXT:    pshuflw {{.*#+}} xmm3 = xmm8[0,1,2,2,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,0,1,3]
; SSE-NEXT:    pandn %xmm3, %xmm6
; SSE-NEXT:    por %xmm0, %xmm6
; SSE-NEXT:    pand %xmm13, %xmm6
; SSE-NEXT:    pandn %xmm12, %xmm13
; SSE-NEXT:    por %xmm6, %xmm13
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm5[0,0,0,0]
; SSE-NEXT:    movdqa %xmm1, %xmm6
; SSE-NEXT:    pandn %xmm0, %xmm6
; SSE-NEXT:    pand %xmm1, %xmm13
; SSE-NEXT:    por %xmm13, %xmm6
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm10[0,1,2,3,7,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,2,2,2]
; SSE-NEXT:    pand %xmm1, %xmm0
; SSE-NEXT:    pshufhw {{.*#+}} xmm3 = xmm14[0,1,2,3,5,7,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,2,3,2]
; SSE-NEXT:    pandn %xmm3, %xmm1
; SSE-NEXT:    por %xmm0, %xmm1
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm9[3,3,3,3]
; SSE-NEXT:    pand %xmm7, %xmm0
; SSE-NEXT:    pshufhw $167, {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Folded Reload
; SSE-NEXT:    # xmm3 = mem[0,1,2,3,7,5,6,6]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[2,3,2,2]
; SSE-NEXT:    pandn %xmm3, %xmm7
; SSE-NEXT:    por %xmm0, %xmm7
; SSE-NEXT:    pand %xmm2, %xmm7
; SSE-NEXT:    pandn %xmm1, %xmm2
; SSE-NEXT:    por %xmm7, %xmm2
; SSE-NEXT:    movdqa {{.*#+}} xmm0 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; SSE-NEXT:    pand %xmm0, %xmm2
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm5[3,3,3,3]
; SSE-NEXT:    pandn %xmm1, %xmm0
; SSE-NEXT:    por %xmm2, %xmm0
; SSE-NEXT:    movdqa %xmm0, 64(%r9)
; SSE-NEXT:    movdqa %xmm6, (%r9)
; SSE-NEXT:    movdqa %xmm15, 16(%r9)
; SSE-NEXT:    movdqa %xmm11, 48(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 32(%r9)
; SSE-NEXT:    retq
;
; AVX1-ONLY-LABEL: store_i8_stride5_vf16:
; AVX1-ONLY:       # %bb.0:
; AVX1-ONLY-NEXT:    vmovdqa (%rdi), %xmm2
; AVX1-ONLY-NEXT:    vmovdqa (%rsi), %xmm3
; AVX1-ONLY-NEXT:    vmovdqa (%rdx), %xmm1
; AVX1-ONLY-NEXT:    vmovdqa (%rcx), %xmm4
; AVX1-ONLY-NEXT:    vmovdqa (%r8), %xmm0
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = zero,xmm4[6,u,u,u],zero,xmm4[7,u,u,u],zero,xmm4[8,u,u,u],zero
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm1[6],zero,xmm1[u,u,u,7],zero,xmm1[u,u,u,8],zero,xmm1[u,u,u,9]
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm6, %xmm5
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm3[u,u,u],zero,xmm3[7,u,u,u],zero,xmm3[8,u,u,u],zero,xmm3[9,u]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm7 = xmm2[u,u,u,7],zero,xmm2[u,u,u,8],zero,xmm2[u,u,u,9],zero,xmm2[u]
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm7, %xmm6
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm7 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255>
; AVX1-ONLY-NEXT:    vpblendvb %xmm7, %xmm5, %xmm6, %xmm5
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = xmm5[0,1],zero,xmm5[3,4,5,6],zero,xmm5[8,9,10,11],zero,xmm5[13,14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = zero,zero,xmm0[6],zero,zero,zero,zero,xmm0[7],zero,zero,zero,zero,xmm0[8],zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm5, %xmm5
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm6 = xmm1[8],xmm4[8],xmm1[9],xmm4[9],xmm1[10],xmm4[10],xmm1[11],xmm4[11],xmm1[12],xmm4[12],xmm1[13],xmm4[13],xmm1[14],xmm4[14],xmm1[15],xmm4[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = zero,zero,zero,xmm6[10,11],zero,zero,zero,xmm6[12,13],zero,zero,zero,xmm6[14,15],zero
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm7 = xmm2[8],xmm3[8],xmm2[9],xmm3[9],xmm2[10],xmm3[10],xmm2[11],xmm3[11],xmm2[12],xmm3[12],xmm2[13],xmm3[13],xmm2[14],xmm3[14],xmm2[15],xmm3[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm8 = zero,xmm7[10,11],zero,zero,zero,xmm7[12,13],zero,zero,zero,xmm7[14,15],zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm8, %xmm6
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm8 = xmm1[0],xmm4[0],xmm1[1],xmm4[1],xmm1[2],xmm4[2],xmm1[3],xmm4[3],xmm1[4],xmm4[4],xmm1[5],xmm4[5],xmm1[6],xmm4[6],xmm1[7],xmm4[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm9 = zero,zero,xmm8[0,1],zero,zero,zero,xmm8[2,3],zero,zero,zero,xmm8[4,5],zero,zero
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm10 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3],xmm2[4],xmm3[4],xmm2[5],xmm3[5],xmm2[6],xmm3[6],xmm2[7],xmm3[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm10 = xmm10[0,1],zero,zero,zero,xmm10[2,3],zero,zero,zero,xmm10[4,5],zero,zero,zero,xmm10[6]
; AVX1-ONLY-NEXT:    vpor %xmm9, %xmm10, %xmm9
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm10 = zero,zero,zero,zero,xmm0[0],zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,xmm0[2],zero
; AVX1-ONLY-NEXT:    vpor %xmm10, %xmm9, %xmm9
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm8 = zero,xmm8[6,7],zero,zero,zero,xmm8[8,9],zero,zero,zero,xmm8[10,11],zero,zero,zero
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm2 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm2[6],zero,zero,zero,xmm2[9,8],zero,zero,zero,xmm2[11,10],zero,zero,zero,xmm2[13,12]
; AVX1-ONLY-NEXT:    vpor %xmm2, %xmm8, %xmm2
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = zero,zero,zero,xmm0[3],zero,zero,zero,zero,xmm0[4],zero,zero,zero,zero,xmm0[5],zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm3, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = zero,zero,xmm7[4,5],zero,zero,zero,xmm7[6,7],zero,zero,zero,xmm7[8,9],zero,zero
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm1 = xmm4[8],xmm1[8],xmm4[9],xmm1[9],xmm4[10],xmm1[10],xmm4[11],xmm1[11],xmm4[12],xmm1[12],xmm4[13],xmm1[13],xmm4[14],xmm1[14],xmm4[15],xmm1[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[2],zero,zero,zero,xmm1[5,4],zero,zero,zero,xmm1[7,6],zero,zero,zero,xmm1[9,8]
; AVX1-ONLY-NEXT:    vpor %xmm3, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = zero,xmm0[9],zero,zero,zero,zero,xmm0[10],zero,zero,zero,zero,xmm0[11],zero,zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm3, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vmovdqa %xmm1, 48(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm2, 16(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm9, (%r9)
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[12],zero,zero,zero,zero,xmm0[13],zero,zero,zero,zero,xmm0[14],zero,zero,zero,zero,xmm0[15]
; AVX1-ONLY-NEXT:    vpor %xmm0, %xmm6, %xmm0
; AVX1-ONLY-NEXT:    vmovdqa %xmm0, 64(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm5, 32(%r9)
; AVX1-ONLY-NEXT:    retq
;
; AVX2-SLOW-LABEL: store_i8_stride5_vf16:
; AVX2-SLOW:       # %bb.0:
; AVX2-SLOW-NEXT:    vmovdqa (%rdi), %xmm0
; AVX2-SLOW-NEXT:    vmovdqa (%rsi), %xmm1
; AVX2-SLOW-NEXT:    vmovdqa (%rdx), %xmm3
; AVX2-SLOW-NEXT:    vmovdqa (%rcx), %xmm4
; AVX2-SLOW-NEXT:    vmovdqa (%r8), %xmm2
; AVX2-SLOW-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm5
; AVX2-SLOW-NEXT:    vinserti128 $1, %xmm4, %ymm3, %ymm6
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm7 = ymm6[6],zero,zero,zero,zero,ymm6[7],zero,zero,zero,zero,ymm6[8],zero,zero,zero,zero,ymm6[9,25],zero,zero,zero,zero,ymm6[26],zero,zero,zero,zero,ymm6[27],zero,zero,zero,zero,ymm6[28]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm6[2,3,0,1]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = zero,ymm8[6],zero,zero,zero,zero,ymm8[7],zero,zero,zero,zero,ymm8[8],zero,zero,zero,zero,zero,zero,zero,zero,ymm8[26],zero,zero,zero,zero,ymm8[27],zero,zero,zero,zero,ymm8[28],zero
; AVX2-SLOW-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = zero,zero,zero,ymm5[7],zero,zero,zero,zero,ymm5[8],zero,zero,zero,zero,ymm5[9],zero,zero,zero,zero,zero,ymm5[26],zero,zero,zero,zero,ymm5[27],zero,zero,zero,zero,ymm5[28],zero,zero
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm5[2,3,0,1]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm9 = zero,zero,zero,zero,ymm9[7],zero,zero,zero,zero,ymm9[8],zero,zero,zero,zero,ymm9[9],zero,zero,zero,ymm9[26],zero,zero,zero,zero,ymm9[27],zero,zero,zero,zero,ymm9[28],zero,zero,zero
; AVX2-SLOW-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm9 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm8 = xmm2[1,1,2,2]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[0,1,1,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,2,0,2]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm6 = zero,zero,ymm6[0,8],zero,zero,zero,ymm6[1,9],zero,zero,zero,ymm6[2,10],zero,zero,zero,ymm6[19,27],zero,zero,zero,ymm6[20,28],zero,zero,zero,ymm6[21,29],zero,zero,zero
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[0,2,2,0]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = ymm5[0,8],zero,zero,zero,ymm5[1,9],zero,zero,zero,ymm5[2,10],zero,zero,zero,ymm5[3,19],zero,zero,zero,ymm5[28,20],zero,zero,zero,ymm5[29,21],zero,zero,zero,ymm5[30,22]
; AVX2-SLOW-NEXT:    vpor %ymm6, %ymm5, %ymm5
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm6 = xmm2[0,0,1,1]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,0,0,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm8 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm8, %ymm5, %ymm6, %ymm5
; AVX2-SLOW-NEXT:    vpunpckhbw {{.*#+}} xmm3 = xmm3[8],xmm4[8],xmm3[9],xmm4[9],xmm3[10],xmm4[10],xmm3[11],xmm4[11],xmm3[12],xmm4[12],xmm3[13],xmm4[13],xmm3[14],xmm4[14],xmm3[15],xmm4[15]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm3 = zero,zero,zero,xmm3[10,11],zero,zero,zero,xmm3[12,13],zero,zero,zero,xmm3[14,15],zero
; AVX2-SLOW-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm0 = zero,xmm0[10,11],zero,zero,zero,xmm0[12,13],zero,zero,zero,xmm0[14,15],zero,zero,zero
; AVX2-SLOW-NEXT:    vpor %xmm3, %xmm0, %xmm0
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm1 = xmm2[12],zero,zero,zero,zero,xmm2[13],zero,zero,zero,zero,xmm2[14],zero,zero,zero,zero,xmm2[15]
; AVX2-SLOW-NEXT:    vpor %xmm1, %xmm0, %xmm0
; AVX2-SLOW-NEXT:    vmovdqa %xmm0, 64(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm5, (%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm7, 32(%r9)
; AVX2-SLOW-NEXT:    vzeroupper
; AVX2-SLOW-NEXT:    retq
;
; AVX2-FAST-LABEL: store_i8_stride5_vf16:
; AVX2-FAST:       # %bb.0:
; AVX2-FAST-NEXT:    vmovdqa (%rdi), %xmm0
; AVX2-FAST-NEXT:    vmovdqa (%rsi), %xmm1
; AVX2-FAST-NEXT:    vmovdqa (%rdx), %xmm2
; AVX2-FAST-NEXT:    vmovdqa (%rcx), %xmm3
; AVX2-FAST-NEXT:    vmovdqa (%r8), %xmm4
; AVX2-FAST-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm5
; AVX2-FAST-NEXT:    vinserti128 $1, %xmm3, %ymm2, %ymm6
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm6[0,2,0,2]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm7 = zero,zero,ymm7[0,8],zero,zero,zero,ymm7[1,9],zero,zero,zero,ymm7[2,10],zero,zero,zero,ymm7[19,27],zero,zero,zero,ymm7[20,28],zero,zero,zero,ymm7[21,29],zero,zero,zero
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm5[0,2,2,0]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = ymm8[0,8],zero,zero,zero,ymm8[1,9],zero,zero,zero,ymm8[2,10],zero,zero,zero,ymm8[3,19],zero,zero,zero,ymm8[28,20],zero,zero,zero,ymm8[29,21],zero,zero,zero,ymm8[30,22]
; AVX2-FAST-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [0,0,0,0,0,0,1,1]
; AVX2-FAST-NEXT:    vpermd %ymm4, %ymm8, %ymm8
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [1,5,2,6,2,6,3,7]
; AVX2-FAST-NEXT:    vpermd %ymm5, %ymm8, %ymm5
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm5 = zero,zero,zero,ymm5[3,7],zero,zero,zero,ymm5[8,12],zero,zero,zero,ymm5[9,13],zero,zero,zero,ymm5[18,22],zero,zero,zero,ymm5[19,23],zero,zero,zero,ymm5[24,28],zero,zero
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [1,5,2,6,6,2,3,7]
; AVX2-FAST-NEXT:    vpermd %ymm6, %ymm8, %ymm6
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm6 = ymm6[2,6],zero,zero,zero,ymm6[3,7],zero,zero,zero,ymm6[8,12],zero,zero,zero,ymm6[9,17],zero,zero,zero,ymm6[22,18],zero,zero,zero,ymm6[23,19],zero,zero,zero,ymm6[24,28]
; AVX2-FAST-NEXT:    vpor %ymm5, %ymm6, %ymm5
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm6 = [1,1,2,2,2,2,2,2]
; AVX2-FAST-NEXT:    vpermd %ymm4, %ymm6, %ymm6
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm8, %ymm5, %ymm6, %ymm5
; AVX2-FAST-NEXT:    vpunpckhbw {{.*#+}} xmm2 = xmm2[8],xmm3[8],xmm2[9],xmm3[9],xmm2[10],xmm3[10],xmm2[11],xmm3[11],xmm2[12],xmm3[12],xmm2[13],xmm3[13],xmm2[14],xmm3[14],xmm2[15],xmm3[15]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm2 = zero,zero,zero,xmm2[10,11],zero,zero,zero,xmm2[12,13],zero,zero,zero,xmm2[14,15],zero
; AVX2-FAST-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm0 = zero,xmm0[10,11],zero,zero,zero,xmm0[12,13],zero,zero,zero,xmm0[14,15],zero,zero,zero
; AVX2-FAST-NEXT:    vpor %xmm2, %xmm0, %xmm0
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm1 = xmm4[12],zero,zero,zero,zero,xmm4[13],zero,zero,zero,zero,xmm4[14],zero,zero,zero,zero,xmm4[15]
; AVX2-FAST-NEXT:    vpor %xmm1, %xmm0, %xmm0
; AVX2-FAST-NEXT:    vmovdqa %xmm0, 64(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm5, 32(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm7, (%r9)
; AVX2-FAST-NEXT:    vzeroupper
; AVX2-FAST-NEXT:    retq
;
; AVX2-FAST-PERLANE-LABEL: store_i8_stride5_vf16:
; AVX2-FAST-PERLANE:       # %bb.0:
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdi), %xmm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rsi), %xmm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdx), %xmm3
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rcx), %xmm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%r8), %xmm2
; AVX2-FAST-PERLANE-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm5
; AVX2-FAST-PERLANE-NEXT:    vinserti128 $1, %xmm4, %ymm3, %ymm6
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm7 = ymm6[6],zero,zero,zero,zero,ymm6[7],zero,zero,zero,zero,ymm6[8],zero,zero,zero,zero,ymm6[9,25],zero,zero,zero,zero,ymm6[26],zero,zero,zero,zero,ymm6[27],zero,zero,zero,zero,ymm6[28]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm6[2,3,0,1]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm8 = zero,ymm8[6],zero,zero,zero,zero,ymm8[7],zero,zero,zero,zero,ymm8[8],zero,zero,zero,zero,zero,zero,zero,zero,ymm8[26],zero,zero,zero,zero,ymm8[27],zero,zero,zero,zero,ymm8[28],zero
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm8 = zero,zero,zero,ymm5[7],zero,zero,zero,zero,ymm5[8],zero,zero,zero,zero,ymm5[9],zero,zero,zero,zero,zero,ymm5[26],zero,zero,zero,zero,ymm5[27],zero,zero,zero,zero,ymm5[28],zero,zero
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm5[2,3,0,1]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm9 = zero,zero,zero,zero,ymm9[7],zero,zero,zero,zero,ymm9[8],zero,zero,zero,zero,ymm9[9],zero,zero,zero,ymm9[26],zero,zero,zero,zero,ymm9[27],zero,zero,zero,zero,ymm9[28],zero,zero,zero
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm9 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} xmm8 = xmm2[1,1,2,2]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[0,1,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,2,0,2]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm6 = zero,zero,ymm6[0,8],zero,zero,zero,ymm6[1,9],zero,zero,zero,ymm6[2,10],zero,zero,zero,ymm6[19,27],zero,zero,zero,ymm6[20,28],zero,zero,zero,ymm6[21,29],zero,zero,zero
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[0,2,2,0]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm5 = ymm5[0,8],zero,zero,zero,ymm5[1,9],zero,zero,zero,ymm5[2,10],zero,zero,zero,ymm5[3,19],zero,zero,zero,ymm5[28,20],zero,zero,zero,ymm5[29,21],zero,zero,zero,ymm5[30,22]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm6, %ymm5, %ymm5
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} xmm6 = xmm2[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,0,0,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm8 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm8, %ymm5, %ymm6, %ymm5
; AVX2-FAST-PERLANE-NEXT:    vpunpckhbw {{.*#+}} xmm3 = xmm3[8],xmm4[8],xmm3[9],xmm4[9],xmm3[10],xmm4[10],xmm3[11],xmm4[11],xmm3[12],xmm4[12],xmm3[13],xmm4[13],xmm3[14],xmm4[14],xmm3[15],xmm4[15]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm3 = zero,zero,zero,xmm3[10,11],zero,zero,zero,xmm3[12,13],zero,zero,zero,xmm3[14,15],zero
; AVX2-FAST-PERLANE-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm0 = zero,xmm0[10,11],zero,zero,zero,xmm0[12,13],zero,zero,zero,xmm0[14,15],zero,zero,zero
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm3, %xmm0, %xmm0
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm1 = xmm2[12],zero,zero,zero,zero,xmm2[13],zero,zero,zero,zero,xmm2[14],zero,zero,zero,zero,xmm2[15]
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm1, %xmm0, %xmm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm0, 64(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm5, (%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm7, 32(%r9)
; AVX2-FAST-PERLANE-NEXT:    vzeroupper
; AVX2-FAST-PERLANE-NEXT:    retq
;
; AVX512F-SLOW-LABEL: store_i8_stride5_vf16:
; AVX512F-SLOW:       # %bb.0:
; AVX512F-SLOW-NEXT:    vmovdqa (%rdi), %xmm0
; AVX512F-SLOW-NEXT:    vmovdqa (%rsi), %xmm2
; AVX512F-SLOW-NEXT:    vmovdqa (%rdx), %xmm3
; AVX512F-SLOW-NEXT:    vmovdqa (%rcx), %xmm4
; AVX512F-SLOW-NEXT:    vmovdqa (%r8), %xmm1
; AVX512F-SLOW-NEXT:    vinserti128 $1, %xmm4, %ymm3, %ymm5
; AVX512F-SLOW-NEXT:    vinserti128 $1, %xmm2, %ymm0, %ymm6
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm7 = ymm6[u,u,u,7],zero,ymm6[u,u,u,8],zero,ymm6[u,u,u,9],zero,ymm6[u,u,u],zero,ymm6[26,u,u,u],zero,ymm6[27,u,u,u],zero,ymm6[28,u,u]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm6[2,3,0,1]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = ymm8[u,u,u],zero,ymm8[7,u,u,u],zero,ymm8[8,u,u,u],zero,ymm8[9,u,u,u,26],zero,ymm8[u,u,u,27],zero,ymm8[u,u,u,28],zero,ymm8[u,u]
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; AVX512F-SLOW-NEXT:    vpternlogq $50, %ymm7, %ymm9, %ymm8
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm7 = ymm5[0,2,0,2]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm7 = zero,zero,ymm7[0,8,u],zero,zero,ymm7[1,9,u],zero,zero,ymm7[2,10,u],zero,zero,ymm7[19,27,u],zero,zero,ymm7[20,28,u],zero,zero,ymm7[21,29,u],zero,zero
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm8, %zmm7, %zmm7
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = ymm5[6],zero,ymm5[u,u,u,7],zero,ymm5[u,u,u,8],zero,ymm5[u,u,u,9,25,u,u,u],zero,ymm5[26,u,u,u],zero,ymm5[27,u,u,u],zero,ymm5[28]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,3,0,1]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = zero,ymm5[6,u,u,u],zero,ymm5[7,u,u,u],zero,ymm5[8,u,u,u],zero,zero,ymm5[u,u,u,26],zero,ymm5[u,u,u,27],zero,ymm5[u,u,u,28],zero
; AVX512F-SLOW-NEXT:    vpternlogq $200, %ymm8, %ymm9, %ymm5
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,2,2,0]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm6 = ymm6[0,8],zero,zero,ymm6[u,1,9],zero,zero,ymm6[u,2,10],zero,zero,ymm6[u,3,19],zero,zero,ymm6[u,28,20],zero,zero,ymm6[u,29,21],zero,zero,ymm6[u,30,22]
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm5, %zmm6, %zmm5
; AVX512F-SLOW-NEXT:    vporq %zmm7, %zmm5, %zmm5
; AVX512F-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm6 = <u,0,0,0,0,u,1,1,1,1,u,2,2,2,2,u>
; AVX512F-SLOW-NEXT:    vpermd %zmm1, %zmm6, %zmm6
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm5, %zmm6
; AVX512F-SLOW-NEXT:    vpunpckhbw {{.*#+}} xmm3 = xmm3[8],xmm4[8],xmm3[9],xmm4[9],xmm3[10],xmm4[10],xmm3[11],xmm4[11],xmm3[12],xmm4[12],xmm3[13],xmm4[13],xmm3[14],xmm4[14],xmm3[15],xmm4[15]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm3 = xmm3[u],zero,zero,xmm3[10,11,u],zero,zero,xmm3[12,13,u],zero,zero,xmm3[14,15,u]
; AVX512F-SLOW-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm2[8],xmm0[9],xmm2[9],xmm0[10],xmm2[10],xmm0[11],xmm2[11],xmm0[12],xmm2[12],xmm0[13],xmm2[13],xmm0[14],xmm2[14],xmm0[15],xmm2[15]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[u,10,11],zero,zero,xmm0[u,12,13],zero,zero,xmm0[u,14,15],zero,zero,xmm0[u]
; AVX512F-SLOW-NEXT:    vpor %xmm3, %xmm0, %xmm0
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm0 = zero,xmm0[1,2,3,4],zero,xmm0[6,7,8,9],zero,xmm0[11,12,13,14],zero
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[12],zero,zero,zero,zero,xmm1[13],zero,zero,zero,zero,xmm1[14],zero,zero,zero,zero,xmm1[15]
; AVX512F-SLOW-NEXT:    vpor %xmm1, %xmm0, %xmm0
; AVX512F-SLOW-NEXT:    vmovdqa %xmm0, 64(%r9)
; AVX512F-SLOW-NEXT:    vmovdqa64 %zmm6, (%r9)
; AVX512F-SLOW-NEXT:    vzeroupper
; AVX512F-SLOW-NEXT:    retq
;
; AVX512F-FAST-LABEL: store_i8_stride5_vf16:
; AVX512F-FAST:       # %bb.0:
; AVX512F-FAST-NEXT:    vmovdqa (%rdi), %xmm1
; AVX512F-FAST-NEXT:    vmovdqa (%rsi), %xmm2
; AVX512F-FAST-NEXT:    vmovdqa (%rdx), %xmm3
; AVX512F-FAST-NEXT:    vmovdqa (%rcx), %xmm4
; AVX512F-FAST-NEXT:    vmovdqa (%r8), %xmm0
; AVX512F-FAST-NEXT:    vinserti128 $1, %xmm2, %ymm1, %ymm5
; AVX512F-FAST-NEXT:    vinserti128 $1, %xmm4, %ymm3, %ymm6
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm6[0,2,0,2]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm7 = zero,zero,ymm7[0,8,u],zero,zero,ymm7[1,9,u],zero,zero,ymm7[2,10,u],zero,zero,ymm7[19,27,u],zero,zero,ymm7[20,28,u],zero,zero,ymm7[21,29,u],zero,zero
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [1,5,2,6,2,6,3,7]
; AVX512F-FAST-NEXT:    vpermd %ymm5, %ymm8, %ymm8
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = zero,zero,ymm8[u,3,7],zero,zero,ymm8[u,8,12],zero,zero,ymm8[u,9,13],zero,zero,ymm8[u,18,22],zero,zero,ymm8[u,19,23],zero,zero,ymm8[u,24,28],zero,zero
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm8, %zmm7, %zmm7
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[0,2,2,0]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm5 = ymm5[0,8],zero,zero,ymm5[u,1,9],zero,zero,ymm5[u,2,10],zero,zero,ymm5[u,3,19],zero,zero,ymm5[u,28,20],zero,zero,ymm5[u,29,21],zero,zero,ymm5[u,30,22]
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [1,5,2,6,6,2,3,7]
; AVX512F-FAST-NEXT:    vpermd %ymm6, %ymm8, %ymm6
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm6 = ymm6[2,6,u],zero,zero,ymm6[3,7,u],zero,zero,ymm6[8,12,u],zero,zero,ymm6[9,17,u],zero,zero,ymm6[22,18,u],zero,zero,ymm6[23,19,u],zero,zero,ymm6[24,28]
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm6, %zmm5, %zmm5
; AVX512F-FAST-NEXT:    vporq %zmm7, %zmm5, %zmm5
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm0, %zmm0, %zmm6
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm7 = <u,0,0,0,0,u,1,1,9,9,u,10,10,10,10,u>
; AVX512F-FAST-NEXT:    vpermd %zmm6, %zmm7, %zmm6
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm5, %zmm6
; AVX512F-FAST-NEXT:    vpunpckhbw {{.*#+}} xmm3 = xmm3[8],xmm4[8],xmm3[9],xmm4[9],xmm3[10],xmm4[10],xmm3[11],xmm4[11],xmm3[12],xmm4[12],xmm3[13],xmm4[13],xmm3[14],xmm4[14],xmm3[15],xmm4[15]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm3 = xmm3[u],zero,zero,xmm3[10,11,u],zero,zero,xmm3[12,13,u],zero,zero,xmm3[14,15,u]
; AVX512F-FAST-NEXT:    vpunpckhbw {{.*#+}} xmm1 = xmm1[8],xmm2[8],xmm1[9],xmm2[9],xmm1[10],xmm2[10],xmm1[11],xmm2[11],xmm1[12],xmm2[12],xmm1[13],xmm2[13],xmm1[14],xmm2[14],xmm1[15],xmm2[15]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[u,10,11],zero,zero,xmm1[u,12,13],zero,zero,xmm1[u,14,15],zero,zero,xmm1[u]
; AVX512F-FAST-NEXT:    vpor %xmm3, %xmm1, %xmm1
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm1 = zero,xmm1[1,2,3,4],zero,xmm1[6,7,8,9],zero,xmm1[11,12,13,14],zero
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[12],zero,zero,zero,zero,xmm0[13],zero,zero,zero,zero,xmm0[14],zero,zero,zero,zero,xmm0[15]
; AVX512F-FAST-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX512F-FAST-NEXT:    vmovdqa %xmm0, 64(%r9)
; AVX512F-FAST-NEXT:    vmovdqa64 %zmm6, (%r9)
; AVX512F-FAST-NEXT:    vzeroupper
; AVX512F-FAST-NEXT:    retq
;
; AVX512BW-SLOW-LABEL: store_i8_stride5_vf16:
; AVX512BW-SLOW:       # %bb.0:
; AVX512BW-SLOW-NEXT:    vmovdqa (%rdi), %xmm0
; AVX512BW-SLOW-NEXT:    vmovdqa (%rsi), %xmm1
; AVX512BW-SLOW-NEXT:    vmovdqa (%rdx), %xmm2
; AVX512BW-SLOW-NEXT:    vmovdqa (%rcx), %xmm3
; AVX512BW-SLOW-NEXT:    vmovdqa (%r8), %xmm4
; AVX512BW-SLOW-NEXT:    vinserti128 $1, %xmm3, %ymm2, %ymm5
; AVX512BW-SLOW-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm6
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm7 = zero,zero,zero,ymm6[7],zero,zero,zero,zero,ymm6[8],zero,zero,zero,zero,ymm6[9],zero,zero,zero,zero,zero,ymm6[26],zero,zero,zero,zero,ymm6[27],zero,zero,zero,zero,ymm6[28],zero,zero
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm6[2,3,0,1]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = zero,zero,zero,zero,ymm8[7],zero,zero,zero,zero,ymm8[8],zero,zero,zero,zero,ymm8[9],zero,zero,zero,ymm8[26],zero,zero,zero,zero,ymm8[27],zero,zero,zero,zero,ymm8[28],zero,zero,zero
; AVX512BW-SLOW-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = ymm5[6],zero,zero,zero,zero,ymm5[7],zero,zero,zero,zero,ymm5[8],zero,zero,zero,zero,ymm5[9,25],zero,zero,zero,zero,ymm5[26],zero,zero,zero,zero,ymm5[27],zero,zero,zero,zero,ymm5[28]
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm5[2,3,0,1]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm9 = zero,ymm9[6],zero,zero,zero,zero,ymm9[7],zero,zero,zero,zero,ymm9[8],zero,zero,zero,zero,zero,zero,zero,zero,ymm9[26],zero,zero,zero,zero,ymm9[27],zero,zero,zero,zero,ymm9[28],zero
; AVX512BW-SLOW-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX512BW-SLOW-NEXT:    movl $831283992, %eax # imm = 0x318C6318
; AVX512BW-SLOW-NEXT:    kmovd %eax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %ymm7, %ymm8 {%k1}
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,2,2,0]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm6 = ymm6[0,8],zero,zero,zero,ymm6[1,9],zero,zero,zero,ymm6[2,10],zero,zero,zero,ymm6[3,19],zero,zero,zero,ymm6[28,20],zero,zero,zero,ymm6[29,21],zero,zero,zero,ymm6[30,22]
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[0,2,0,2]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = zero,zero,ymm5[0,8],zero,zero,zero,ymm5[1,9],zero,zero,zero,ymm5[2,10],zero,zero,zero,ymm5[19,27],zero,zero,zero,ymm5[20,28],zero,zero,zero,ymm5[21,29],zero,zero,zero
; AVX512BW-SLOW-NEXT:    vpor %ymm6, %ymm5, %ymm5
; AVX512BW-SLOW-NEXT:    vinserti64x4 $1, %ymm8, %zmm5, %zmm5
; AVX512BW-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm6 = [0,0,0,0,0,0,1,1,1,1,2,2,2,2,2,2]
; AVX512BW-SLOW-NEXT:    vpermd %zmm4, %zmm6, %zmm6
; AVX512BW-SLOW-NEXT:    movabsq $595056260442243600, %rax # imm = 0x842108421084210
; AVX512BW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %zmm6, %zmm5 {%k1}
; AVX512BW-SLOW-NEXT:    vpunpckhbw {{.*#+}} xmm2 = xmm2[8],xmm3[8],xmm2[9],xmm3[9],xmm2[10],xmm3[10],xmm2[11],xmm3[11],xmm2[12],xmm3[12],xmm2[13],xmm3[13],xmm2[14],xmm3[14],xmm2[15],xmm3[15]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm2 = zero,zero,zero,xmm2[10,11],zero,zero,zero,xmm2[12,13],zero,zero,zero,xmm2[14,15],zero
; AVX512BW-SLOW-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm0 = zero,xmm0[10,11],zero,zero,zero,xmm0[12,13],zero,zero,zero,xmm0[14,15],zero,zero,zero
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm1 = xmm4[12],zero,zero,zero,zero,xmm4[13],zero,zero,zero,zero,xmm4[14],zero,zero,zero,zero,xmm4[15]
; AVX512BW-SLOW-NEXT:    vpternlogq $254, %xmm2, %xmm0, %xmm1
; AVX512BW-SLOW-NEXT:    vmovdqa %xmm1, 64(%r9)
; AVX512BW-SLOW-NEXT:    vmovdqa64 %zmm5, (%r9)
; AVX512BW-SLOW-NEXT:    vzeroupper
; AVX512BW-SLOW-NEXT:    retq
;
; AVX512BW-FAST-LABEL: store_i8_stride5_vf16:
; AVX512BW-FAST:       # %bb.0:
; AVX512BW-FAST-NEXT:    vmovdqa (%rdi), %xmm0
; AVX512BW-FAST-NEXT:    vmovdqa (%rsi), %xmm1
; AVX512BW-FAST-NEXT:    vmovdqa (%rdx), %xmm2
; AVX512BW-FAST-NEXT:    vmovdqa (%rcx), %xmm3
; AVX512BW-FAST-NEXT:    vmovdqa (%r8), %xmm4
; AVX512BW-FAST-NEXT:    vinserti128 $1, %xmm1, %ymm0, %ymm5
; AVX512BW-FAST-NEXT:    vinserti128 $1, %xmm3, %ymm2, %ymm6
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} ymm7 = [1,5,2,6,6,2,3,7]
; AVX512BW-FAST-NEXT:    vpermd %ymm6, %ymm7, %ymm7
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm5[0,2,2,0]
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm7, %zmm8, %zmm7
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} zmm7 = zmm7[0,8],zero,zero,zero,zmm7[1,9],zero,zero,zero,zmm7[2,10],zero,zero,zero,zmm7[3,19],zero,zero,zero,zmm7[28,20],zero,zero,zero,zmm7[29,21],zero,zero,zero,zmm7[30,22,34,38],zero,zero,zero,zmm7[35,39],zero,zero,zero,zmm7[40,44],zero,zero,zero,zmm7[41,49],zero,zero,zero,zmm7[54,50],zero,zero,zero,zmm7[55,51],zero,zero,zero,zmm7[56,60]
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [1,5,2,6,2,6,3,7]
; AVX512BW-FAST-NEXT:    vpermd %ymm5, %ymm8, %ymm5
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,2,0,2]
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm5, %zmm6, %zmm5
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} zmm5 = zero,zero,zmm5[0,8],zero,zero,zero,zmm5[1,9],zero,zero,zero,zmm5[2,10],zero,zero,zero,zmm5[19,27],zero,zero,zero,zmm5[20,28],zero,zero,zero,zmm5[21,29],zero,zero,zero,zero,zero,zero,zmm5[35,39],zero,zero,zero,zmm5[40,44],zero,zero,zero,zmm5[41,45],zero,zero,zero,zmm5[50,54],zero,zero,zero,zmm5[51,55],zero,zero,zero,zmm5[56,60],zero,zero
; AVX512BW-FAST-NEXT:    vporq %zmm7, %zmm5, %zmm5
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm4, %zmm4, %zmm6
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm7 = [0,0,0,0,0,0,1,1,9,9,10,10,10,10,10,10]
; AVX512BW-FAST-NEXT:    vpermd %zmm6, %zmm7, %zmm6
; AVX512BW-FAST-NEXT:    movabsq $595056260442243600, %rax # imm = 0x842108421084210
; AVX512BW-FAST-NEXT:    kmovq %rax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm6, %zmm5 {%k1}
; AVX512BW-FAST-NEXT:    vpunpckhbw {{.*#+}} xmm2 = xmm2[8],xmm3[8],xmm2[9],xmm3[9],xmm2[10],xmm3[10],xmm2[11],xmm3[11],xmm2[12],xmm3[12],xmm2[13],xmm3[13],xmm2[14],xmm3[14],xmm2[15],xmm3[15]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm2 = zero,zero,zero,xmm2[10,11],zero,zero,zero,xmm2[12,13],zero,zero,zero,xmm2[14,15],zero
; AVX512BW-FAST-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm0 = zero,xmm0[10,11],zero,zero,zero,xmm0[12,13],zero,zero,zero,xmm0[14,15],zero,zero,zero
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm1 = xmm4[12],zero,zero,zero,zero,xmm4[13],zero,zero,zero,zero,xmm4[14],zero,zero,zero,zero,xmm4[15]
; AVX512BW-FAST-NEXT:    vpternlogq $254, %xmm2, %xmm0, %xmm1
; AVX512BW-FAST-NEXT:    vmovdqa %xmm1, 64(%r9)
; AVX512BW-FAST-NEXT:    vmovdqa64 %zmm5, (%r9)
; AVX512BW-FAST-NEXT:    vzeroupper
; AVX512BW-FAST-NEXT:    retq
  %in.vec0 = load <16 x i8>, ptr %in.vecptr0, align 64
  %in.vec1 = load <16 x i8>, ptr %in.vecptr1, align 64
  %in.vec2 = load <16 x i8>, ptr %in.vecptr2, align 64
  %in.vec3 = load <16 x i8>, ptr %in.vecptr3, align 64
  %in.vec4 = load <16 x i8>, ptr %in.vecptr4, align 64
  %1 = shufflevector <16 x i8> %in.vec0, <16 x i8> %in.vec1, <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 31>
  %2 = shufflevector <16 x i8> %in.vec2, <16 x i8> %in.vec3, <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 31>
  %3 = shufflevector <32 x i8> %1, <32 x i8> %2, <64 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
  %4 = shufflevector <16 x i8> %in.vec4, <16 x i8> poison, <64 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 undef, 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 undef, 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 undef, 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 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5 = shufflevector <64 x i8> %3, <64 x i8> %4, <80 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79>
  %interleaved.vec = shufflevector <80 x i8> %5, <80 x i8> poison, <80 x i32> <i32 0, i32 16, i32 32, i32 48, i32 64, i32 1, i32 17, i32 33, i32 49, i32 65, i32 2, i32 18, i32 34, i32 50, i32 66, i32 3, i32 19, i32 35, i32 51, i32 67, i32 4, i32 20, i32 36, i32 52, i32 68, i32 5, i32 21, i32 37, i32 53, i32 69, i32 6, i32 22, i32 38, i32 54, i32 70, i32 7, i32 23, i32 39, i32 55, i32 71, i32 8, i32 24, i32 40, i32 56, i32 72, i32 9, i32 25, i32 41, i32 57, i32 73, i32 10, i32 26, i32 42, i32 58, i32 74, i32 11, i32 27, i32 43, i32 59, i32 75, i32 12, i32 28, i32 44, i32 60, i32 76, i32 13, i32 29, i32 45, i32 61, i32 77, i32 14, i32 30, i32 46, i32 62, i32 78, i32 15, i32 31, i32 47, i32 63, i32 79>
  store <80 x i8> %interleaved.vec, ptr %out.vec, align 64
  ret void
}

define void @store_i8_stride5_vf32(ptr %in.vecptr0, ptr %in.vecptr1, ptr %in.vecptr2, ptr %in.vecptr3, ptr %in.vecptr4, ptr %out.vec) nounwind {
; SSE-LABEL: store_i8_stride5_vf32:
; SSE:       # %bb.0:
; SSE-NEXT:    subq $152, %rsp
; SSE-NEXT:    movdqa 16(%rdi), %xmm3
; SSE-NEXT:    movdqa (%rsi), %xmm9
; SSE-NEXT:    movdqa %xmm9, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa 16(%rsi), %xmm7
; SSE-NEXT:    movdqa (%rdx), %xmm2
; SSE-NEXT:    movdqa %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa 16(%rdx), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa (%rcx), %xmm11
; SSE-NEXT:    movdqa 16(%rcx), %xmm12
; SSE-NEXT:    movdqa 16(%r8), %xmm14
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,3,3,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa {{.*#+}} xmm13 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; SSE-NEXT:    pand %xmm13, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm12[2,1,2,3]
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[3,3,0,3]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa %xmm13, %xmm4
; SSE-NEXT:    pandn %xmm1, %xmm4
; SSE-NEXT:    por %xmm0, %xmm4
; SSE-NEXT:    movdqa {{.*#+}} xmm8 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255]
; SSE-NEXT:    pand %xmm8, %xmm4
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm3[1,1,2,2]
; SSE-NEXT:    movdqa %xmm3, %xmm15
; SSE-NEXT:    movdqa {{.*#+}} xmm3 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; SSE-NEXT:    pand %xmm3, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm7[0,1,2,1]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,7,6,7]
; SSE-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,0,1,0]
; SSE-NEXT:    movdqa %xmm3, %xmm5
; SSE-NEXT:    pandn %xmm1, %xmm5
; SSE-NEXT:    por %xmm0, %xmm5
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    por %xmm4, %xmm0
; SSE-NEXT:    movdqa {{.*#+}} xmm10 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; SSE-NEXT:    pand %xmm10, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm14[1,1,2,2]
; SSE-NEXT:    movdqa %xmm10, %xmm4
; SSE-NEXT:    pandn %xmm1, %xmm4
; SSE-NEXT:    por %xmm0, %xmm4
; SSE-NEXT:    movdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm2[3,3,3,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
; SSE-NEXT:    pand %xmm13, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm11[2,1,2,3]
; SSE-NEXT:    movdqa %xmm11, %xmm2
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[3,3,0,3]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa %xmm13, %xmm4
; SSE-NEXT:    pandn %xmm1, %xmm4
; SSE-NEXT:    por %xmm0, %xmm4
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm9[0,1,2,1]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,7,6,7]
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,1,0]
; SSE-NEXT:    movdqa %xmm3, %xmm1
; SSE-NEXT:    pandn %xmm0, %xmm1
; SSE-NEXT:    movdqa (%rdi), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,2,2]
; SSE-NEXT:    pand %xmm3, %xmm0
; SSE-NEXT:    por %xmm0, %xmm1
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    pandn %xmm1, %xmm0
; SSE-NEXT:    pand %xmm8, %xmm4
; SSE-NEXT:    por %xmm4, %xmm0
; SSE-NEXT:    movdqa (%r8), %xmm1
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,2,2]
; SSE-NEXT:    movdqa %xmm10, %xmm4
; SSE-NEXT:    pandn %xmm1, %xmm4
; SSE-NEXT:    pand %xmm10, %xmm0
; SSE-NEXT:    por %xmm0, %xmm4
; SSE-NEXT:    movdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa %xmm7, %xmm11
; SSE-NEXT:    movdqa %xmm7, %xmm0
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm7[8],xmm0[9],xmm7[9],xmm0[10],xmm7[10],xmm0[11],xmm7[11],xmm0[12],xmm7[12],xmm0[13],xmm7[13],xmm0[14],xmm7[14],xmm0[15],xmm7[15]
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[2,2,2,2,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,5,4,7]
; SSE-NEXT:    movdqa {{.*#+}} xmm9 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; SSE-NEXT:    movdqa %xmm9, %xmm4
; SSE-NEXT:    pandn %xmm0, %xmm4
; SSE-NEXT:    movdqa %xmm15, %xmm6
; SSE-NEXT:    movdqa %xmm15, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm15[2,2,3,3]
; SSE-NEXT:    pand %xmm9, %xmm0
; SSE-NEXT:    por %xmm0, %xmm4
; SSE-NEXT:    movdqa {{.*#+}} xmm8 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    pandn %xmm4, %xmm0
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm7 # 16-byte Reload
; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm7[0,1,2,3,5,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[2,2,2,2]
; SSE-NEXT:    movdqa %xmm3, %xmm5
; SSE-NEXT:    pandn %xmm4, %xmm5
; SSE-NEXT:    movdqa %xmm12, %xmm15
; SSE-NEXT:    punpckhbw {{.*#+}} xmm15 = xmm15[8],xmm12[8],xmm15[9],xmm12[9],xmm15[10],xmm12[10],xmm15[11],xmm12[11],xmm15[12],xmm12[12],xmm15[13],xmm12[13],xmm15[14],xmm12[14],xmm15[15],xmm12[15]
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm15[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[1,1,2,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,4,7,6,4]
; SSE-NEXT:    pand %xmm3, %xmm4
; SSE-NEXT:    por %xmm5, %xmm4
; SSE-NEXT:    pand %xmm8, %xmm4
; SSE-NEXT:    por %xmm0, %xmm4
; SSE-NEXT:    movdqa %xmm14, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm14[2,2,2,2]
; SSE-NEXT:    movdqa %xmm13, %xmm1
; SSE-NEXT:    pandn %xmm0, %xmm1
; SSE-NEXT:    pand %xmm13, %xmm4
; SSE-NEXT:    por %xmm4, %xmm1
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    punpcklbw {{.*#+}} xmm12 = xmm12[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    movdqa %xmm12, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm12[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,3,2,0,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,5,5,5]
; SSE-NEXT:    movdqa %xmm10, %xmm4
; SSE-NEXT:    pandn %xmm0, %xmm4
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm7[1,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE-NEXT:    pand %xmm10, %xmm0
; SSE-NEXT:    por %xmm0, %xmm4
; SSE-NEXT:    movdqa {{.*#+}} xmm1 = [255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; SSE-NEXT:    movdqa %xmm1, %xmm5
; SSE-NEXT:    pandn %xmm4, %xmm5
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm6[1,1,1,1]
; SSE-NEXT:    movdqa %xmm3, %xmm7
; SSE-NEXT:    pandn %xmm4, %xmm7
; SSE-NEXT:    movdqa %xmm11, %xmm0
; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm11[0],xmm0[1],xmm11[1],xmm0[2],xmm11[2],xmm0[3],xmm11[3],xmm0[4],xmm11[4],xmm0[5],xmm11[5],xmm0[6],xmm11[6],xmm0[7],xmm11[7]
; SSE-NEXT:    movdqa %xmm0, (%rsp) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,4,5,6,6]
; SSE-NEXT:    pand %xmm3, %xmm4
; SSE-NEXT:    por %xmm7, %xmm4
; SSE-NEXT:    pand %xmm1, %xmm4
; SSE-NEXT:    por %xmm5, %xmm4
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm14[0,0,1,1]
; SSE-NEXT:    movdqa %xmm9, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    pand %xmm9, %xmm4
; SSE-NEXT:    por %xmm4, %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm11 # 16-byte Reload
; SSE-NEXT:    movdqa %xmm11, %xmm0
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm11[8],xmm0[9],xmm11[9],xmm0[10],xmm11[10],xmm0[11],xmm11[11],xmm0[12],xmm11[12],xmm0[13],xmm11[13],xmm0[14],xmm11[14],xmm0[15],xmm11[15]
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm0[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[2,2,2,2,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,7,5,4,7]
; SSE-NEXT:    movdqa %xmm9, %xmm5
; SSE-NEXT:    pandn %xmm4, %xmm5
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm6[2,2,3,3]
; SSE-NEXT:    pand %xmm9, %xmm4
; SSE-NEXT:    por %xmm4, %xmm5
; SSE-NEXT:    movdqa %xmm8, %xmm4
; SSE-NEXT:    pandn %xmm5, %xmm4
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm0[0,1,2,3,5,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[2,2,2,2]
; SSE-NEXT:    movdqa %xmm3, %xmm7
; SSE-NEXT:    pandn %xmm5, %xmm7
; SSE-NEXT:    movdqa %xmm2, %xmm12
; SSE-NEXT:    punpckhbw {{.*#+}} xmm12 = xmm12[8],xmm2[8],xmm12[9],xmm2[9],xmm12[10],xmm2[10],xmm12[11],xmm2[11],xmm12[12],xmm2[12],xmm12[13],xmm2[13],xmm12[14],xmm2[14],xmm12[15],xmm2[15]
; SSE-NEXT:    pshufd {{.*#+}} xmm14 = xmm12[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm14 = xmm14[1,1,2,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm14 = xmm14[0,1,2,3,4,7,6,4]
; SSE-NEXT:    pand %xmm3, %xmm14
; SSE-NEXT:    por %xmm7, %xmm14
; SSE-NEXT:    pand %xmm8, %xmm14
; SSE-NEXT:    por %xmm4, %xmm14
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm5[2,2,2,2]
; SSE-NEXT:    movdqa %xmm13, %xmm7
; SSE-NEXT:    pandn %xmm4, %xmm7
; SSE-NEXT:    pand %xmm13, %xmm14
; SSE-NEXT:    por %xmm14, %xmm7
; SSE-NEXT:    movdqa %xmm7, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    punpcklbw {{.*#+}} xmm2 = xmm2[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    movdqa %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm2[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[0,3,2,0,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,5,5,5,5]
; SSE-NEXT:    movdqa %xmm10, %xmm14
; SSE-NEXT:    pandn %xmm4, %xmm14
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm0[1,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm4[0,0,0,0]
; SSE-NEXT:    pand %xmm10, %xmm4
; SSE-NEXT:    por %xmm4, %xmm14
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm6[1,1,1,1]
; SSE-NEXT:    movdqa %xmm3, %xmm2
; SSE-NEXT:    pandn %xmm4, %xmm2
; SSE-NEXT:    punpcklbw {{.*#+}} xmm11 = xmm11[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    movdqa %xmm11, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm4 = xmm11[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm4 = xmm4[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm4 = xmm4[0,1,2,3,4,5,6,6]
; SSE-NEXT:    pand %xmm3, %xmm4
; SSE-NEXT:    por %xmm2, %xmm4
; SSE-NEXT:    pand %xmm1, %xmm4
; SSE-NEXT:    pandn %xmm14, %xmm1
; SSE-NEXT:    por %xmm4, %xmm1
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm5[0,0,1,1]
; SSE-NEXT:    movdqa %xmm9, %xmm7
; SSE-NEXT:    pandn %xmm2, %xmm7
; SSE-NEXT:    pand %xmm9, %xmm1
; SSE-NEXT:    por %xmm1, %xmm7
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm15[0,1,2,3,5,7,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,2,3,2]
; SSE-NEXT:    movdqa %xmm3, %xmm2
; SSE-NEXT:    pandn %xmm0, %xmm2
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Reload
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm5[0,1,2,3,7,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,2,2,2]
; SSE-NEXT:    pand %xmm3, %xmm0
; SSE-NEXT:    por %xmm0, %xmm2
; SSE-NEXT:    movdqa {{.*#+}} xmm6 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255]
; SSE-NEXT:    movdqa %xmm6, %xmm0
; SSE-NEXT:    pandn %xmm2, %xmm0
; SSE-NEXT:    pshufhw $167, {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 # 16-byte Folded Reload
; SSE-NEXT:    # xmm2 = mem[0,1,2,3,7,5,6,6]
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[2,3,2,2]
; SSE-NEXT:    movdqa %xmm10, %xmm14
; SSE-NEXT:    pandn %xmm2, %xmm14
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm4 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm4[3,3,3,3]
; SSE-NEXT:    pand %xmm10, %xmm2
; SSE-NEXT:    por %xmm2, %xmm14
; SSE-NEXT:    pand %xmm6, %xmm14
; SSE-NEXT:    por %xmm0, %xmm14
; SSE-NEXT:    movdqa {{.*#+}} xmm11 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm1[3,3,3,3]
; SSE-NEXT:    movdqa %xmm11, %xmm15
; SSE-NEXT:    pandn %xmm0, %xmm15
; SSE-NEXT:    pand %xmm11, %xmm14
; SSE-NEXT:    por %xmm14, %xmm15
; SSE-NEXT:    pshuflw $225, {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Folded Reload
; SSE-NEXT:    # xmm0 = mem[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,1,0,1]
; SSE-NEXT:    movdqa %xmm9, %xmm2
; SSE-NEXT:    pandn %xmm0, %xmm2
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm5[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE-NEXT:    pand %xmm9, %xmm0
; SSE-NEXT:    por %xmm0, %xmm2
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    pandn %xmm2, %xmm0
; SSE-NEXT:    pshuflw $164, (%rsp), %xmm2 # 16-byte Folded Reload
; SSE-NEXT:    # xmm2 = mem[0,1,2,2,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,0,1,3]
; SSE-NEXT:    movdqa %xmm13, %xmm14
; SSE-NEXT:    pandn %xmm2, %xmm14
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm4[0,0,0,0]
; SSE-NEXT:    pand %xmm13, %xmm2
; SSE-NEXT:    por %xmm2, %xmm14
; SSE-NEXT:    pand %xmm8, %xmm14
; SSE-NEXT:    por %xmm0, %xmm14
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[0,0,0,0]
; SSE-NEXT:    movdqa %xmm3, %xmm0
; SSE-NEXT:    pandn %xmm2, %xmm0
; SSE-NEXT:    pand %xmm3, %xmm14
; SSE-NEXT:    por %xmm14, %xmm0
; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm12[0,1,2,3,5,7,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,2,3,2]
; SSE-NEXT:    movdqa %xmm3, %xmm5
; SSE-NEXT:    pandn %xmm2, %xmm5
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm4 # 16-byte Reload
; SSE-NEXT:    pshufhw {{.*#+}} xmm2 = xmm4[0,1,2,3,7,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[2,2,2,2]
; SSE-NEXT:    pand %xmm3, %xmm2
; SSE-NEXT:    por %xmm2, %xmm5
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm14 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm14[3,3,3,3]
; SSE-NEXT:    pand %xmm10, %xmm2
; SSE-NEXT:    pshufhw $167, {{[-0-9]+}}(%r{{[sb]}}p), %xmm12 # 16-byte Folded Reload
; SSE-NEXT:    # xmm12 = mem[0,1,2,3,7,5,6,6]
; SSE-NEXT:    pshufd {{.*#+}} xmm12 = xmm12[2,3,2,2]
; SSE-NEXT:    pandn %xmm12, %xmm10
; SSE-NEXT:    por %xmm2, %xmm10
; SSE-NEXT:    movdqa %xmm6, %xmm1
; SSE-NEXT:    pand %xmm6, %xmm10
; SSE-NEXT:    pandn %xmm5, %xmm1
; SSE-NEXT:    por %xmm10, %xmm1
; SSE-NEXT:    pand %xmm11, %xmm1
; SSE-NEXT:    movdqa %xmm1, %xmm5
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[3,3,3,3]
; SSE-NEXT:    pandn %xmm2, %xmm11
; SSE-NEXT:    por %xmm5, %xmm11
; SSE-NEXT:    pshuflw {{.*#+}} xmm2 = xmm4[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm2[0,0,0,0]
; SSE-NEXT:    pand %xmm9, %xmm2
; SSE-NEXT:    pshuflw $225, {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Folded Reload
; SSE-NEXT:    # xmm5 = mem[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,1,0,1]
; SSE-NEXT:    pandn %xmm5, %xmm9
; SSE-NEXT:    por %xmm2, %xmm9
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm14[0,0,0,0]
; SSE-NEXT:    pand %xmm13, %xmm2
; SSE-NEXT:    pshuflw $164, {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Folded Reload
; SSE-NEXT:    # xmm5 = mem[0,1,2,2,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,0,1,3]
; SSE-NEXT:    pandn %xmm5, %xmm13
; SSE-NEXT:    por %xmm2, %xmm13
; SSE-NEXT:    pand %xmm8, %xmm13
; SSE-NEXT:    pandn %xmm9, %xmm8
; SSE-NEXT:    por %xmm13, %xmm8
; SSE-NEXT:    pand %xmm3, %xmm8
; SSE-NEXT:    pshufd {{.*#+}} xmm2 = xmm1[0,0,0,0]
; SSE-NEXT:    pandn %xmm2, %xmm3
; SSE-NEXT:    por %xmm8, %xmm3
; SSE-NEXT:    movdqa %xmm3, (%r9)
; SSE-NEXT:    movdqa %xmm11, 64(%r9)
; SSE-NEXT:    movdqa %xmm0, 80(%r9)
; SSE-NEXT:    movdqa %xmm15, 144(%r9)
; SSE-NEXT:    movdqa %xmm7, 16(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 48(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 96(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 128(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 32(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 112(%r9)
; SSE-NEXT:    addq $152, %rsp
; SSE-NEXT:    retq
;
; AVX1-ONLY-LABEL: store_i8_stride5_vf32:
; AVX1-ONLY:       # %bb.0:
; AVX1-ONLY-NEXT:    vmovdqa 16(%rsi), %xmm10
; AVX1-ONLY-NEXT:    vmovdqa 16(%rdi), %xmm11
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm11[8],xmm10[8],xmm11[9],xmm10[9],xmm11[10],xmm10[10],xmm11[11],xmm10[11],xmm11[12],xmm10[12],xmm11[13],xmm10[13],xmm11[14],xmm10[14],xmm11[15],xmm10[15]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm5 = [0,10,11,14,15,0,12,13,0,10,11,14,15,0,12,13]
; AVX1-ONLY-NEXT:    # xmm5 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm5, %xmm0, %xmm1
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[u,u,4,5,u,u,u,6,7,u,u,u,8,9,u,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm1 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; AVX1-ONLY-NEXT:    vandnps %ymm0, %ymm1, %ymm2
; AVX1-ONLY-NEXT:    vmovaps %ymm1, %ymm7
; AVX1-ONLY-NEXT:    vmovdqa 16(%rcx), %xmm0
; AVX1-ONLY-NEXT:    vmovdqa 16(%rdx), %xmm1
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm3 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm4 = [12,13,0,10,11,14,15,0,12,13,0,10,11,14,15,0]
; AVX1-ONLY-NEXT:    # xmm4 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm4, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm6 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm6[2,u,u,u,5,4,u,u,u,7,6,u,u,u,9,8]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm3, %ymm6, %ymm3
; AVX1-ONLY-NEXT:    vandps %ymm7, %ymm3, %ymm3
; AVX1-ONLY-NEXT:    vorps %ymm2, %ymm3, %ymm2
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm2, %xmm3
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = zero,xmm3[1,2,3,4],zero,xmm3[6,7,8,9],zero,xmm3[11,12,13,14],zero
; AVX1-ONLY-NEXT:    vmovdqa 16(%r8), %xmm15
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm12 = [12,128,128,128,128,13,128,128,128,128,14,128,128,128,128,15]
; AVX1-ONLY-NEXT:    vpshufb %xmm12, %xmm15, %xmm6
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vmovdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm2[0],zero,xmm2[2,3,4,5],zero,xmm2[7,8,9,10],zero,xmm2[12,13,14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = zero,xmm15[9],zero,zero,zero,zero,xmm15[10],zero,zero,zero,zero,xmm15[11],zero,zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm3, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vmovdqa (%rcx), %xmm7
; AVX1-ONLY-NEXT:    vmovdqa (%rdx), %xmm8
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm2 = xmm8[8],xmm7[8],xmm8[9],xmm7[9],xmm8[10],xmm7[10],xmm8[11],xmm7[11],xmm8[12],xmm7[12],xmm8[13],xmm7[13],xmm8[14],xmm7[14],xmm8[15],xmm7[15]
; AVX1-ONLY-NEXT:    vpshufb %xmm4, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm6 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = xmm6[u,u,0,1,u,u,u,2,3,u,u,u,4,5,u,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm3, %ymm2, %ymm14
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm2 = xmm11[0],xmm10[0],xmm11[1],xmm10[1],xmm11[2],xmm10[2],xmm11[3],xmm10[3],xmm11[4],xmm10[4],xmm11[5],xmm10[5],xmm11[6],xmm10[6],xmm11[7],xmm10[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm13 = xmm2[0,1,u,u,u,2,3,u,u,u,4,5,u,u,u,6]
; AVX1-ONLY-NEXT:    vmovdqa (%rsi), %xmm2
; AVX1-ONLY-NEXT:    vmovdqa (%rdi), %xmm3
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm4 = xmm3[8],xmm2[8],xmm3[9],xmm2[9],xmm3[10],xmm2[10],xmm3[11],xmm2[11],xmm3[12],xmm2[12],xmm3[13],xmm2[13],xmm3[14],xmm2[14],xmm3[15],xmm2[15]
; AVX1-ONLY-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb %xmm5, %xmm4, %xmm5
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm13, %ymm5, %ymm5
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm13 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; AVX1-ONLY-NEXT:    vandnps %ymm14, %ymm13, %ymm14
; AVX1-ONLY-NEXT:    vandps %ymm5, %ymm13, %ymm5
; AVX1-ONLY-NEXT:    vorps %ymm5, %ymm14, %ymm5
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm5, %xmm14
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm14 = xmm14[0,1,2,3],zero,xmm14[5,6,7,8],zero,xmm14[10,11,12,13],zero,xmm14[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm9 = zero,zero,zero,zero,xmm15[0],zero,zero,zero,zero,xmm15[1],zero,zero,zero,zero,xmm15[2],zero
; AVX1-ONLY-NEXT:    vpor %xmm9, %xmm14, %xmm4
; AVX1-ONLY-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = zero,xmm5[1,2,3,4],zero,xmm5[6,7,8,9],zero,xmm5[11,12,13,14],zero
; AVX1-ONLY-NEXT:    vmovdqa (%r8), %xmm14
; AVX1-ONLY-NEXT:    vpshufb %xmm12, %xmm14, %xmm9
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm9, %xmm4
; AVX1-ONLY-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm0 = zero,xmm0[6,u,u,u],zero,xmm0[7,u,u,u],zero,xmm0[8,u,u,u],zero
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[6],zero,xmm1[u,u,u,7],zero,xmm1[u,u,u,8],zero,xmm1[u,u,u,9]
; AVX1-ONLY-NEXT:    vpor %xmm0, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm9 = [0,6,7,10,11,0,8,9,0,6,7,10,11,0,8,9]
; AVX1-ONLY-NEXT:    # xmm9 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm9, %xmm6, %xmm5
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm1, %ymm5, %ymm1
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = xmm10[u,u,u],zero,xmm10[7,u,u,u],zero,xmm10[8,u,u,u],zero,xmm10[9,u]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm6 = [8,128,0,7,128,9,128,0,8,128,0,7,128,9,128,0]
; AVX1-ONLY-NEXT:    # xmm6 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm6, %xmm11, %xmm12
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm12, %xmm5
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm10 = xmm10[0],xmm11[0],xmm10[1],xmm11[1],xmm10[2],xmm11[2],xmm10[3],xmm11[3],xmm10[4],xmm11[4],xmm10[5],xmm11[5],xmm10[6],xmm11[6],xmm10[7],xmm11[7]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm12 = [6,11,10,0,9,8,13,12,6,11,10,0,9,8,13,12]
; AVX1-ONLY-NEXT:    # xmm12 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm12, %xmm10, %xmm10
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm5, %ymm10, %ymm5
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm10 = [255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0]
; AVX1-ONLY-NEXT:    vandnps %ymm1, %ymm10, %ymm1
; AVX1-ONLY-NEXT:    vandps %ymm5, %ymm10, %ymm5
; AVX1-ONLY-NEXT:    vorps %ymm1, %ymm5, %ymm5
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm10 = [0,1,2,128,4,5,6,7,128,9,10,11,12,128,14,15]
; AVX1-ONLY-NEXT:    vpshufb %xmm10, %xmm5, %xmm1
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm4 = [128,128,128,3,128,128,128,128,4,128,128,128,128,5,128,128]
; AVX1-ONLY-NEXT:    vpshufb %xmm4, %xmm15, %xmm11
; AVX1-ONLY-NEXT:    vpor %xmm1, %xmm11, %xmm1
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm5, %xmm5
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm11 = [0,1,128,3,4,5,6,128,8,9,10,11,128,13,14,15]
; AVX1-ONLY-NEXT:    vpshufb %xmm11, %xmm5, %xmm0
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm5 = [128,128,6,128,128,128,128,7,128,128,128,128,8,128,128,128]
; AVX1-ONLY-NEXT:    vpshufb %xmm5, %xmm15, %xmm15
; AVX1-ONLY-NEXT:    vpor %xmm0, %xmm15, %xmm15
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm8[0],xmm7[0],xmm8[1],xmm7[1],xmm8[2],xmm7[2],xmm8[3],xmm7[3],xmm8[4],xmm7[4],xmm8[5],xmm7[5],xmm8[6],xmm7[6],xmm8[7],xmm7[7]
; AVX1-ONLY-NEXT:    vpshufb %xmm9, %xmm0, %xmm9
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[u,u,0,1,u,u,u,2,3,u,u,u,4,5,u,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm9, %ymm0, %ymm0
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm9 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3],xmm2[4],xmm3[4],xmm2[5],xmm3[5],xmm2[6],xmm3[6],xmm2[7],xmm3[7]
; AVX1-ONLY-NEXT:    vpshufb %xmm12, %xmm9, %xmm9
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm12 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3],xmm3[4],xmm2[4],xmm3[5],xmm2[5],xmm3[6],xmm2[6],xmm3[7],xmm2[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm12 = xmm12[0,1,u,u,u,2,3,u,u,u,4,5,u,u,u,6]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm9, %ymm12, %ymm9
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm12 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; AVX1-ONLY-NEXT:    vandnps %ymm0, %ymm12, %ymm0
; AVX1-ONLY-NEXT:    vandps %ymm12, %ymm9, %ymm9
; AVX1-ONLY-NEXT:    vorps %ymm0, %ymm9, %ymm9
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm9, %xmm0
; AVX1-ONLY-NEXT:    vpshufb %xmm10, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vpshufb %xmm4, %xmm14, %xmm4
; AVX1-ONLY-NEXT:    vpor %xmm4, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm9[0,1,2,3],zero,xmm9[5,6,7,8],zero,xmm9[10,11,12,13],zero,xmm9[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm9 = zero,zero,zero,zero,xmm14[0],zero,zero,zero,zero,xmm14[1],zero,zero,zero,zero,xmm14[2],zero
; AVX1-ONLY-NEXT:    vpor %xmm4, %xmm9, %xmm4
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm2[u,u,u],zero,xmm2[7,u,u,u],zero,xmm2[8,u,u,u],zero,xmm2[9,u]
; AVX1-ONLY-NEXT:    vpshufb %xmm6, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vpor %xmm2, %xmm3, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Reload
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = xmm3[u,u,4,5,u,u,u,6,7,u,u,u,8,9,u,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm3, %ymm2, %ymm2
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = zero,xmm7[6,u,u,u],zero,xmm7[7,u,u,u],zero,xmm7[8,u,u,u],zero
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm8[6],zero,xmm8[u,u,u,7],zero,xmm8[u,u,u,8],zero,xmm8[u,u,u,9]
; AVX1-ONLY-NEXT:    vpor %xmm3, %xmm6, %xmm3
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm6 = xmm7[8],xmm8[8],xmm7[9],xmm8[9],xmm7[10],xmm8[10],xmm7[11],xmm8[11],xmm7[12],xmm8[12],xmm7[13],xmm8[13],xmm7[14],xmm8[14],xmm7[15],xmm8[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm6[2,u,u,u,5,4,u,u,u,7,6,u,u,u,9,8]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm6, %ymm3, %ymm3
; AVX1-ONLY-NEXT:    vandnps %ymm2, %ymm13, %ymm2
; AVX1-ONLY-NEXT:    vandps %ymm3, %ymm13, %ymm3
; AVX1-ONLY-NEXT:    vorps %ymm2, %ymm3, %ymm2
; AVX1-ONLY-NEXT:    vpshufb %xmm11, %xmm2, %xmm3
; AVX1-ONLY-NEXT:    vpshufb %xmm5, %xmm14, %xmm5
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm2, %xmm2
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm2[0],zero,xmm2[2,3,4,5],zero,xmm2[7,8,9,10],zero,xmm2[12,13,14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = zero,xmm14[9],zero,zero,zero,zero,xmm14[10],zero,zero,zero,zero,xmm14[11],zero,zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa %xmm2, 48(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm3, 32(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm4, (%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm0, 16(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm15, 112(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm1, 96(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 64(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 80(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 128(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 144(%r9)
; AVX1-ONLY-NEXT:    vzeroupper
; AVX1-ONLY-NEXT:    retq
;
; AVX2-SLOW-LABEL: store_i8_stride5_vf32:
; AVX2-SLOW:       # %bb.0:
; AVX2-SLOW-NEXT:    vmovdqa (%rdi), %ymm3
; AVX2-SLOW-NEXT:    vmovdqa (%rsi), %ymm4
; AVX2-SLOW-NEXT:    vmovdqa (%rdx), %ymm1
; AVX2-SLOW-NEXT:    vmovdqa (%rcx), %ymm2
; AVX2-SLOW-NEXT:    vmovdqa (%r8), %ymm0
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,25,u,27,u,u,26,u,28,u,30,u,u,29,u,31,u]
; AVX2-SLOW-NEXT:    vpshufhw {{.*#+}} ymm6 = ymm1[0,1,2,3,5,6,7,6,8,9,10,11,13,14,15,14]
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm6 = ymm6[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vbroadcasti128 {{.*#+}} ymm7 = [255,0,255,0,0,255,0,255,0,255,0,0,255,0,255,0,255,0,255,0,0,255,0,255,0,255,0,0,255,0,255,0]
; AVX2-SLOW-NEXT:    # ymm7 = mem[0,1,0,1]
; AVX2-SLOW-NEXT:    vpblendvb %ymm7, %ymm5, %ymm6, %ymm5
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm6 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,27,u,u,26,u,28,u,u,u,u,29,u,31,u,u,30]
; AVX2-SLOW-NEXT:    vpshufhw {{.*#+}} ymm7 = ymm3[0,1,2,3,6,5,6,7,8,9,10,11,14,13,14,15]
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm7 = ymm7[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vbroadcasti128 {{.*#+}} ymm8 = [255,0,0,255,0,255,0,0,0,0,255,0,255,0,0,255,255,0,0,255,0,255,0,0,0,0,255,0,255,0,0,255]
; AVX2-SLOW-NEXT:    # ymm8 = mem[0,1,0,1]
; AVX2-SLOW-NEXT:    vpblendvb %ymm8, %ymm6, %ymm7, %ymm6
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[2,2,3,3]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm7 = <255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u>
; AVX2-SLOW-NEXT:    vpblendvb %ymm7, %ymm5, %ymm6, %ymm5
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm6 = ymm0[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[2,2,3,3]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm7 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; AVX2-SLOW-NEXT:    vpblendvb %ymm7, %ymm5, %ymm6, %ymm5
; AVX2-SLOW-NEXT:    vmovdqa (%rsi), %xmm7
; AVX2-SLOW-NEXT:    vmovdqa (%rdi), %xmm8
; AVX2-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm6 = xmm8[0],xmm7[0],xmm8[1],xmm7[1],xmm8[2],xmm7[2],xmm8[3],xmm7[3],xmm8[4],xmm7[4],xmm8[5],xmm7[5],xmm8[6],xmm7[6],xmm8[7],xmm7[7]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm6 = xmm6[0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,0,1,1]
; AVX2-SLOW-NEXT:    vmovdqa (%rdx), %xmm9
; AVX2-SLOW-NEXT:    vmovdqa (%rcx), %xmm10
; AVX2-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm11 = xmm10[0],xmm9[0],xmm10[1],xmm9[1],xmm10[2],xmm9[2],xmm10[3],xmm9[3],xmm10[4],xmm9[4],xmm10[5],xmm9[5],xmm10[6],xmm9[6],xmm10[7],xmm9[7]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm11 = xmm11[2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm11 = ymm11[0,0,1,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm12 = <255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm12, %ymm6, %ymm11, %ymm6
; AVX2-SLOW-NEXT:    vmovdqa (%r8), %xmm11
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm12 = xmm11[0,0,1,1]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm12 = ymm12[0,0,0,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm13 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm13, %ymm6, %ymm12, %ymm6
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm10 = zero,xmm10[6],zero,xmm10[8,u],zero,xmm10[7],zero,xmm10[9],zero,xmm10[11,u],zero,xmm10[10],zero,xmm10[12]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm9 = xmm9[6],zero,xmm9[8],zero,xmm9[u,7],zero,xmm9[9],zero,xmm9[11],zero,xmm9[u,10],zero,xmm9[12],zero
; AVX2-SLOW-NEXT:    vpor %xmm10, %xmm9, %xmm9
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[0,0,1,1]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm8 = xmm8[8],zero,xmm8[u,7],zero,xmm8[9],zero,xmm8[u],zero,xmm8[u,10],zero,xmm8[12],zero,xmm8[u,11]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} xmm7 = zero,xmm7[8,u],zero,xmm7[7],zero,xmm7[9,u,11,u],zero,xmm7[10],zero,xmm7[12,u],zero
; AVX2-SLOW-NEXT:    vpor %xmm7, %xmm8, %xmm7
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[0,0,1,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm8 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm8, %ymm9, %ymm7, %ymm7
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm8 = xmm11[1,1,2,2]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[0,1,1,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm3[21],zero,ymm3[19,20],zero,ymm3[22],zero,ymm3[24],zero,ymm3[22,23],zero,ymm3[25],zero,ymm3[23]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm9 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19],zero,ymm4[21],zero,zero,ymm4[20],zero,ymm4[22],zero,ymm4[24],zero,zero,ymm4[23],zero,ymm4[25],zero
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-SLOW-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm9 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm2[19],zero,ymm2[21],zero,zero,ymm2[20],zero,ymm2[22],zero,ymm2[24],zero,zero,ymm2[23],zero
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm10 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm1[19],zero,ymm1[21],zero,zero,ymm1[20],zero,ymm1[22],zero,ymm1[24],zero,zero,ymm1[23],zero,ymm1[25]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[2,2,3,3]
; AVX2-SLOW-NEXT:    vpor %ymm9, %ymm10, %ymm9
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm10 = <255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0>
; AVX2-SLOW-NEXT:    vpblendvb %ymm10, %ymm8, %ymm9, %ymm8
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm9 = ymm0[0,2,1,1,4,6,5,5]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,3,3,2]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm10 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm10, %ymm8, %ymm9, %ymm8
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm9 = <3,3,3,u,4,4,4,4>
; AVX2-SLOW-NEXT:    vpermd %ymm3, %ymm9, %ymm3
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm4 = ymm4[u,u,13,u,u,u,u,14,u,u,u,u,15,u,u,u,u,16,u,u,u,u,17,u,u,u,u,18,u,u,u,u]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm9 = <u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm9, %ymm3, %ymm4, %ymm3
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,zero,zero,ymm2[13],zero,zero,zero,zero,ymm2[14],zero,zero,zero,zero,ymm2[15],zero,zero,zero,zero,ymm2[16],zero,zero,zero,zero,ymm2[17],zero,zero,zero,zero,ymm2[18],zero,zero
; AVX2-SLOW-NEXT:    vpshufb {{.*#+}} ymm1 = zero,zero,zero,ymm1[13],zero,zero,zero,zero,ymm1[14],zero,zero,zero,zero,ymm1[15],zero,zero,zero,zero,ymm1[16],zero,zero,zero,zero,ymm1[17],zero,zero,zero,zero,ymm1[18],zero,zero,zero
; AVX2-SLOW-NEXT:    vpor %ymm2, %ymm1, %ymm1
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm2 = <u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm2, %ymm3, %ymm1, %ymm1
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm2 = <3,3,3,3,u,4,4,4>
; AVX2-SLOW-NEXT:    vpermd %ymm0, %ymm2, %ymm0
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm2 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
; AVX2-SLOW-NEXT:    vmovdqa %ymm0, 64(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm8, 96(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm5, 128(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm7, 32(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm6, (%r9)
; AVX2-SLOW-NEXT:    vzeroupper
; AVX2-SLOW-NEXT:    retq
;
; AVX2-FAST-LABEL: store_i8_stride5_vf32:
; AVX2-FAST:       # %bb.0:
; AVX2-FAST-NEXT:    vmovdqa (%rdi), %ymm3
; AVX2-FAST-NEXT:    vmovdqa (%rsi), %ymm4
; AVX2-FAST-NEXT:    vmovdqa (%rdx), %ymm1
; AVX2-FAST-NEXT:    vmovdqa (%rcx), %ymm2
; AVX2-FAST-NEXT:    vmovdqa (%r8), %ymm0
; AVX2-FAST-NEXT:    vmovdqa (%rsi), %xmm6
; AVX2-FAST-NEXT:    vmovdqa (%rdi), %xmm7
; AVX2-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm5 = xmm7[0],xmm6[0],xmm7[1],xmm6[1],xmm7[2],xmm6[2],xmm7[3],xmm6[3],xmm7[4],xmm6[4],xmm7[5],xmm6[5],xmm7[6],xmm6[6],xmm7[7],xmm6[7]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm5 = xmm5[0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[0,0,1,1]
; AVX2-FAST-NEXT:    vmovdqa (%rdx), %xmm8
; AVX2-FAST-NEXT:    vmovdqa (%rcx), %xmm9
; AVX2-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm10 = xmm9[0],xmm8[0],xmm9[1],xmm8[1],xmm9[2],xmm8[2],xmm9[3],xmm8[3],xmm9[4],xmm8[4],xmm9[5],xmm8[5],xmm9[6],xmm8[6],xmm9[7],xmm8[7]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm10 = xmm10[2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[0,0,1,1]
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm11 = <255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255>
; AVX2-FAST-NEXT:    vpblendvb %ymm11, %ymm5, %ymm10, %ymm5
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm10 = [0,0,0,0,0,0,1,1]
; AVX2-FAST-NEXT:    vpermd %ymm0, %ymm10, %ymm10
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm11 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm11, %ymm5, %ymm10, %ymm5
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm9 = zero,xmm9[6],zero,xmm9[8,u],zero,xmm9[7],zero,xmm9[9],zero,xmm9[11,u],zero,xmm9[10],zero,xmm9[12]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm8 = xmm8[6],zero,xmm8[8],zero,xmm8[u,7],zero,xmm8[9],zero,xmm8[11],zero,xmm8[u,10],zero,xmm8[12],zero
; AVX2-FAST-NEXT:    vpor %xmm9, %xmm8, %xmm8
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[0,0,1,1]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm7 = xmm7[8],zero,xmm7[u,7],zero,xmm7[9],zero,xmm7[u],zero,xmm7[u,10],zero,xmm7[12],zero,xmm7[u,11]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} xmm6 = zero,xmm6[8,u],zero,xmm6[7],zero,xmm6[9,u,11,u],zero,xmm6[10],zero,xmm6[12,u],zero
; AVX2-FAST-NEXT:    vpor %xmm7, %xmm6, %xmm6
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,0,1,1]
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm7 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255>
; AVX2-FAST-NEXT:    vpblendvb %ymm7, %ymm8, %ymm6, %ymm6
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm7 = [1,1,2,2,2,2,2,2]
; AVX2-FAST-NEXT:    vpermd %ymm0, %ymm7, %ymm7
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm8, %ymm6, %ymm7, %ymm6
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm7 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm3[21],zero,ymm3[19,20],zero,ymm3[22],zero,ymm3[24],zero,ymm3[22,23],zero,ymm3[25],zero,ymm3[23]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19],zero,ymm4[21],zero,zero,ymm4[20],zero,ymm4[22],zero,ymm4[24],zero,zero,ymm4[23],zero,ymm4[25],zero
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm2[19],zero,ymm2[21],zero,zero,ymm2[20],zero,ymm2[22],zero,ymm2[24],zero,zero,ymm2[23],zero
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm9 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm1[19],zero,ymm1[21],zero,zero,ymm1[20],zero,ymm1[22],zero,ymm1[24],zero,zero,ymm1[23],zero,ymm1[25]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm9 = <255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0>
; AVX2-FAST-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [4,6,5,5,5,5,4,6]
; AVX2-FAST-NEXT:    vpermd %ymm0, %ymm8, %ymm8
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm1[27],zero,ymm1[29,26],zero,ymm1[28],zero,ymm1[30],zero,ymm1[28,29],zero,ymm1[31],zero,ymm1[29]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm9 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,25],zero,ymm2[27],zero,zero,ymm2[26],zero,ymm2[28],zero,ymm2[30],zero,zero,ymm2[29],zero,ymm2[31],zero
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm9 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm3[29,26],zero,ymm3[28],zero,ymm3[26,27,28,29],zero,ymm3[31],zero,ymm3[29,30],zero
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm10 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,27],zero,zero,ymm4[26],zero,ymm4[28],zero,zero,zero,zero,ymm4[29],zero,ymm4[31],zero,zero,ymm4[30]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm9, %ymm10, %ymm9
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm10 = <255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u>
; AVX2-FAST-NEXT:    vpblendvb %ymm10, %ymm8, %ymm9, %ymm8
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm9 = [6,6,6,6,7,7,7,7]
; AVX2-FAST-NEXT:    vpermd %ymm0, %ymm9, %ymm9
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm10 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; AVX2-FAST-NEXT:    vpblendvb %ymm10, %ymm8, %ymm9, %ymm8
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm9 = <3,3,3,u,4,4,4,4>
; AVX2-FAST-NEXT:    vpermd %ymm3, %ymm9, %ymm3
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm4 = ymm4[u,u,13,u,u,u,u,14,u,u,u,u,15,u,u,u,u,16,u,u,u,u,17,u,u,u,u,18,u,u,u,u]
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm9 = <u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255>
; AVX2-FAST-NEXT:    vpblendvb %ymm9, %ymm3, %ymm4, %ymm3
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,zero,zero,ymm2[13],zero,zero,zero,zero,ymm2[14],zero,zero,zero,zero,ymm2[15],zero,zero,zero,zero,ymm2[16],zero,zero,zero,zero,ymm2[17],zero,zero,zero,zero,ymm2[18],zero,zero
; AVX2-FAST-NEXT:    vpshufb {{.*#+}} ymm1 = zero,zero,zero,ymm1[13],zero,zero,zero,zero,ymm1[14],zero,zero,zero,zero,ymm1[15],zero,zero,zero,zero,ymm1[16],zero,zero,zero,zero,ymm1[17],zero,zero,zero,zero,ymm1[18],zero,zero,zero
; AVX2-FAST-NEXT:    vpor %ymm2, %ymm1, %ymm1
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm2 = <u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255>
; AVX2-FAST-NEXT:    vpblendvb %ymm2, %ymm3, %ymm1, %ymm1
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm2 = <3,3,3,3,u,4,4,4>
; AVX2-FAST-NEXT:    vpermd %ymm0, %ymm2, %ymm0
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm2 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
; AVX2-FAST-NEXT:    vmovdqa %ymm0, 64(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm8, 128(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm7, 96(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm6, 32(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm5, (%r9)
; AVX2-FAST-NEXT:    vzeroupper
; AVX2-FAST-NEXT:    retq
;
; AVX2-FAST-PERLANE-LABEL: store_i8_stride5_vf32:
; AVX2-FAST-PERLANE:       # %bb.0:
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdi), %ymm3
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rsi), %ymm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdx), %ymm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rcx), %ymm2
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%r8), %ymm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rsi), %xmm6
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdi), %xmm7
; AVX2-FAST-PERLANE-NEXT:    vpunpcklbw {{.*#+}} xmm5 = xmm7[0],xmm6[0],xmm7[1],xmm6[1],xmm7[2],xmm6[2],xmm7[3],xmm6[3],xmm7[4],xmm6[4],xmm7[5],xmm6[5],xmm7[6],xmm6[6],xmm7[7],xmm6[7]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm5 = xmm5[0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdx), %xmm8
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rcx), %xmm9
; AVX2-FAST-PERLANE-NEXT:    vpunpcklbw {{.*#+}} xmm10 = xmm9[0],xmm8[0],xmm9[1],xmm8[1],xmm9[2],xmm8[2],xmm9[3],xmm8[3],xmm9[4],xmm8[4],xmm9[5],xmm8[5],xmm9[6],xmm8[6],xmm9[7],xmm8[7]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm10 = xmm10[2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm11 = <255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm11, %ymm5, %ymm10, %ymm5
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%r8), %xmm10
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} xmm11 = xmm10[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm11 = ymm11[0,0,0,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm12 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm12, %ymm5, %ymm11, %ymm5
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm9 = zero,xmm9[6],zero,xmm9[8,u],zero,xmm9[7],zero,xmm9[9],zero,xmm9[11,u],zero,xmm9[10],zero,xmm9[12]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm8 = xmm8[6],zero,xmm8[8],zero,xmm8[u,7],zero,xmm8[9],zero,xmm8[11],zero,xmm8[u,10],zero,xmm8[12],zero
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm9, %xmm8, %xmm8
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm7 = xmm7[8],zero,xmm7[u,7],zero,xmm7[9],zero,xmm7[u],zero,xmm7[u,10],zero,xmm7[12],zero,xmm7[u,11]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} xmm6 = zero,xmm6[8,u],zero,xmm6[7],zero,xmm6[9,u,11,u],zero,xmm6[10],zero,xmm6[12,u],zero
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm7, %xmm6, %xmm6
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm7 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm7, %ymm8, %ymm6, %ymm6
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} xmm7 = xmm10[1,1,2,2]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[0,1,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm8 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm8, %ymm6, %ymm7, %ymm6
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm7 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm3[21],zero,ymm3[19,20],zero,ymm3[22],zero,ymm3[24],zero,ymm3[22,23],zero,ymm3[25],zero,ymm3[23]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm8 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19],zero,ymm4[21],zero,zero,ymm4[20],zero,ymm4[22],zero,ymm4[24],zero,zero,ymm4[23],zero,ymm4[25],zero
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm8 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm2[19],zero,ymm2[21],zero,zero,ymm2[20],zero,ymm2[22],zero,ymm2[24],zero,zero,ymm2[23],zero
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm9 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm1[19],zero,ymm1[21],zero,zero,ymm1[20],zero,ymm1[22],zero,ymm1[24],zero,zero,ymm1[23],zero,ymm1[25]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm9 = <255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} ymm8 = ymm0[0,2,1,1,4,6,5,5]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,3,3,2]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm9, %ymm7, %ymm8, %ymm7
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm8 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm1[27],zero,ymm1[29,26],zero,ymm1[28],zero,ymm1[30],zero,ymm1[28,29],zero,ymm1[31],zero,ymm1[29]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm9 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,25],zero,ymm2[27],zero,zero,ymm2[26],zero,ymm2[28],zero,ymm2[30],zero,zero,ymm2[29],zero,ymm2[31],zero
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm9 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm3[29,26],zero,ymm3[28],zero,ymm3[26,27,28,29],zero,ymm3[31],zero,ymm3[29,30],zero
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm10 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,27],zero,zero,ymm4[26],zero,ymm4[28],zero,zero,zero,zero,ymm4[29],zero,ymm4[31],zero,zero,ymm4[30]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm9, %ymm10, %ymm9
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm10 = <255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm10, %ymm8, %ymm9, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} ymm9 = ymm0[2,2,3,3,6,6,7,7]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm10 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm10, %ymm8, %ymm9, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm9 = <3,3,3,u,4,4,4,4>
; AVX2-FAST-PERLANE-NEXT:    vpermd %ymm3, %ymm9, %ymm3
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm4 = ymm4[u,u,13,u,u,u,u,14,u,u,u,u,15,u,u,u,u,16,u,u,u,u,17,u,u,u,u,18,u,u,u,u]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm9 = <u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm9, %ymm3, %ymm4, %ymm3
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm2 = zero,zero,zero,zero,ymm2[13],zero,zero,zero,zero,ymm2[14],zero,zero,zero,zero,ymm2[15],zero,zero,zero,zero,ymm2[16],zero,zero,zero,zero,ymm2[17],zero,zero,zero,zero,ymm2[18],zero,zero
; AVX2-FAST-PERLANE-NEXT:    vpshufb {{.*#+}} ymm1 = zero,zero,zero,ymm1[13],zero,zero,zero,zero,ymm1[14],zero,zero,zero,zero,ymm1[15],zero,zero,zero,zero,ymm1[16],zero,zero,zero,zero,ymm1[17],zero,zero,zero,zero,ymm1[18],zero,zero,zero
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm2, %ymm1, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm2 = <u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm2, %ymm3, %ymm1, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm2 = <3,3,3,3,u,4,4,4>
; AVX2-FAST-PERLANE-NEXT:    vpermd %ymm0, %ymm2, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm2 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm2, %ymm1, %ymm0, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm0, 64(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm8, 128(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm7, 96(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm6, 32(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm5, (%r9)
; AVX2-FAST-PERLANE-NEXT:    vzeroupper
; AVX2-FAST-PERLANE-NEXT:    retq
;
; AVX512F-SLOW-LABEL: store_i8_stride5_vf32:
; AVX512F-SLOW:       # %bb.0:
; AVX512F-SLOW-NEXT:    vmovdqa (%rdi), %ymm1
; AVX512F-SLOW-NEXT:    vmovdqa (%rsi), %ymm2
; AVX512F-SLOW-NEXT:    vmovdqa (%rdx), %ymm3
; AVX512F-SLOW-NEXT:    vmovdqa (%rcx), %ymm4
; AVX512F-SLOW-NEXT:    vmovdqa (%r8), %ymm0
; AVX512F-SLOW-NEXT:    vmovdqa (%rdi), %xmm5
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm6 = xmm5[8],zero,xmm5[u,7],zero,xmm5[9],zero,xmm5[u],zero,xmm5[u,10],zero,xmm5[12],zero,xmm5[u,11]
; AVX512F-SLOW-NEXT:    vmovdqa (%rsi), %xmm7
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm8 = zero,xmm7[8,u],zero,xmm7[7],zero,xmm7[9,u,11,u],zero,xmm7[10],zero,xmm7[12,u],zero
; AVX512F-SLOW-NEXT:    vpor %xmm6, %xmm8, %xmm6
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,0,1,1]
; AVX512F-SLOW-NEXT:    vmovdqa (%rcx), %xmm8
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm9 = zero,xmm8[6],zero,xmm8[8,u],zero,xmm8[7],zero,xmm8[9],zero,xmm8[11,u],zero,xmm8[10],zero,xmm8[12]
; AVX512F-SLOW-NEXT:    vmovdqa (%rdx), %xmm10
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm11 = xmm10[6],zero,xmm10[8],zero,xmm10[u,7],zero,xmm10[9],zero,xmm10[11],zero,xmm10[u,10],zero,xmm10[12],zero
; AVX512F-SLOW-NEXT:    vpor %xmm9, %xmm11, %xmm9
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[0,0,1,1]
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm11 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; AVX512F-SLOW-NEXT:    vpternlogq $226, %ymm6, %ymm11, %ymm9
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm9, %zmm0, %zmm6
; AVX512F-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm8 = xmm8[0],xmm10[0],xmm8[1],xmm10[1],xmm8[2],xmm10[2],xmm8[3],xmm10[3],xmm8[4],xmm10[4],xmm8[5],xmm10[5],xmm8[6],xmm10[6],xmm8[7],xmm10[7]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm8 = xmm8[2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[0,0,1,1]
; AVX512F-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm5 = xmm5[0],xmm7[0],xmm5[1],xmm7[1],xmm5[2],xmm7[2],xmm5[3],xmm7[3],xmm5[4],xmm7[4],xmm5[5],xmm7[5],xmm5[6],xmm7[6],xmm5[7],xmm7[7]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} xmm5 = xmm5[0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[0,0,1,1]
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm7 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; AVX512F-SLOW-NEXT:    vpternlogq $226, %ymm8, %ymm7, %ymm5
; AVX512F-SLOW-NEXT:    vshufi64x2 {{.*#+}} zmm5 = zmm5[0,1,2,3],zmm6[4,5,6,7]
; AVX512F-SLOW-NEXT:    vmovdqa (%r8), %xmm6
; AVX512F-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm8 = <u,0,0,0,0,u,1,1,1,1,u,2,2,2,2,u>
; AVX512F-SLOW-NEXT:    vpermd %zmm6, %zmm8, %zmm6
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm5, %zmm6
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = ymm4[u,u,u],zero,ymm4[13,u,u,u],zero,ymm4[14,u,u,u],zero,ymm4[15,u,u,u],zero,ymm4[16,u,u,u],zero,ymm4[17,u,u,u],zero,ymm4[18,u,u]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = ymm3[u,u,u,13],zero,ymm3[u,u,u,14],zero,ymm3[u,u,u,15],zero,ymm3[u,u,u,16],zero,ymm3[u,u,u,17],zero,ymm3[u,u,u,18],zero,ymm3[u,u]
; AVX512F-SLOW-NEXT:    vpor %ymm5, %ymm8, %ymm5
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = ymm2[u],zero,ymm2[13,u,u,u],zero,ymm2[14,u,u,u],zero,ymm2[15,u,u,u],zero,ymm2[16,u,u,u],zero,ymm2[17,u,u,u],zero,ymm2[18,u,u,u],zero
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm9 = ymm1[u,13],zero,ymm1[u,u,u,14],zero,ymm1[u,u,u,15],zero,ymm1[u,u,u,16],zero,ymm1[u,u,u,17],zero,ymm1[u,u,u,18],zero,ymm1[u,u,u,19]
; AVX512F-SLOW-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX512F-SLOW-NEXT:    vpternlogq $226, %ymm5, %ymm11, %ymm8
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19,u,21,u,u,20,u,22,u,24,u,u,23,u,25,u]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm9 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm1[21],zero,zero,ymm1[20],zero,ymm1[22],zero,ymm1[24],zero,zero,ymm1[23],zero,ymm1[25],zero,zero
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm10 = [18374966859431608575,18374966859431608575,18446463693966278400,18446463693966278400]
; AVX512F-SLOW-NEXT:    vpternlogq $248, %ymm10, %ymm5, %ymm9
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm4[19],zero,ymm4[21],zero,zero,ymm4[20],zero,ymm4[22],zero,ymm4[24],zero,zero,ymm4[23],zero
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm11 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,18,19],zero,ymm3[21],zero,ymm3[21,20],zero,ymm3[22],zero,ymm3[24],zero,ymm3[22,23],zero,ymm3[25]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm11 = ymm11[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpor %ymm5, %ymm11, %ymm5
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm9, %ymm5
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm5, %zmm0, %zmm5
; AVX512F-SLOW-NEXT:    vshufi64x2 {{.*#+}} zmm5 = zmm8[0,1,2,3],zmm5[4,5,6,7]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = ymm0[12],zero,zero,zero,zero,ymm0[13],zero,zero,zero,zero,ymm0[14],zero,zero,zero,zero,ymm0[15],zero,zero,zero,zero,ymm0[16],zero,zero,zero,zero,ymm0[17],zero,zero,zero,zero,ymm0[18],zero
; AVX512F-SLOW-NEXT:    vpshufd {{.*#+}} ymm9 = ymm0[0,2,1,1,4,6,5,5]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,3,3,2]
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm11 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX512F-SLOW-NEXT:    vpandn %ymm9, %ymm11, %ymm9
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm9, %zmm8, %zmm8
; AVX512F-SLOW-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm5, %zmm8
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm4 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,25,u,27,u,u,26,u,28,u,30,u,u,29,u,31,u]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm3 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm3[27],zero,zero,ymm3[26],zero,ymm3[28],zero,ymm3[30],zero,zero,ymm3[29],zero,ymm3[31],zero,zero
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpternlogq $248, %ymm10, %ymm4, %ymm3
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm2 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,27,u,u,26,u,28,u,u,u,u,29,u,31,u,u,30]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm1 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm1[26],zero,ymm1[28],zero,zero,ymm1[27],zero,ymm1[29],zero,ymm1[31],zero,zero,ymm1[30],zero
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm2, %ymm1
; AVX512F-SLOW-NEXT:    vpternlogq $184, %ymm3, %ymm7, %ymm1
; AVX512F-SLOW-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[2,2,3,3,6,6,7,7]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm1, %ymm0
; AVX512F-SLOW-NEXT:    vmovdqa %ymm0, 128(%r9)
; AVX512F-SLOW-NEXT:    vmovdqa64 %zmm8, 64(%r9)
; AVX512F-SLOW-NEXT:    vmovdqa64 %zmm6, (%r9)
; AVX512F-SLOW-NEXT:    vzeroupper
; AVX512F-SLOW-NEXT:    retq
;
; AVX512F-FAST-LABEL: store_i8_stride5_vf32:
; AVX512F-FAST:       # %bb.0:
; AVX512F-FAST-NEXT:    vmovdqa (%rdi), %ymm0
; AVX512F-FAST-NEXT:    vmovdqa (%rsi), %ymm1
; AVX512F-FAST-NEXT:    vmovdqa (%rdx), %ymm2
; AVX512F-FAST-NEXT:    vmovdqa (%rcx), %ymm3
; AVX512F-FAST-NEXT:    vmovdqa (%rdi), %xmm4
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm5 = xmm4[8],zero,xmm4[u,7],zero,xmm4[9],zero,xmm4[u],zero,xmm4[u,10],zero,xmm4[12],zero,xmm4[u,11]
; AVX512F-FAST-NEXT:    vmovdqa (%rsi), %xmm6
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm7 = zero,xmm6[8,u],zero,xmm6[7],zero,xmm6[9,u,11,u],zero,xmm6[10],zero,xmm6[12,u],zero
; AVX512F-FAST-NEXT:    vpor %xmm5, %xmm7, %xmm5
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[0,0,1,1]
; AVX512F-FAST-NEXT:    vmovdqa (%rcx), %xmm7
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm8 = zero,xmm7[6],zero,xmm7[8,u],zero,xmm7[7],zero,xmm7[9],zero,xmm7[11,u],zero,xmm7[10],zero,xmm7[12]
; AVX512F-FAST-NEXT:    vmovdqa (%rdx), %xmm9
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm10 = xmm9[6],zero,xmm9[8],zero,xmm9[u,7],zero,xmm9[9],zero,xmm9[11],zero,xmm9[u,10],zero,xmm9[12],zero
; AVX512F-FAST-NEXT:    vpor %xmm8, %xmm10, %xmm8
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[0,0,1,1]
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm10 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; AVX512F-FAST-NEXT:    vpternlogq $226, %ymm5, %ymm10, %ymm8
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm8, %zmm0, %zmm5
; AVX512F-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm7 = xmm7[0],xmm9[0],xmm7[1],xmm9[1],xmm7[2],xmm9[2],xmm7[3],xmm9[3],xmm7[4],xmm9[4],xmm7[5],xmm9[5],xmm7[6],xmm9[6],xmm7[7],xmm9[7]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm7 = xmm7[2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[0,0,1,1]
; AVX512F-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm4 = xmm4[0],xmm6[0],xmm4[1],xmm6[1],xmm4[2],xmm6[2],xmm4[3],xmm6[3],xmm4[4],xmm6[4],xmm4[5],xmm6[5],xmm4[6],xmm6[6],xmm4[7],xmm6[7]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} xmm4 = xmm4[0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,1,1]
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm6 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; AVX512F-FAST-NEXT:    vpternlogq $226, %ymm7, %ymm6, %ymm4
; AVX512F-FAST-NEXT:    vshufi64x2 {{.*#+}} zmm5 = zmm4[0,1,2,3],zmm5[4,5,6,7]
; AVX512F-FAST-NEXT:    vbroadcasti64x4 {{.*#+}} zmm4 = mem[0,1,2,3,0,1,2,3]
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm7 = <u,0,0,0,0,u,1,1,9,9,u,10,10,10,10,u>
; AVX512F-FAST-NEXT:    vpermd %zmm4, %zmm7, %zmm7
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm5, %zmm7
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm5 = ymm3[u,u,u],zero,ymm3[13,u,u,u],zero,ymm3[14,u,u,u],zero,ymm3[15,u,u,u],zero,ymm3[16,u,u,u],zero,ymm3[17,u,u,u],zero,ymm3[18,u,u]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = ymm2[u,u,u,13],zero,ymm2[u,u,u,14],zero,ymm2[u,u,u,15],zero,ymm2[u,u,u,16],zero,ymm2[u,u,u,17],zero,ymm2[u,u,u,18],zero,ymm2[u,u]
; AVX512F-FAST-NEXT:    vpor %ymm5, %ymm8, %ymm5
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = ymm1[u],zero,ymm1[13,u,u,u],zero,ymm1[14,u,u,u],zero,ymm1[15,u,u,u],zero,ymm1[16,u,u,u],zero,ymm1[17,u,u,u],zero,ymm1[18,u,u,u],zero
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm9 = ymm0[u,13],zero,ymm0[u,u,u,14],zero,ymm0[u,u,u,15],zero,ymm0[u,u,u,16],zero,ymm0[u,u,u,17],zero,ymm0[u,u,u,18],zero,ymm0[u,u,u,19]
; AVX512F-FAST-NEXT:    vpor %ymm8, %ymm9, %ymm8
; AVX512F-FAST-NEXT:    vpternlogq $226, %ymm5, %ymm10, %ymm8
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm5 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19,u,21,u,u,20,u,22,u,24,u,u,23,u,25,u]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm9 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm0[21],zero,zero,ymm0[20],zero,ymm0[22],zero,ymm0[24],zero,zero,ymm0[23],zero,ymm0[25],zero,zero
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm10 = [18374966859431608575,18374966859431608575,18446463693966278400,18446463693966278400]
; AVX512F-FAST-NEXT:    vpternlogq $248, %ymm10, %ymm5, %ymm9
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm5 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm3[19],zero,ymm3[21],zero,zero,ymm3[20],zero,ymm3[22],zero,ymm3[24],zero,zero,ymm3[23],zero
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm11 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,18,19],zero,ymm2[21],zero,ymm2[21,20],zero,ymm2[22],zero,ymm2[24],zero,ymm2[22,23],zero,ymm2[25]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm11 = ymm11[2,2,3,3]
; AVX512F-FAST-NEXT:    vpor %ymm5, %ymm11, %ymm5
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm9, %ymm5
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm5, %zmm0, %zmm5
; AVX512F-FAST-NEXT:    vshufi64x2 {{.*#+}} zmm5 = zmm8[0,1,2,3],zmm5[4,5,6,7]
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = <4,u,5,5,5,5,u,6>
; AVX512F-FAST-NEXT:    vpermd %ymm4, %ymm8, %ymm8
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX512F-FAST-NEXT:    vpandn %ymm8, %ymm9, %ymm8
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm9 = ymm4[12],zero,zero,zero,zero,ymm4[13],zero,zero,zero,zero,ymm4[14],zero,zero,zero,zero,ymm4[15],zero,zero,zero,zero,ymm4[16],zero,zero,zero,zero,ymm4[17],zero,zero,zero,zero,ymm4[18],zero
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm8, %zmm9, %zmm8
; AVX512F-FAST-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm5, %zmm8
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm3 = ymm3[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,25,u,27,u,u,26,u,28,u,30,u,u,29,u,31,u]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[2,2,3,3]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm2[27],zero,zero,ymm2[26],zero,ymm2[28],zero,ymm2[30],zero,zero,ymm2[29],zero,ymm2[31],zero,zero
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512F-FAST-NEXT:    vpternlogq $248, %ymm10, %ymm3, %ymm2
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm1 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,27,u,u,26,u,28,u,u,u,u,29,u,31,u,u,30]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm0[26],zero,ymm0[28],zero,zero,ymm0[27],zero,ymm0[29],zero,ymm0[31],zero,zero,ymm0[30],zero
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,2,3,3]
; AVX512F-FAST-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm1, %ymm0
; AVX512F-FAST-NEXT:    vpternlogq $184, %ymm2, %ymm6, %ymm0
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm1 = <6,6,6,u,7,7,7,7>
; AVX512F-FAST-NEXT:    vpermd %ymm4, %ymm1, %ymm1
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm1
; AVX512F-FAST-NEXT:    vmovdqa %ymm1, 128(%r9)
; AVX512F-FAST-NEXT:    vmovdqa64 %zmm8, 64(%r9)
; AVX512F-FAST-NEXT:    vmovdqa64 %zmm7, (%r9)
; AVX512F-FAST-NEXT:    vzeroupper
; AVX512F-FAST-NEXT:    retq
;
; AVX512BW-SLOW-LABEL: store_i8_stride5_vf32:
; AVX512BW-SLOW:       # %bb.0:
; AVX512BW-SLOW-NEXT:    vmovdqa (%rdi), %ymm4
; AVX512BW-SLOW-NEXT:    vmovdqa (%rsi), %ymm5
; AVX512BW-SLOW-NEXT:    vmovdqa (%rdx), %ymm1
; AVX512BW-SLOW-NEXT:    vmovdqa (%rcx), %ymm2
; AVX512BW-SLOW-NEXT:    vmovdqa (%r8), %ymm0
; AVX512BW-SLOW-NEXT:    vmovdqa (%rdi), %xmm3
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm6 = xmm3[8],zero,xmm3[u,7],zero,xmm3[9],zero,xmm3[u],zero,xmm3[u,10],zero,xmm3[12],zero,xmm3[u,11]
; AVX512BW-SLOW-NEXT:    vmovdqa (%rsi), %xmm7
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm8 = zero,xmm7[8,u],zero,xmm7[7],zero,xmm7[9,u,11,u],zero,xmm7[10],zero,xmm7[12,u],zero
; AVX512BW-SLOW-NEXT:    vpor %xmm6, %xmm8, %xmm6
; AVX512BW-SLOW-NEXT:    vmovdqa (%rdx), %xmm8
; AVX512BW-SLOW-NEXT:    vmovdqa (%rcx), %xmm9
; AVX512BW-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm10 = xmm9[0],xmm8[0],xmm9[1],xmm8[1],xmm9[2],xmm8[2],xmm9[3],xmm8[3],xmm9[4],xmm8[4],xmm9[5],xmm8[5],xmm9[6],xmm8[6],xmm9[7],xmm8[7]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm10 = xmm10[2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8]
; AVX512BW-SLOW-NEXT:    vinserti32x4 $2, %xmm6, %zmm10, %zmm6
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} zmm6 = zmm6[0,0,1,1,4,4,5,5]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm9 = zero,xmm9[6],zero,xmm9[8,u],zero,xmm9[7],zero,xmm9[9],zero,xmm9[11,u],zero,xmm9[10],zero,xmm9[12]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm8 = xmm8[6],zero,xmm8[8],zero,xmm8[u,7],zero,xmm8[9],zero,xmm8[11],zero,xmm8[u,10],zero,xmm8[12],zero
; AVX512BW-SLOW-NEXT:    vpor %xmm9, %xmm8, %xmm8
; AVX512BW-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm3 = xmm3[0],xmm7[0],xmm3[1],xmm7[1],xmm3[2],xmm7[2],xmm3[3],xmm7[3],xmm3[4],xmm7[4],xmm3[5],xmm7[5],xmm3[6],xmm7[6],xmm3[7],xmm7[7]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} xmm3 = xmm3[0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13]
; AVX512BW-SLOW-NEXT:    vinserti32x4 $2, %xmm8, %zmm3, %zmm3
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} zmm3 = zmm3[0,0,1,1,4,4,5,5]
; AVX512BW-SLOW-NEXT:    movabsq $3570337559743967628, %rax # imm = 0x318C631818C6318C
; AVX512BW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %zmm6, %zmm3 {%k1}
; AVX512BW-SLOW-NEXT:    vmovdqa (%r8), %xmm6
; AVX512BW-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm7 = [0,0,0,0,0,0,1,1,1,1,2,2,2,2,2,2]
; AVX512BW-SLOW-NEXT:    vpermd %zmm6, %zmm7, %zmm6
; AVX512BW-SLOW-NEXT:    movabsq $595056260442243600, %rax # imm = 0x842108421084210
; AVX512BW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %zmm6, %zmm3 {%k1}
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm6 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,18,19],zero,ymm1[21],zero,ymm1[21,20],zero,ymm1[22],zero,ymm1[24],zero,ymm1[22,23],zero,ymm1[25]
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[2,2,3,3]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm7 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm2[19],zero,ymm2[21],zero,zero,ymm2[20],zero,ymm2[22],zero,ymm2[24],zero,zero,ymm2[23],zero
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512BW-SLOW-NEXT:    vpor %ymm6, %ymm7, %ymm6
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm7 = zero,zero,ymm1[12,13],zero,zero,zero,zero,ymm1[14],zero,zero,zero,ymm1[14,15],zero,zero,zero,zero,ymm1[16],zero,zero,zero,ymm1[16,17],zero,zero,zero,zero,ymm1[18],zero,zero,zero
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = zero,zero,zero,zero,ymm2[13],zero,zero,zero,zero,ymm2[14],zero,zero,zero,zero,ymm2[15],zero,zero,zero,zero,ymm2[16],zero,zero,zero,zero,ymm2[17],zero,zero,zero,zero,ymm2[18],zero,zero
; AVX512BW-SLOW-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX512BW-SLOW-NEXT:    vinserti64x4 $1, %ymm6, %zmm7, %zmm6
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm7 = ymm5[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19],zero,ymm5[21],zero,zero,ymm5[20],zero,ymm5[22],zero,ymm5[24],zero,zero,ymm5[23],zero,ymm5[25],zero
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm4[21],zero,zero,ymm4[20],zero,ymm4[22],zero,ymm4[24],zero,zero,ymm4[23],zero,ymm4[25],zero,zero
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX512BW-SLOW-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX512BW-SLOW-NEXT:    vmovdqa {{.*#+}} ymm8 = <3,3,3,u,4,4,4,4>
; AVX512BW-SLOW-NEXT:    vpermd %ymm4, %ymm8, %ymm8
; AVX512BW-SLOW-NEXT:    movl $138547332, %eax # imm = 0x8421084
; AVX512BW-SLOW-NEXT:    kmovd %eax, %k1
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm8 {%k1} = ymm5[u,u,13,u,u,u,u,14,u,u,u,u,15,u,u,u,u,16,u,u,u,u,17,u,u,u,u,18,u,u,u,u]
; AVX512BW-SLOW-NEXT:    vinserti64x4 $1, %ymm7, %zmm8, %zmm7
; AVX512BW-SLOW-NEXT:    movabsq $-8330787646191410408, %rax # imm = 0x8C6318C6318C6318
; AVX512BW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %zmm6, %zmm7 {%k1}
; AVX512BW-SLOW-NEXT:    vmovdqa {{.*#+}} ymm6 = <3,3,3,3,u,4,4,4>
; AVX512BW-SLOW-NEXT:    vpermd %ymm0, %ymm6, %ymm6
; AVX512BW-SLOW-NEXT:    vpshufd {{.*#+}} ymm8 = ymm0[0,2,1,1,4,6,5,5]
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,3,3,2]
; AVX512BW-SLOW-NEXT:    vinserti64x4 $1, %ymm8, %zmm6, %zmm6
; AVX512BW-SLOW-NEXT:    movabsq $1190112520884487201, %rax # imm = 0x1084210842108421
; AVX512BW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %zmm6, %zmm7 {%k1}
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = ymm5[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,27,u,u,26,u,28,u,u,u,u,29,u,31,u,u,30]
; AVX512BW-SLOW-NEXT:    vpshufhw {{.*#+}} ymm4 = ymm4[0,1,2,3,6,5,6,7,8,9,10,11,14,13,14,15]
; AVX512BW-SLOW-NEXT:    vpshufd {{.*#+}} ymm4 = ymm4[2,2,3,3,6,6,7,7]
; AVX512BW-SLOW-NEXT:    movl $1251232404, %eax # imm = 0x4A944A94
; AVX512BW-SLOW-NEXT:    kmovd %eax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %ymm4, %ymm5 {%k1}
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm5[2,2,3,3]
; AVX512BW-SLOW-NEXT:    vpshufb {{.*#+}} ymm2 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,25,u,27,u,u,26,u,28,u,30,u,u,29,u,31,u]
; AVX512BW-SLOW-NEXT:    vpshufhw {{.*#+}} ymm1 = ymm1[0,1,2,3,5,6,7,6,8,9,10,11,13,14,15,14]
; AVX512BW-SLOW-NEXT:    vpshufd {{.*#+}} ymm1 = ymm1[2,2,3,3,6,6,7,7]
; AVX512BW-SLOW-NEXT:    movl $693250386, %eax # imm = 0x29522952
; AVX512BW-SLOW-NEXT:    kmovd %eax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %ymm1, %ymm2 {%k1}
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm2[2,2,3,3]
; AVX512BW-SLOW-NEXT:    movl $415641996, %eax # imm = 0x18C6318C
; AVX512BW-SLOW-NEXT:    kmovd %eax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %ymm4, %ymm1 {%k1}
; AVX512BW-SLOW-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[2,2,3,3,6,6,7,7]
; AVX512BW-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,2,3,3]
; AVX512BW-SLOW-NEXT:    movl $-2078209982, %eax # imm = 0x84210842
; AVX512BW-SLOW-NEXT:    kmovd %eax, %k1
; AVX512BW-SLOW-NEXT:    vmovdqu8 %ymm0, %ymm1 {%k1}
; AVX512BW-SLOW-NEXT:    vmovdqa %ymm1, 128(%r9)
; AVX512BW-SLOW-NEXT:    vmovdqa64 %zmm7, 64(%r9)
; AVX512BW-SLOW-NEXT:    vmovdqa64 %zmm3, (%r9)
; AVX512BW-SLOW-NEXT:    vzeroupper
; AVX512BW-SLOW-NEXT:    retq
;
; AVX512BW-FAST-LABEL: store_i8_stride5_vf32:
; AVX512BW-FAST:       # %bb.0:
; AVX512BW-FAST-NEXT:    vmovdqa (%rdi), %ymm1
; AVX512BW-FAST-NEXT:    vmovdqa (%rsi), %ymm4
; AVX512BW-FAST-NEXT:    vmovdqa (%rdx), %ymm0
; AVX512BW-FAST-NEXT:    vmovdqa (%rcx), %ymm2
; AVX512BW-FAST-NEXT:    vmovdqa (%rdi), %xmm3
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm5 = xmm3[8],zero,xmm3[u,7],zero,xmm3[9],zero,xmm3[u],zero,xmm3[u,10],zero,xmm3[12],zero,xmm3[u,11]
; AVX512BW-FAST-NEXT:    vmovdqa (%rsi), %xmm6
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm7 = zero,xmm6[8,u],zero,xmm6[7],zero,xmm6[9,u,11,u],zero,xmm6[10],zero,xmm6[12,u],zero
; AVX512BW-FAST-NEXT:    vpor %xmm5, %xmm7, %xmm5
; AVX512BW-FAST-NEXT:    vmovdqa (%rdx), %xmm7
; AVX512BW-FAST-NEXT:    vmovdqa (%rcx), %xmm8
; AVX512BW-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm9 = xmm8[0],xmm7[0],xmm8[1],xmm7[1],xmm8[2],xmm7[2],xmm8[3],xmm7[3],xmm8[4],xmm7[4],xmm8[5],xmm7[5],xmm8[6],xmm7[6],xmm8[7],xmm7[7]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm9 = xmm9[2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8]
; AVX512BW-FAST-NEXT:    vinserti32x4 $2, %xmm5, %zmm9, %zmm5
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} zmm5 = zmm5[0,0,1,1,4,4,5,5]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm8 = zero,xmm8[6],zero,xmm8[8,u],zero,xmm8[7],zero,xmm8[9],zero,xmm8[11,u],zero,xmm8[10],zero,xmm8[12]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm7 = xmm7[6],zero,xmm7[8],zero,xmm7[u,7],zero,xmm7[9],zero,xmm7[11],zero,xmm7[u,10],zero,xmm7[12],zero
; AVX512BW-FAST-NEXT:    vpor %xmm7, %xmm8, %xmm7
; AVX512BW-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm3 = xmm3[0],xmm6[0],xmm3[1],xmm6[1],xmm3[2],xmm6[2],xmm3[3],xmm6[3],xmm3[4],xmm6[4],xmm3[5],xmm6[5],xmm3[6],xmm6[6],xmm3[7],xmm6[7]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} xmm3 = xmm3[0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13]
; AVX512BW-FAST-NEXT:    vinserti32x4 $2, %xmm7, %zmm3, %zmm3
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} zmm3 = zmm3[0,0,1,1,4,4,5,5]
; AVX512BW-FAST-NEXT:    movabsq $3570337559743967628, %rax # imm = 0x318C631818C6318C
; AVX512BW-FAST-NEXT:    kmovq %rax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm5, %zmm3 {%k1}
; AVX512BW-FAST-NEXT:    vbroadcasti64x4 {{.*#+}} zmm5 = mem[0,1,2,3,0,1,2,3]
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm6 = [0,0,0,0,0,0,1,1,9,9,10,10,10,10,10,10]
; AVX512BW-FAST-NEXT:    vpermd %zmm5, %zmm6, %zmm6
; AVX512BW-FAST-NEXT:    movabsq $595056260442243600, %rax # imm = 0x842108421084210
; AVX512BW-FAST-NEXT:    kmovq %rax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm6, %zmm3 {%k1}
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm6 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,18,19],zero,ymm0[21],zero,ymm0[21,20],zero,ymm0[22],zero,ymm0[24],zero,ymm0[22,23],zero,ymm0[25]
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm7 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm2[19],zero,ymm2[21],zero,zero,ymm2[20],zero,ymm2[22],zero,ymm2[24],zero,zero,ymm2[23],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpor %ymm6, %ymm7, %ymm6
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm7 = zero,zero,ymm0[12,13],zero,zero,zero,zero,ymm0[14],zero,zero,zero,ymm0[14,15],zero,zero,zero,zero,ymm0[16],zero,zero,zero,ymm0[16,17],zero,zero,zero,zero,ymm0[18],zero,zero,zero
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = zero,zero,zero,zero,ymm2[13],zero,zero,zero,zero,ymm2[14],zero,zero,zero,zero,ymm2[15],zero,zero,zero,zero,ymm2[16],zero,zero,zero,zero,ymm2[17],zero,zero,zero,zero,ymm2[18],zero,zero
; AVX512BW-FAST-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm6, %zmm7, %zmm6
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm7 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19],zero,ymm4[21],zero,zero,ymm4[20],zero,ymm4[22],zero,ymm4[24],zero,zero,ymm4[23],zero,ymm4[25],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm8 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm1[21],zero,zero,ymm1[20],zero,ymm1[22],zero,ymm1[24],zero,zero,ymm1[23],zero,ymm1[25],zero,zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpor %ymm7, %ymm8, %ymm7
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = <3,3,3,u,4,4,4,4>
; AVX512BW-FAST-NEXT:    vpermd %ymm1, %ymm8, %ymm8
; AVX512BW-FAST-NEXT:    movl $138547332, %eax # imm = 0x8421084
; AVX512BW-FAST-NEXT:    kmovd %eax, %k1
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm8 {%k1} = ymm4[u,u,13,u,u,u,u,14,u,u,u,u,15,u,u,u,u,16,u,u,u,u,17,u,u,u,u,18,u,u,u,u]
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm7, %zmm8, %zmm7
; AVX512BW-FAST-NEXT:    movabsq $-8330787646191410408, %rax # imm = 0x8C6318C6318C6318
; AVX512BW-FAST-NEXT:    kmovq %rax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm6, %zmm7 {%k1}
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm6 = <3,3,3,3,u,4,4,4,12,14,13,13,13,13,12,14>
; AVX512BW-FAST-NEXT:    vpermd %zmm5, %zmm6, %zmm6
; AVX512BW-FAST-NEXT:    movabsq $1190112520884487201, %rax # imm = 0x1084210842108421
; AVX512BW-FAST-NEXT:    kmovq %rax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm6, %zmm7 {%k1}
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm4 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,27],zero,zero,ymm4[26],zero,ymm4[28],zero,zero,zero,zero,ymm4[29],zero,ymm4[31],zero,zero,ymm4[30]
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm1 = ymm1[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm1[26],zero,ymm1[28],zero,zero,ymm1[27],zero,ymm1[29],zero,ymm1[31],zero,zero,ymm1[30],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpor %ymm4, %ymm1, %ymm1
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,25],zero,ymm2[27],zero,zero,ymm2[26],zero,ymm2[28],zero,ymm2[30],zero,zero,ymm2[29],zero,ymm2[31],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm0 = ymm0[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm0[27],zero,zero,ymm0[26],zero,ymm0[28],zero,ymm0[30],zero,zero,ymm0[29],zero,ymm0[31],zero,zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpor %ymm2, %ymm0, %ymm0
; AVX512BW-FAST-NEXT:    movl $415641996, %eax # imm = 0x18C6318C
; AVX512BW-FAST-NEXT:    kmovd %eax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %ymm1, %ymm0 {%k1}
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} ymm1 = [6,6,6,6,7,7,7,7]
; AVX512BW-FAST-NEXT:    vpermd %ymm5, %ymm1, %ymm1
; AVX512BW-FAST-NEXT:    movl $-2078209982, %eax # imm = 0x84210842
; AVX512BW-FAST-NEXT:    kmovd %eax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %ymm1, %ymm0 {%k1}
; AVX512BW-FAST-NEXT:    vmovdqa %ymm0, 128(%r9)
; AVX512BW-FAST-NEXT:    vmovdqa64 %zmm7, 64(%r9)
; AVX512BW-FAST-NEXT:    vmovdqa64 %zmm3, (%r9)
; AVX512BW-FAST-NEXT:    vzeroupper
; AVX512BW-FAST-NEXT:    retq
  %in.vec0 = load <32 x i8>, ptr %in.vecptr0, align 64
  %in.vec1 = load <32 x i8>, ptr %in.vecptr1, align 64
  %in.vec2 = load <32 x i8>, ptr %in.vecptr2, align 64
  %in.vec3 = load <32 x i8>, ptr %in.vecptr3, align 64
  %in.vec4 = load <32 x i8>, ptr %in.vecptr4, align 64
  %1 = shufflevector <32 x i8> %in.vec0, <32 x i8> %in.vec1, <64 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
  %2 = shufflevector <32 x i8> %in.vec2, <32 x i8> %in.vec3, <64 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63>
  %3 = shufflevector <64 x i8> %1, <64 x i8> %2, <128 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
  %4 = shufflevector <32 x i8> %in.vec4, <32 x i8> poison, <128 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 31, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5 = shufflevector <128 x i8> %3, <128 x i8> %4, <160 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127, i32 128, i32 129, i32 130, i32 131, i32 132, i32 133, i32 134, i32 135, i32 136, i32 137, i32 138, i32 139, i32 140, i32 141, i32 142, i32 143, i32 144, i32 145, i32 146, i32 147, i32 148, i32 149, i32 150, i32 151, i32 152, i32 153, i32 154, i32 155, i32 156, i32 157, i32 158, i32 159>
  %interleaved.vec = shufflevector <160 x i8> %5, <160 x i8> poison, <160 x i32> <i32 0, i32 32, i32 64, i32 96, i32 128, i32 1, i32 33, i32 65, i32 97, i32 129, i32 2, i32 34, i32 66, i32 98, i32 130, i32 3, i32 35, i32 67, i32 99, i32 131, i32 4, i32 36, i32 68, i32 100, i32 132, i32 5, i32 37, i32 69, i32 101, i32 133, i32 6, i32 38, i32 70, i32 102, i32 134, i32 7, i32 39, i32 71, i32 103, i32 135, i32 8, i32 40, i32 72, i32 104, i32 136, i32 9, i32 41, i32 73, i32 105, i32 137, i32 10, i32 42, i32 74, i32 106, i32 138, i32 11, i32 43, i32 75, i32 107, i32 139, i32 12, i32 44, i32 76, i32 108, i32 140, i32 13, i32 45, i32 77, i32 109, i32 141, i32 14, i32 46, i32 78, i32 110, i32 142, i32 15, i32 47, i32 79, i32 111, i32 143, i32 16, i32 48, i32 80, i32 112, i32 144, i32 17, i32 49, i32 81, i32 113, i32 145, i32 18, i32 50, i32 82, i32 114, i32 146, i32 19, i32 51, i32 83, i32 115, i32 147, i32 20, i32 52, i32 84, i32 116, i32 148, i32 21, i32 53, i32 85, i32 117, i32 149, i32 22, i32 54, i32 86, i32 118, i32 150, i32 23, i32 55, i32 87, i32 119, i32 151, i32 24, i32 56, i32 88, i32 120, i32 152, i32 25, i32 57, i32 89, i32 121, i32 153, i32 26, i32 58, i32 90, i32 122, i32 154, i32 27, i32 59, i32 91, i32 123, i32 155, i32 28, i32 60, i32 92, i32 124, i32 156, i32 29, i32 61, i32 93, i32 125, i32 157, i32 30, i32 62, i32 94, i32 126, i32 158, i32 31, i32 63, i32 95, i32 127, i32 159>
  store <160 x i8> %interleaved.vec, ptr %out.vec, align 64
  ret void
}

define void @store_i8_stride5_vf64(ptr %in.vecptr0, ptr %in.vecptr1, ptr %in.vecptr2, ptr %in.vecptr3, ptr %in.vecptr4, ptr %out.vec) nounwind {
; SSE-LABEL: store_i8_stride5_vf64:
; SSE:       # %bb.0:
; SSE-NEXT:    subq $504, %rsp # imm = 0x1F8
; SSE-NEXT:    movdqa (%rdi), %xmm7
; SSE-NEXT:    movdqa (%rsi), %xmm9
; SSE-NEXT:    movdqa 16(%rsi), %xmm14
; SSE-NEXT:    movdqa (%rdx), %xmm0
; SSE-NEXT:    movdqa %xmm0, (%rsp) # 16-byte Spill
; SSE-NEXT:    movdqa 16(%rdx), %xmm11
; SSE-NEXT:    movdqa %xmm11, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa (%rcx), %xmm10
; SSE-NEXT:    movdqa 16(%rcx), %xmm6
; SSE-NEXT:    movdqa %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa (%r8), %xmm13
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,3,3,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa {{.*#+}} xmm12 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; SSE-NEXT:    pand %xmm12, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm10[2,1,2,3]
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[3,3,0,3]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa %xmm12, %xmm4
; SSE-NEXT:    pandn %xmm1, %xmm4
; SSE-NEXT:    por %xmm0, %xmm4
; SSE-NEXT:    movdqa {{.*#+}} xmm8 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255]
; SSE-NEXT:    pand %xmm8, %xmm4
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm7[1,1,2,2]
; SSE-NEXT:    movdqa {{.*#+}} xmm2 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; SSE-NEXT:    pand %xmm2, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm9[0,1,2,1]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,7,6,7]
; SSE-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[0,0,1,0]
; SSE-NEXT:    movdqa %xmm2, %xmm5
; SSE-NEXT:    pandn %xmm1, %xmm5
; SSE-NEXT:    por %xmm0, %xmm5
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    por %xmm4, %xmm0
; SSE-NEXT:    movdqa {{.*#+}} xmm15 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; SSE-NEXT:    pand %xmm15, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm13[1,1,2,2]
; SSE-NEXT:    movdqa %xmm15, %xmm3
; SSE-NEXT:    pandn %xmm1, %xmm3
; SSE-NEXT:    por %xmm0, %xmm3
; SSE-NEXT:    movdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm11[3,3,3,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
; SSE-NEXT:    pand %xmm12, %xmm0
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm6[2,1,2,3]
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[3,3,0,3]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa %xmm12, %xmm5
; SSE-NEXT:    pandn %xmm1, %xmm5
; SSE-NEXT:    por %xmm0, %xmm5
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm14[0,1,2,1]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,7,6,7]
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,1,0]
; SSE-NEXT:    movdqa %xmm2, %xmm1
; SSE-NEXT:    pandn %xmm0, %xmm1
; SSE-NEXT:    movdqa 16(%rdi), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,2,2]
; SSE-NEXT:    pand %xmm2, %xmm0
; SSE-NEXT:    por %xmm0, %xmm1
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    pandn %xmm1, %xmm0
; SSE-NEXT:    pand %xmm8, %xmm5
; SSE-NEXT:    por %xmm5, %xmm0
; SSE-NEXT:    movdqa 16(%r8), %xmm1
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,2,2]
; SSE-NEXT:    movdqa %xmm15, %xmm3
; SSE-NEXT:    pandn %xmm1, %xmm3
; SSE-NEXT:    pand %xmm15, %xmm0
; SSE-NEXT:    por %xmm0, %xmm3
; SSE-NEXT:    movdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa 32(%rcx), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,1,2,3]
; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[3,3,0,3]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa %xmm12, %xmm1
; SSE-NEXT:    pandn %xmm0, %xmm1
; SSE-NEXT:    movdqa 32(%rdx), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,3,3,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
; SSE-NEXT:    pand %xmm12, %xmm0
; SSE-NEXT:    por %xmm0, %xmm1
; SSE-NEXT:    movdqa 32(%rsi), %xmm11
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm11[0,1,2,1]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,7,6,7]
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,1,0]
; SSE-NEXT:    movdqa %xmm2, %xmm5
; SSE-NEXT:    pandn %xmm0, %xmm5
; SSE-NEXT:    movdqa 32(%rdi), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,2,2]
; SSE-NEXT:    pand %xmm2, %xmm0
; SSE-NEXT:    por %xmm0, %xmm5
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    pand %xmm8, %xmm1
; SSE-NEXT:    por %xmm1, %xmm0
; SSE-NEXT:    movdqa 32(%r8), %xmm1
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,2,2]
; SSE-NEXT:    movdqa %xmm15, %xmm3
; SSE-NEXT:    pandn %xmm1, %xmm3
; SSE-NEXT:    pand %xmm15, %xmm0
; SSE-NEXT:    por %xmm0, %xmm3
; SSE-NEXT:    movdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa 48(%rcx), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,1,2,3]
; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[3,3,0,3]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
; SSE-NEXT:    movdqa %xmm12, %xmm1
; SSE-NEXT:    pandn %xmm0, %xmm1
; SSE-NEXT:    movdqa 48(%rdx), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,3,3,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,4,4,4]
; SSE-NEXT:    pand %xmm12, %xmm0
; SSE-NEXT:    por %xmm0, %xmm1
; SSE-NEXT:    movdqa 48(%rsi), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,1,2,1]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,4,7,6,7]
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,1,0]
; SSE-NEXT:    movdqa %xmm2, %xmm5
; SSE-NEXT:    pandn %xmm0, %xmm5
; SSE-NEXT:    movdqa 48(%rdi), %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[1,1,2,2]
; SSE-NEXT:    pand %xmm2, %xmm0
; SSE-NEXT:    por %xmm0, %xmm5
; SSE-NEXT:    movdqa %xmm8, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    pand %xmm8, %xmm1
; SSE-NEXT:    por %xmm1, %xmm0
; SSE-NEXT:    movdqa 48(%r8), %xmm1
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[1,1,2,2]
; SSE-NEXT:    movdqa %xmm15, %xmm5
; SSE-NEXT:    pandn %xmm1, %xmm5
; SSE-NEXT:    pand %xmm15, %xmm0
; SSE-NEXT:    por %xmm0, %xmm5
; SSE-NEXT:    movdqa %xmm5, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa %xmm10, %xmm0
; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm10[0],xmm0[1],xmm10[1],xmm0[2],xmm10[2],xmm0[3],xmm10[3],xmm0[4],xmm10[4],xmm0[5],xmm10[5],xmm0[6],xmm10[6],xmm0[7],xmm10[7]
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm0[0,3,2,0,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,5,5,5]
; SSE-NEXT:    movdqa %xmm15, %xmm1
; SSE-NEXT:    pandn %xmm0, %xmm1
; SSE-NEXT:    movdqa (%rsp), %xmm3 # 16-byte Reload
; SSE-NEXT:    pshuflw {{.*#+}} xmm0 = xmm3[1,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm0 = xmm0[0,0,0,0]
; SSE-NEXT:    pand %xmm15, %xmm0
; SSE-NEXT:    por %xmm0, %xmm1
; SSE-NEXT:    movdqa {{.*#+}} xmm6 = [255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; SSE-NEXT:    movdqa %xmm6, %xmm0
; SSE-NEXT:    pandn %xmm1, %xmm0
; SSE-NEXT:    movdqa %xmm7, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm7[1,1,1,1]
; SSE-NEXT:    movdqa %xmm2, %xmm5
; SSE-NEXT:    pandn %xmm1, %xmm5
; SSE-NEXT:    movdqa %xmm9, %xmm1
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm9[0],xmm1[1],xmm9[1],xmm1[2],xmm9[2],xmm1[3],xmm9[3],xmm1[4],xmm9[4],xmm1[5],xmm9[5],xmm1[6],xmm9[6],xmm1[7],xmm9[7]
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,4,5,6,6]
; SSE-NEXT:    pand %xmm2, %xmm1
; SSE-NEXT:    por %xmm5, %xmm1
; SSE-NEXT:    pand %xmm6, %xmm1
; SSE-NEXT:    por %xmm0, %xmm1
; SSE-NEXT:    movdqa {{.*#+}} xmm4 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; SSE-NEXT:    movdqa %xmm13, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm13[0,0,1,1]
; SSE-NEXT:    movdqa %xmm4, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    pand %xmm4, %xmm1
; SSE-NEXT:    por %xmm1, %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    punpckhbw {{.*#+}} xmm9 = xmm9[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    movdqa %xmm9, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm9[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm1 = xmm1[2,2,2,2,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,7,5,4,7]
; SSE-NEXT:    movdqa %xmm4, %xmm5
; SSE-NEXT:    pandn %xmm1, %xmm5
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm7[2,2,3,3]
; SSE-NEXT:    pand %xmm4, %xmm1
; SSE-NEXT:    por %xmm1, %xmm5
; SSE-NEXT:    movdqa {{.*#+}} xmm8 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; SSE-NEXT:    movdqa %xmm8, %xmm7
; SSE-NEXT:    pandn %xmm5, %xmm7
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm3[0,1,2,3,5,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[2,2,2,2]
; SSE-NEXT:    movdqa %xmm2, %xmm9
; SSE-NEXT:    pandn %xmm5, %xmm9
; SSE-NEXT:    punpckhbw {{.*#+}} xmm10 = xmm10[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    movdqa %xmm10, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm10[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[1,1,2,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,4,7,6,4]
; SSE-NEXT:    pand %xmm2, %xmm5
; SSE-NEXT:    por %xmm9, %xmm5
; SSE-NEXT:    pand %xmm8, %xmm5
; SSE-NEXT:    por %xmm7, %xmm5
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm13[2,2,2,2]
; SSE-NEXT:    movdqa %xmm12, %xmm0
; SSE-NEXT:    pandn %xmm7, %xmm0
; SSE-NEXT:    pand %xmm12, %xmm5
; SSE-NEXT:    por %xmm5, %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm13 # 16-byte Reload
; SSE-NEXT:    movdqa %xmm13, %xmm0
; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm13[0],xmm0[1],xmm13[1],xmm0[2],xmm13[2],xmm0[3],xmm13[3],xmm0[4],xmm13[4],xmm0[5],xmm13[5],xmm0[6],xmm13[6],xmm0[7],xmm13[7]
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[0,3,2,0,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,5,5,5,5]
; SSE-NEXT:    movdqa %xmm15, %xmm7
; SSE-NEXT:    pandn %xmm5, %xmm7
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm10 # 16-byte Reload
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm10[1,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,0,0,0]
; SSE-NEXT:    pand %xmm15, %xmm5
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    movdqa %xmm6, %xmm5
; SSE-NEXT:    pandn %xmm7, %xmm5
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm0[1,1,1,1]
; SSE-NEXT:    movdqa %xmm2, %xmm9
; SSE-NEXT:    pandn %xmm7, %xmm9
; SSE-NEXT:    movdqa %xmm14, %xmm1
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm14[0],xmm1[1],xmm14[1],xmm1[2],xmm14[2],xmm1[3],xmm14[3],xmm1[4],xmm14[4],xmm1[5],xmm14[5],xmm1[6],xmm14[6],xmm1[7],xmm14[7]
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm1[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm7 = xmm7[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm7 = xmm7[0,1,2,3,4,5,6,6]
; SSE-NEXT:    pand %xmm2, %xmm7
; SSE-NEXT:    por %xmm9, %xmm7
; SSE-NEXT:    pand %xmm6, %xmm7
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[0,0,1,1]
; SSE-NEXT:    movdqa %xmm4, %xmm3
; SSE-NEXT:    pandn %xmm5, %xmm3
; SSE-NEXT:    pand %xmm4, %xmm7
; SSE-NEXT:    por %xmm7, %xmm3
; SSE-NEXT:    movdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    punpckhbw {{.*#+}} xmm14 = xmm14[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    movdqa %xmm14, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm14[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[2,2,2,2,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,7,5,4,7]
; SSE-NEXT:    movdqa %xmm4, %xmm7
; SSE-NEXT:    pandn %xmm5, %xmm7
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[2,2,3,3]
; SSE-NEXT:    pand %xmm4, %xmm5
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    movdqa %xmm8, %xmm5
; SSE-NEXT:    pandn %xmm7, %xmm5
; SSE-NEXT:    pshufhw {{.*#+}} xmm7 = xmm10[0,1,2,3,5,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm7[2,2,2,2]
; SSE-NEXT:    movdqa %xmm2, %xmm9
; SSE-NEXT:    pandn %xmm7, %xmm9
; SSE-NEXT:    movdqa %xmm13, %xmm0
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8],xmm13[8],xmm0[9],xmm13[9],xmm0[10],xmm13[10],xmm0[11],xmm13[11],xmm0[12],xmm13[12],xmm0[13],xmm13[13],xmm0[14],xmm13[14],xmm0[15],xmm13[15]
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm0[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm7 = xmm7[1,1,2,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm7 = xmm7[0,1,2,3,4,7,6,4]
; SSE-NEXT:    pand %xmm2, %xmm7
; SSE-NEXT:    por %xmm9, %xmm7
; SSE-NEXT:    pand %xmm8, %xmm7
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[2,2,2,2]
; SSE-NEXT:    movdqa %xmm12, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    pand %xmm12, %xmm7
; SSE-NEXT:    por %xmm7, %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Reload
; SSE-NEXT:    movdqa %xmm3, %xmm0
; SSE-NEXT:    punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3],xmm0[4],xmm3[4],xmm0[5],xmm3[5],xmm0[6],xmm3[6],xmm0[7],xmm3[7]
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[0,3,2,0,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,5,5,5,5]
; SSE-NEXT:    movdqa %xmm15, %xmm7
; SSE-NEXT:    pandn %xmm5, %xmm7
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm10 # 16-byte Reload
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm10[1,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,0,0,0]
; SSE-NEXT:    pand %xmm15, %xmm5
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    movdqa %xmm6, %xmm5
; SSE-NEXT:    pandn %xmm7, %xmm5
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm13 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm13[1,1,1,1]
; SSE-NEXT:    movdqa %xmm2, %xmm9
; SSE-NEXT:    pandn %xmm7, %xmm9
; SSE-NEXT:    movdqa %xmm11, %xmm14
; SSE-NEXT:    punpcklbw {{.*#+}} xmm14 = xmm14[0],xmm11[0],xmm14[1],xmm11[1],xmm14[2],xmm11[2],xmm14[3],xmm11[3],xmm14[4],xmm11[4],xmm14[5],xmm11[5],xmm14[6],xmm11[6],xmm14[7],xmm11[7]
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm14[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm7 = xmm7[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm7 = xmm7[0,1,2,3,4,5,6,6]
; SSE-NEXT:    pand %xmm2, %xmm7
; SSE-NEXT:    por %xmm9, %xmm7
; SSE-NEXT:    pand %xmm6, %xmm7
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[0,0,1,1]
; SSE-NEXT:    movdqa %xmm4, %xmm1
; SSE-NEXT:    pandn %xmm5, %xmm1
; SSE-NEXT:    pand %xmm4, %xmm7
; SSE-NEXT:    por %xmm7, %xmm1
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    punpckhbw {{.*#+}} xmm11 = xmm11[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    movdqa %xmm11, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm11[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[2,2,2,2,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,7,5,4,7]
; SSE-NEXT:    movdqa %xmm4, %xmm7
; SSE-NEXT:    pandn %xmm5, %xmm7
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm13[2,2,3,3]
; SSE-NEXT:    pand %xmm4, %xmm5
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    movdqa %xmm8, %xmm5
; SSE-NEXT:    pandn %xmm7, %xmm5
; SSE-NEXT:    pshufhw {{.*#+}} xmm7 = xmm10[0,1,2,3,5,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm7[2,2,2,2]
; SSE-NEXT:    movdqa %xmm2, %xmm9
; SSE-NEXT:    pandn %xmm7, %xmm9
; SSE-NEXT:    movdqa %xmm3, %xmm1
; SSE-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm3[8],xmm1[9],xmm3[9],xmm1[10],xmm3[10],xmm1[11],xmm3[11],xmm1[12],xmm3[12],xmm1[13],xmm3[13],xmm1[14],xmm3[14],xmm1[15],xmm3[15]
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm7 = xmm1[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm7 = xmm7[1,1,2,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm7 = xmm7[0,1,2,3,4,7,6,4]
; SSE-NEXT:    pand %xmm2, %xmm7
; SSE-NEXT:    por %xmm9, %xmm7
; SSE-NEXT:    pand %xmm8, %xmm7
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[2,2,2,2]
; SSE-NEXT:    movdqa %xmm12, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    pand %xmm12, %xmm7
; SSE-NEXT:    por %xmm7, %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movdqa %xmm0, %xmm1
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[0,3,2,0,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,5,5,5,5]
; SSE-NEXT:    movdqa %xmm15, %xmm7
; SSE-NEXT:    pandn %xmm5, %xmm7
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm10 # 16-byte Reload
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm10[1,2,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,0,0,0]
; SSE-NEXT:    pand %xmm15, %xmm5
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm11 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm11[1,1,1,1]
; SSE-NEXT:    movdqa %xmm2, %xmm9
; SSE-NEXT:    pandn %xmm5, %xmm9
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Reload
; SSE-NEXT:    movdqa %xmm3, %xmm1
; SSE-NEXT:    punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[2,1,2,3]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[3,1,0,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,4,5,6,6]
; SSE-NEXT:    pand %xmm2, %xmm5
; SSE-NEXT:    por %xmm9, %xmm5
; SSE-NEXT:    pand %xmm6, %xmm5
; SSE-NEXT:    pandn %xmm7, %xmm6
; SSE-NEXT:    por %xmm5, %xmm6
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm9 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm9[0,0,1,1]
; SSE-NEXT:    movdqa %xmm4, %xmm1
; SSE-NEXT:    pandn %xmm5, %xmm1
; SSE-NEXT:    pand %xmm4, %xmm6
; SSE-NEXT:    por %xmm6, %xmm1
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    movdqa %xmm3, %xmm1
; SSE-NEXT:    punpckhbw {{.*#+}} xmm1 = xmm1[8],xmm3[8],xmm1[9],xmm3[9],xmm1[10],xmm3[10],xmm1[11],xmm3[11],xmm1[12],xmm3[12],xmm1[13],xmm3[13],xmm1[14],xmm3[14],xmm1[15],xmm3[15]
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm5[2,2,2,2,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm5[0,1,2,3,7,5,4,7]
; SSE-NEXT:    movdqa %xmm4, %xmm6
; SSE-NEXT:    pandn %xmm5, %xmm6
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm11[2,2,3,3]
; SSE-NEXT:    pand %xmm4, %xmm5
; SSE-NEXT:    por %xmm5, %xmm6
; SSE-NEXT:    movdqa %xmm8, %xmm5
; SSE-NEXT:    pandn %xmm6, %xmm5
; SSE-NEXT:    pshufhw {{.*#+}} xmm6 = xmm10[0,1,2,3,5,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[2,2,2,2]
; SSE-NEXT:    movdqa %xmm2, %xmm7
; SSE-NEXT:    pandn %xmm6, %xmm7
; SSE-NEXT:    punpckhbw {{.*#+}} xmm0 = xmm0[8,8,9,9,10,10,11,11,12,12,13,13,14,14,15,15]
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm0[0,1,2,1]
; SSE-NEXT:    pshuflw {{.*#+}} xmm6 = xmm6[1,1,2,3,4,5,6,7]
; SSE-NEXT:    pshufhw {{.*#+}} xmm6 = xmm6[0,1,2,3,4,7,6,4]
; SSE-NEXT:    pand %xmm2, %xmm6
; SSE-NEXT:    por %xmm7, %xmm6
; SSE-NEXT:    pand %xmm8, %xmm6
; SSE-NEXT:    por %xmm5, %xmm6
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm9[2,2,2,2]
; SSE-NEXT:    movdqa %xmm12, %xmm0
; SSE-NEXT:    pandn %xmm5, %xmm0
; SSE-NEXT:    pand %xmm12, %xmm6
; SSE-NEXT:    por %xmm6, %xmm0
; SSE-NEXT:    movdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshuflw $225, {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Folded Reload
; SSE-NEXT:    # xmm5 = mem[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,1,0,1]
; SSE-NEXT:    movdqa %xmm4, %xmm6
; SSE-NEXT:    pandn %xmm5, %xmm6
; SSE-NEXT:    movdqa (%rsp), %xmm3 # 16-byte Reload
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm3[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,0,0,0]
; SSE-NEXT:    pand %xmm4, %xmm5
; SSE-NEXT:    por %xmm5, %xmm6
; SSE-NEXT:    movdqa %xmm8, %xmm5
; SSE-NEXT:    pandn %xmm6, %xmm5
; SSE-NEXT:    pshuflw $164, {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Folded Reload
; SSE-NEXT:    # xmm6 = mem[0,1,2,2,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[0,0,1,3]
; SSE-NEXT:    movdqa %xmm12, %xmm7
; SSE-NEXT:    pandn %xmm6, %xmm7
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm9 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm9[0,0,0,0]
; SSE-NEXT:    pand %xmm12, %xmm6
; SSE-NEXT:    por %xmm6, %xmm7
; SSE-NEXT:    pand %xmm8, %xmm7
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[0,0,0,0]
; SSE-NEXT:    movdqa %xmm2, %xmm1
; SSE-NEXT:    pandn %xmm5, %xmm1
; SSE-NEXT:    pand %xmm2, %xmm7
; SSE-NEXT:    por %xmm7, %xmm1
; SSE-NEXT:    movdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; SSE-NEXT:    pshufhw $237, {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Folded Reload
; SSE-NEXT:    # xmm5 = mem[0,1,2,3,5,7,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,2,3,2]
; SSE-NEXT:    movdqa %xmm2, %xmm6
; SSE-NEXT:    pandn %xmm5, %xmm6
; SSE-NEXT:    pshufhw {{.*#+}} xmm5 = xmm3[0,1,2,3,7,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[2,2,2,2]
; SSE-NEXT:    pand %xmm2, %xmm5
; SSE-NEXT:    por %xmm5, %xmm6
; SSE-NEXT:    movdqa {{.*#+}} xmm1 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255]
; SSE-NEXT:    movdqa %xmm1, %xmm5
; SSE-NEXT:    pandn %xmm6, %xmm5
; SSE-NEXT:    pshufhw $167, {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Folded Reload
; SSE-NEXT:    # xmm6 = mem[0,1,2,3,7,5,6,6]
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[2,3,2,2]
; SSE-NEXT:    movdqa %xmm15, %xmm7
; SSE-NEXT:    pandn %xmm6, %xmm7
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm9[3,3,3,3]
; SSE-NEXT:    pand %xmm15, %xmm6
; SSE-NEXT:    por %xmm6, %xmm7
; SSE-NEXT:    pand %xmm1, %xmm7
; SSE-NEXT:    por %xmm5, %xmm7
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[3,3,3,3]
; SSE-NEXT:    movdqa {{.*#+}} xmm9 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; SSE-NEXT:    movdqa %xmm9, %xmm13
; SSE-NEXT:    pandn %xmm5, %xmm13
; SSE-NEXT:    pand %xmm9, %xmm7
; SSE-NEXT:    por %xmm7, %xmm13
; SSE-NEXT:    pshuflw $225, {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Folded Reload
; SSE-NEXT:    # xmm5 = mem[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,1,0,1]
; SSE-NEXT:    movdqa %xmm4, %xmm6
; SSE-NEXT:    pandn %xmm5, %xmm6
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm11 # 16-byte Reload
; SSE-NEXT:    pshuflw {{.*#+}} xmm5 = xmm11[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[0,0,0,0]
; SSE-NEXT:    pand %xmm4, %xmm5
; SSE-NEXT:    por %xmm5, %xmm6
; SSE-NEXT:    movdqa %xmm8, %xmm5
; SSE-NEXT:    pandn %xmm6, %xmm5
; SSE-NEXT:    pshuflw $164, {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Folded Reload
; SSE-NEXT:    # xmm6 = mem[0,1,2,2,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[0,0,1,3]
; SSE-NEXT:    movdqa %xmm12, %xmm3
; SSE-NEXT:    pandn %xmm6, %xmm3
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm1[0,0,0,0]
; SSE-NEXT:    pand %xmm12, %xmm6
; SSE-NEXT:    por %xmm6, %xmm3
; SSE-NEXT:    pand %xmm8, %xmm3
; SSE-NEXT:    por %xmm5, %xmm3
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm0[0,0,0,0]
; SSE-NEXT:    movdqa %xmm2, %xmm10
; SSE-NEXT:    pandn %xmm5, %xmm10
; SSE-NEXT:    pand %xmm2, %xmm3
; SSE-NEXT:    por %xmm3, %xmm10
; SSE-NEXT:    pshufhw $237, {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Folded Reload
; SSE-NEXT:    # xmm3 = mem[0,1,2,3,5,7,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,2,3,2]
; SSE-NEXT:    movdqa %xmm2, %xmm5
; SSE-NEXT:    pandn %xmm3, %xmm5
; SSE-NEXT:    pshufhw {{.*#+}} xmm3 = xmm11[0,1,2,3,7,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[2,2,2,2]
; SSE-NEXT:    pand %xmm2, %xmm3
; SSE-NEXT:    por %xmm3, %xmm5
; SSE-NEXT:    movdqa {{.*#+}} xmm7 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255]
; SSE-NEXT:    movdqa %xmm7, %xmm3
; SSE-NEXT:    pandn %xmm5, %xmm3
; SSE-NEXT:    pshufhw $167, {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Folded Reload
; SSE-NEXT:    # xmm5 = mem[0,1,2,3,7,5,6,6]
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm5[2,3,2,2]
; SSE-NEXT:    movdqa %xmm15, %xmm6
; SSE-NEXT:    pandn %xmm5, %xmm6
; SSE-NEXT:    pshufd {{.*#+}} xmm5 = xmm1[3,3,3,3]
; SSE-NEXT:    pand %xmm15, %xmm5
; SSE-NEXT:    por %xmm5, %xmm6
; SSE-NEXT:    pand %xmm7, %xmm6
; SSE-NEXT:    por %xmm3, %xmm6
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm0[3,3,3,3]
; SSE-NEXT:    movdqa %xmm9, %xmm7
; SSE-NEXT:    pandn %xmm3, %xmm7
; SSE-NEXT:    pand %xmm9, %xmm6
; SSE-NEXT:    por %xmm6, %xmm7
; SSE-NEXT:    pshuflw $225, {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Folded Reload
; SSE-NEXT:    # xmm3 = mem[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,1,0,1]
; SSE-NEXT:    movdqa %xmm4, %xmm6
; SSE-NEXT:    pandn %xmm3, %xmm6
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Reload
; SSE-NEXT:    pshuflw {{.*#+}} xmm3 = xmm5[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,0,0,0]
; SSE-NEXT:    pand %xmm4, %xmm3
; SSE-NEXT:    por %xmm3, %xmm6
; SSE-NEXT:    movdqa %xmm8, %xmm3
; SSE-NEXT:    pandn %xmm6, %xmm3
; SSE-NEXT:    pshuflw {{.*#+}} xmm6 = xmm14[0,1,2,2,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[0,0,1,3]
; SSE-NEXT:    movdqa %xmm12, %xmm11
; SSE-NEXT:    pandn %xmm6, %xmm11
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm0[0,0,0,0]
; SSE-NEXT:    pand %xmm12, %xmm6
; SSE-NEXT:    por %xmm6, %xmm11
; SSE-NEXT:    pand %xmm8, %xmm11
; SSE-NEXT:    por %xmm3, %xmm11
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[0,0,0,0]
; SSE-NEXT:    movdqa %xmm2, %xmm14
; SSE-NEXT:    pandn %xmm3, %xmm14
; SSE-NEXT:    pand %xmm2, %xmm11
; SSE-NEXT:    por %xmm11, %xmm14
; SSE-NEXT:    pshufhw $237, {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Folded Reload
; SSE-NEXT:    # xmm3 = mem[0,1,2,3,5,7,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,2,3,2]
; SSE-NEXT:    movdqa %xmm2, %xmm6
; SSE-NEXT:    pandn %xmm3, %xmm6
; SSE-NEXT:    pshufhw {{.*#+}} xmm3 = xmm5[0,1,2,3,7,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[2,2,2,2]
; SSE-NEXT:    pand %xmm2, %xmm3
; SSE-NEXT:    por %xmm3, %xmm6
; SSE-NEXT:    movdqa {{.*#+}} xmm5 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255]
; SSE-NEXT:    movdqa %xmm5, %xmm3
; SSE-NEXT:    pandn %xmm6, %xmm3
; SSE-NEXT:    pshufhw $167, {{[-0-9]+}}(%r{{[sb]}}p), %xmm6 # 16-byte Folded Reload
; SSE-NEXT:    # xmm6 = mem[0,1,2,3,7,5,6,6]
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm6[2,3,2,2]
; SSE-NEXT:    movdqa %xmm15, %xmm11
; SSE-NEXT:    pandn %xmm6, %xmm11
; SSE-NEXT:    pshufd {{.*#+}} xmm6 = xmm0[3,3,3,3]
; SSE-NEXT:    pand %xmm15, %xmm6
; SSE-NEXT:    por %xmm6, %xmm11
; SSE-NEXT:    pand %xmm5, %xmm11
; SSE-NEXT:    por %xmm3, %xmm11
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm1[3,3,3,3]
; SSE-NEXT:    movdqa %xmm9, %xmm6
; SSE-NEXT:    pandn %xmm3, %xmm6
; SSE-NEXT:    pand %xmm9, %xmm11
; SSE-NEXT:    por %xmm11, %xmm6
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
; SSE-NEXT:    pshuflw {{.*#+}} xmm3 = xmm1[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,0,0,0]
; SSE-NEXT:    pand %xmm4, %xmm3
; SSE-NEXT:    pshuflw $225, {{[-0-9]+}}(%r{{[sb]}}p), %xmm11 # 16-byte Folded Reload
; SSE-NEXT:    # xmm11 = mem[1,0,2,3,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm11 = xmm11[0,1,0,1]
; SSE-NEXT:    pandn %xmm11, %xmm4
; SSE-NEXT:    por %xmm3, %xmm4
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm5 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm5[0,0,0,0]
; SSE-NEXT:    pand %xmm12, %xmm3
; SSE-NEXT:    pshuflw $164, {{[-0-9]+}}(%r{{[sb]}}p), %xmm11 # 16-byte Folded Reload
; SSE-NEXT:    # xmm11 = mem[0,1,2,2,4,5,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm11 = xmm11[0,0,1,3]
; SSE-NEXT:    pandn %xmm11, %xmm12
; SSE-NEXT:    por %xmm3, %xmm12
; SSE-NEXT:    pand %xmm8, %xmm12
; SSE-NEXT:    pandn %xmm4, %xmm8
; SSE-NEXT:    por %xmm12, %xmm8
; SSE-NEXT:    movdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm4 # 16-byte Reload
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm4[0,0,0,0]
; SSE-NEXT:    movdqa %xmm2, %xmm0
; SSE-NEXT:    pandn %xmm3, %xmm0
; SSE-NEXT:    pand %xmm2, %xmm8
; SSE-NEXT:    por %xmm8, %xmm0
; SSE-NEXT:    pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,7,6,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm1[2,2,2,2]
; SSE-NEXT:    pand %xmm2, %xmm1
; SSE-NEXT:    pshufhw $237, {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Folded Reload
; SSE-NEXT:    # xmm3 = mem[0,1,2,3,5,7,6,7]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[0,2,3,2]
; SSE-NEXT:    pandn %xmm3, %xmm2
; SSE-NEXT:    por %xmm1, %xmm2
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm5[3,3,3,3]
; SSE-NEXT:    pand %xmm15, %xmm1
; SSE-NEXT:    pshufhw $167, {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Folded Reload
; SSE-NEXT:    # xmm3 = mem[0,1,2,3,7,5,6,6]
; SSE-NEXT:    pshufd {{.*#+}} xmm3 = xmm3[2,3,2,2]
; SSE-NEXT:    pandn %xmm3, %xmm15
; SSE-NEXT:    por %xmm1, %xmm15
; SSE-NEXT:    movdqa {{.*#+}} xmm1 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255]
; SSE-NEXT:    pand %xmm1, %xmm15
; SSE-NEXT:    pandn %xmm2, %xmm1
; SSE-NEXT:    por %xmm15, %xmm1
; SSE-NEXT:    pand %xmm9, %xmm1
; SSE-NEXT:    movdqa %xmm1, %xmm2
; SSE-NEXT:    pshufd {{.*#+}} xmm1 = xmm4[3,3,3,3]
; SSE-NEXT:    pandn %xmm1, %xmm9
; SSE-NEXT:    por %xmm2, %xmm9
; SSE-NEXT:    movdqa %xmm9, 304(%r9)
; SSE-NEXT:    movdqa %xmm0, 240(%r9)
; SSE-NEXT:    movdqa %xmm6, 224(%r9)
; SSE-NEXT:    movdqa %xmm14, 160(%r9)
; SSE-NEXT:    movdqa %xmm7, 144(%r9)
; SSE-NEXT:    movdqa %xmm10, 80(%r9)
; SSE-NEXT:    movdqa %xmm13, 64(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, (%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 288(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 256(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 208(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 176(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 128(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 96(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 48(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 16(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 272(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 192(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 112(%r9)
; SSE-NEXT:    movaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; SSE-NEXT:    movaps %xmm0, 32(%r9)
; SSE-NEXT:    addq $504, %rsp # imm = 0x1F8
; SSE-NEXT:    retq
;
; AVX1-ONLY-LABEL: store_i8_stride5_vf64:
; AVX1-ONLY:       # %bb.0:
; AVX1-ONLY-NEXT:    subq $104, %rsp
; AVX1-ONLY-NEXT:    vmovdqa 48(%rcx), %xmm0
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm1 = [128,6,128,8,0,128,7,128,128,6,128,8,0,128,7,128]
; AVX1-ONLY-NEXT:    # xmm1 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm1, %xmm0, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa %xmm1, %xmm14
; AVX1-ONLY-NEXT:    vmovdqa 48(%rdx), %xmm1
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = xmm1[6],zero,xmm1[u,u,u,7],zero,xmm1[u,u,u,8],zero,xmm1[u,u,u,9]
; AVX1-ONLY-NEXT:    vpor %xmm2, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm9 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3],xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm9[u,6,7,u,u,u,8,9,u,u,u,10,11,u,u,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm3, %ymm4, %ymm3
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm5 = [255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0]
; AVX1-ONLY-NEXT:    vandnps %ymm3, %ymm5, %ymm4
; AVX1-ONLY-NEXT:    vmovaps %ymm5, %ymm2
; AVX1-ONLY-NEXT:    vmovdqa 48(%rsi), %xmm3
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm5 = [128,8,0,128,7,128,9,0,128,8,0,128,7,128,9,0]
; AVX1-ONLY-NEXT:    # xmm5 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm5, %xmm3, %xmm6
; AVX1-ONLY-NEXT:    vmovdqa %xmm5, %xmm10
; AVX1-ONLY-NEXT:    vmovdqa 48(%rdi), %xmm5
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm8 = [8,128,0,7,128,9,128,0,8,128,0,7,128,9,128,0]
; AVX1-ONLY-NEXT:    # xmm8 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm8, %xmm5, %xmm7
; AVX1-ONLY-NEXT:    vmovdqa %xmm8, %xmm12
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm7, %xmm6
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm7 = xmm3[0],xmm5[0],xmm3[1],xmm5[1],xmm3[2],xmm5[2],xmm3[3],xmm5[3],xmm3[4],xmm5[4],xmm3[5],xmm5[5],xmm3[6],xmm5[6],xmm3[7],xmm5[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm7 = xmm7[6,u,u,u,9,8,u,u,u,11,10,u,u,u,13,12]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm6, %ymm7, %ymm6
; AVX1-ONLY-NEXT:    vandps %ymm2, %ymm6, %ymm6
; AVX1-ONLY-NEXT:    vorps %ymm4, %ymm6, %ymm4
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm4, %xmm6
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm6[0,1],zero,xmm6[3,4,5,6],zero,xmm6[8,9,10,11],zero,xmm6[13,14,15]
; AVX1-ONLY-NEXT:    vmovdqa 48(%r8), %xmm7
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm8 = zero,zero,xmm7[6],zero,zero,zero,zero,xmm7[7],zero,zero,zero,zero,xmm7[8],zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm8, %xmm6
; AVX1-ONLY-NEXT:    vmovdqa %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm4[0,1,2],zero,xmm4[4,5,6,7],zero,xmm4[9,10,11,12],zero,xmm4[14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = zero,zero,zero,xmm7[3],zero,zero,zero,zero,xmm7[4],zero,zero,zero,zero,xmm7[5],zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm4, %xmm4
; AVX1-ONLY-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm4 = xmm5[8],xmm3[8],xmm5[9],xmm3[9],xmm5[10],xmm3[10],xmm5[11],xmm3[11],xmm5[12],xmm3[12],xmm5[13],xmm3[13],xmm5[14],xmm3[14],xmm5[15],xmm3[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm4[u,10,11,u,u,u,12,13,u,u,u,14,15,u,u,u]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm8 = [7,0,4,5,8,9,0,6,7,0,4,5,8,9,0,6]
; AVX1-ONLY-NEXT:    # xmm8 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm8, %xmm4, %xmm4
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm6, %ymm4, %ymm4
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm6 = xmm1[8],xmm0[8],xmm1[9],xmm0[9],xmm1[10],xmm0[10],xmm1[11],xmm0[11],xmm1[12],xmm0[12],xmm1[13],xmm0[13],xmm1[14],xmm0[14],xmm1[15],xmm0[15]
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm1[8],xmm0[9],xmm1[9],xmm0[10],xmm1[10],xmm0[11],xmm1[11],xmm0[12],xmm1[12],xmm0[13],xmm1[13],xmm0[14],xmm1[14],xmm0[15],xmm1[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm1 = xmm6[u,u,u,10,11,u,u,u,12,13,u,u,u,14,15,u]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm6 = [2,7,6,0,5,4,9,8,2,7,6,0,5,4,9,8]
; AVX1-ONLY-NEXT:    # xmm6 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm6, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vmovdqa %xmm6, %xmm15
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm11 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; AVX1-ONLY-NEXT:    vandnps %ymm4, %ymm11, %ymm1
; AVX1-ONLY-NEXT:    vandps %ymm0, %ymm11, %ymm0
; AVX1-ONLY-NEXT:    vorps %ymm1, %ymm0, %ymm0
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm0, %xmm1
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm1 = zero,xmm1[1,2,3,4],zero,xmm1[6,7,8,9],zero,xmm1[11,12,13,14],zero
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm7[12],zero,zero,zero,zero,xmm7[13],zero,zero,zero,zero,xmm7[14],zero,zero,zero,zero,xmm7[15]
; AVX1-ONLY-NEXT:    vpor %xmm4, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vmovdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm13 = [0,128,2,3,4,5,128,7,8,9,10,128,12,13,14,15]
; AVX1-ONLY-NEXT:    vpshufb %xmm13, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm2 = [128,9,128,128,128,128,10,128,128,128,128,11,128,128,128,128]
; AVX1-ONLY-NEXT:    vpshufb %xmm2, %xmm7, %xmm1
; AVX1-ONLY-NEXT:    vpor %xmm1, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vmovdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vmovdqa 32(%rsi), %xmm0
; AVX1-ONLY-NEXT:    vmovdqa 32(%rdi), %xmm4
; AVX1-ONLY-NEXT:    vpshufb %xmm10, %xmm0, %xmm1
; AVX1-ONLY-NEXT:    vpshufb %xmm12, %xmm4, %xmm6
; AVX1-ONLY-NEXT:    vpor %xmm1, %xmm6, %xmm1
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm12 = xmm4[8],xmm0[8],xmm4[9],xmm0[9],xmm4[10],xmm0[10],xmm4[11],xmm0[11],xmm4[12],xmm0[12],xmm4[13],xmm0[13],xmm4[14],xmm0[14],xmm4[15],xmm0[15]
; AVX1-ONLY-NEXT:    vpshufb %xmm8, %xmm12, %xmm6
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm6, %ymm1, %ymm6
; AVX1-ONLY-NEXT:    vmovdqa 32(%rcx), %xmm1
; AVX1-ONLY-NEXT:    vmovdqa 32(%rdx), %xmm8
; AVX1-ONLY-NEXT:    vpshufb %xmm14, %xmm1, %xmm10
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm14 = xmm8[6],zero,xmm8[u,u,u,7],zero,xmm8[u,u,u,8],zero,xmm8[u,u,u,9]
; AVX1-ONLY-NEXT:    vpor %xmm10, %xmm14, %xmm10
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm14 = xmm1[8],xmm8[8],xmm1[9],xmm8[9],xmm1[10],xmm8[10],xmm1[11],xmm8[11],xmm1[12],xmm8[12],xmm1[13],xmm8[13],xmm1[14],xmm8[14],xmm1[15],xmm8[15]
; AVX1-ONLY-NEXT:    vpshufb %xmm15, %xmm14, %xmm14
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm14, %ymm10, %ymm14
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm10 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; AVX1-ONLY-NEXT:    vandnps %ymm6, %ymm10, %ymm6
; AVX1-ONLY-NEXT:    vandps %ymm10, %ymm14, %ymm14
; AVX1-ONLY-NEXT:    vorps %ymm6, %ymm14, %ymm14
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm14, %xmm6
; AVX1-ONLY-NEXT:    vpshufb %xmm13, %xmm6, %xmm15
; AVX1-ONLY-NEXT:    vmovdqa 32(%r8), %xmm6
; AVX1-ONLY-NEXT:    vpshufb %xmm2, %xmm6, %xmm13
; AVX1-ONLY-NEXT:    vpor %xmm13, %xmm15, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm13 = xmm14[0,1],zero,xmm14[3,4,5,6],zero,xmm14[8,9,10,11],zero,xmm14[13,14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm14 = zero,zero,xmm6[6],zero,zero,zero,zero,xmm6[7],zero,zero,zero,zero,xmm6[8],zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm14, %xmm13, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa %xmm2, (%rsp) # 16-byte Spill
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm14 = [3,0,0,1,4,5,0,2,3,0,0,1,4,5,0,2]
; AVX1-ONLY-NEXT:    # xmm14 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm14, %xmm9, %xmm2
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm13 = xmm8[8],xmm1[8],xmm8[9],xmm1[9],xmm8[10],xmm1[10],xmm8[11],xmm1[11],xmm8[12],xmm1[12],xmm8[13],xmm1[13],xmm8[14],xmm1[14],xmm8[15],xmm1[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm13 = xmm13[u,u,u,10,11,u,u,u,12,13,u,u,u,14,15,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm2, %ymm13, %ymm2
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm3 = xmm5[0],xmm3[0],xmm5[1],xmm3[1],xmm5[2],xmm3[2],xmm5[3],xmm3[3],xmm5[4],xmm3[4],xmm5[5],xmm3[5],xmm5[6],xmm3[6],xmm5[7],xmm3[7]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm13 = [0,1,4,5,0,2,3,6,0,1,4,5,0,2,3,6]
; AVX1-ONLY-NEXT:    # xmm13 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm13, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm9 = [0,10,11,14,15,0,12,13,0,10,11,14,15,0,12,13]
; AVX1-ONLY-NEXT:    # xmm9 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm9, %xmm12, %xmm5
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm3, %ymm5, %ymm3
; AVX1-ONLY-NEXT:    vandnps %ymm2, %ymm10, %ymm2
; AVX1-ONLY-NEXT:    vandps %ymm3, %ymm10, %ymm3
; AVX1-ONLY-NEXT:    vorps %ymm2, %ymm3, %ymm2
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm2, %xmm3
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm5 = [0,1,2,3,128,5,6,7,8,128,10,11,12,13,128,15]
; AVX1-ONLY-NEXT:    vpshufb %xmm5, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vmovdqa %xmm5, %xmm12
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm10 = [128,128,128,128,0,128,128,128,128,1,128,128,128,128,2,128]
; AVX1-ONLY-NEXT:    vpshufb %xmm10, %xmm7, %xmm5
; AVX1-ONLY-NEXT:    vmovdqa %xmm10, %xmm7
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vmovdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm5 = [128,1,2,3,4,128,6,7,8,9,128,11,12,13,14,128]
; AVX1-ONLY-NEXT:    vpshufb %xmm5, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm10 = [12,128,128,128,128,13,128,128,128,128,14,128,128,128,128,15]
; AVX1-ONLY-NEXT:    vpshufb %xmm10, %xmm6, %xmm3
; AVX1-ONLY-NEXT:    vpor %xmm3, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm1 = xmm8[0],xmm1[0],xmm8[1],xmm1[1],xmm8[2],xmm1[2],xmm8[3],xmm1[3],xmm8[4],xmm1[4],xmm8[5],xmm1[5],xmm8[6],xmm1[6],xmm8[7],xmm1[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm1[u,6,7,u,u,u,8,9,u,u,u,10,11,u,u,u]
; AVX1-ONLY-NEXT:    vpshufb %xmm14, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm2, %ymm1, %ymm1
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm2 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3],xmm0[4],xmm4[4],xmm0[5],xmm4[5],xmm0[6],xmm4[6],xmm0[7],xmm4[7]
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm4[0],xmm0[0],xmm4[1],xmm0[1],xmm4[2],xmm0[2],xmm4[3],xmm0[3],xmm4[4],xmm0[4],xmm4[5],xmm0[5],xmm4[6],xmm0[6],xmm4[7],xmm0[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm2[6,u,u,u,9,8,u,u,u,11,10,u,u,u,13,12]
; AVX1-ONLY-NEXT:    vpshufb %xmm13, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm2, %ymm0, %ymm0
; AVX1-ONLY-NEXT:    vandnps %ymm1, %ymm11, %ymm1
; AVX1-ONLY-NEXT:    vandps %ymm0, %ymm11, %ymm0
; AVX1-ONLY-NEXT:    vorps %ymm1, %ymm0, %ymm0
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm0, %xmm1
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm1 = xmm1[0,1,2],zero,xmm1[4,5,6,7],zero,xmm1[9,10,11,12],zero,xmm1[14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = zero,zero,zero,xmm6[3],zero,zero,zero,zero,xmm6[4],zero,zero,zero,zero,xmm6[5],zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm2, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vmovdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb %xmm12, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vpshufb %xmm7, %xmm6, %xmm1
; AVX1-ONLY-NEXT:    vpor %xmm1, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vmovdqa %xmm0, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vmovdqa 16(%rsi), %xmm8
; AVX1-ONLY-NEXT:    vmovdqa 16(%rdi), %xmm6
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm0 = xmm6[8],xmm8[8],xmm6[9],xmm8[9],xmm6[10],xmm8[10],xmm6[11],xmm8[11],xmm6[12],xmm8[12],xmm6[13],xmm8[13],xmm6[14],xmm8[14],xmm6[15],xmm8[15]
; AVX1-ONLY-NEXT:    vpshufb %xmm9, %xmm0, %xmm1
; AVX1-ONLY-NEXT:    vmovdqa %xmm9, %xmm15
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm0 = xmm0[u,u,4,5,u,u,u,6,7,u,u,u,8,9,u,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm1, %ymm0, %ymm0
; AVX1-ONLY-NEXT:    vmovdqa 16(%rcx), %xmm1
; AVX1-ONLY-NEXT:    vmovdqa 16(%rdx), %xmm2
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm3 = xmm2[8],xmm1[8],xmm2[9],xmm1[9],xmm2[10],xmm1[10],xmm2[11],xmm1[11],xmm2[12],xmm1[12],xmm2[13],xmm1[13],xmm2[14],xmm1[14],xmm2[15],xmm1[15]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm12 = [12,13,0,10,11,14,15,0,12,13,0,10,11,14,15,0]
; AVX1-ONLY-NEXT:    # xmm12 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm12, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm4 = xmm1[8],xmm2[8],xmm1[9],xmm2[9],xmm1[10],xmm2[10],xmm1[11],xmm2[11],xmm1[12],xmm2[12],xmm1[13],xmm2[13],xmm1[14],xmm2[14],xmm1[15],xmm2[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm4[2,u,u,u,5,4,u,u,u,7,6,u,u,u,9,8]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm3, %ymm4, %ymm3
; AVX1-ONLY-NEXT:    vandnps %ymm0, %ymm11, %ymm0
; AVX1-ONLY-NEXT:    vandps %ymm3, %ymm11, %ymm3
; AVX1-ONLY-NEXT:    vorps %ymm0, %ymm3, %ymm3
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm3, %xmm0
; AVX1-ONLY-NEXT:    vpshufb %xmm5, %xmm0, %xmm4
; AVX1-ONLY-NEXT:    vmovdqa %xmm5, %xmm11
; AVX1-ONLY-NEXT:    vmovdqa 16(%r8), %xmm0
; AVX1-ONLY-NEXT:    vpshufb %xmm10, %xmm0, %xmm7
; AVX1-ONLY-NEXT:    vpor %xmm7, %xmm4, %xmm4
; AVX1-ONLY-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = xmm3[0],zero,xmm3[2,3,4,5],zero,xmm3[7,8,9,10],zero,xmm3[12,13,14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = zero,xmm0[9],zero,zero,zero,zero,xmm0[10],zero,zero,zero,zero,xmm0[11],zero,zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm4, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vmovdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vmovdqa (%rcx), %xmm9
; AVX1-ONLY-NEXT:    vmovdqa (%rdx), %xmm7
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm3 = xmm7[8],xmm9[8],xmm7[9],xmm9[9],xmm7[10],xmm9[10],xmm7[11],xmm9[11],xmm7[12],xmm9[12],xmm7[13],xmm9[13],xmm7[14],xmm9[14],xmm7[15],xmm9[15]
; AVX1-ONLY-NEXT:    vpshufb %xmm12, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm4 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3],xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],xmm1[7]
; AVX1-ONLY-NEXT:    vpshufb %xmm14, %xmm4, %xmm5
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm5, %ymm3, %ymm12
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm3 = xmm6[0],xmm8[0],xmm6[1],xmm8[1],xmm6[2],xmm8[2],xmm6[3],xmm8[3],xmm6[4],xmm8[4],xmm6[5],xmm8[5],xmm6[6],xmm8[6],xmm6[7],xmm8[7]
; AVX1-ONLY-NEXT:    vpshufb %xmm13, %xmm3, %xmm13
; AVX1-ONLY-NEXT:    vmovdqa (%rsi), %xmm5
; AVX1-ONLY-NEXT:    vmovdqa (%rdi), %xmm3
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm10 = xmm3[8],xmm5[8],xmm3[9],xmm5[9],xmm3[10],xmm5[10],xmm3[11],xmm5[11],xmm3[12],xmm5[12],xmm3[13],xmm5[13],xmm3[14],xmm5[14],xmm3[15],xmm5[15]
; AVX1-ONLY-NEXT:    vpshufb %xmm15, %xmm10, %xmm14
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm13, %ymm14, %ymm13
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm14 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; AVX1-ONLY-NEXT:    vandnps %ymm12, %ymm14, %ymm12
; AVX1-ONLY-NEXT:    vandps %ymm14, %ymm13, %ymm13
; AVX1-ONLY-NEXT:    vorps %ymm12, %ymm13, %ymm12
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm12, %xmm13
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm13 = xmm13[0,1,2,3],zero,xmm13[5,6,7,8],zero,xmm13[10,11,12,13],zero,xmm13[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm14 = zero,zero,zero,zero,xmm0[0],zero,zero,zero,zero,xmm0[1],zero,zero,zero,zero,xmm0[2],zero
; AVX1-ONLY-NEXT:    vpor %xmm14, %xmm13, %xmm13
; AVX1-ONLY-NEXT:    vmovdqa %xmm13, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vpshufb %xmm11, %xmm12, %xmm12
; AVX1-ONLY-NEXT:    vmovdqa (%r8), %xmm13
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm15 = xmm13[12],zero,zero,zero,zero,xmm13[13],zero,zero,zero,zero,xmm13[14],zero,zero,zero,zero,xmm13[15]
; AVX1-ONLY-NEXT:    vpor %xmm15, %xmm12, %xmm11
; AVX1-ONLY-NEXT:    vmovdqa %xmm11, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm11 = [128,6,128,8,0,128,7,128,128,6,128,8,0,128,7,128]
; AVX1-ONLY-NEXT:    # xmm11 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm11, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm2[6],zero,xmm2[u,u,u,7],zero,xmm2[u,u,u,8],zero,xmm2[u,u,u,9]
; AVX1-ONLY-NEXT:    vpor %xmm1, %xmm2, %xmm1
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm11 = [0,6,7,10,11,0,8,9,0,6,7,10,11,0,8,9]
; AVX1-ONLY-NEXT:    # xmm11 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm11, %xmm4, %xmm2
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm1, %ymm2, %ymm1
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm8[u,u,u],zero,xmm8[7,u,u,u],zero,xmm8[8,u,u,u],zero,xmm8[9,u]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm6[u,u,u,7],zero,xmm6[u,u,u,8],zero,xmm6[u,u,u,9],zero,xmm6[u]
; AVX1-ONLY-NEXT:    vpor %xmm2, %xmm4, %xmm2
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm4 = xmm8[0],xmm6[0],xmm8[1],xmm6[1],xmm8[2],xmm6[2],xmm8[3],xmm6[3],xmm8[4],xmm6[4],xmm8[5],xmm6[5],xmm8[6],xmm6[6],xmm8[7],xmm6[7]
; AVX1-ONLY-NEXT:    vmovddup {{.*#+}} xmm6 = [6,11,10,0,9,8,13,12,6,11,10,0,9,8,13,12]
; AVX1-ONLY-NEXT:    # xmm6 = mem[0,0]
; AVX1-ONLY-NEXT:    vpshufb %xmm6, %xmm4, %xmm4
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm2, %ymm4, %ymm2
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm4 = [255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0]
; AVX1-ONLY-NEXT:    vandnps %ymm1, %ymm4, %ymm1
; AVX1-ONLY-NEXT:    vandps %ymm4, %ymm2, %ymm2
; AVX1-ONLY-NEXT:    vorps %ymm1, %ymm2, %ymm2
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm2, %xmm1
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm14 = [0,1,128,3,4,5,6,128,8,9,10,11,128,13,14,15]
; AVX1-ONLY-NEXT:    vpshufb %xmm14, %xmm1, %xmm1
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm15 = [128,128,6,128,128,128,128,7,128,128,128,128,8,128,128,128]
; AVX1-ONLY-NEXT:    vpshufb %xmm15, %xmm0, %xmm4
; AVX1-ONLY-NEXT:    vpor %xmm4, %xmm1, %xmm12
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm1 = [0,1,2,128,4,5,6,7,128,9,10,11,12,128,14,15]
; AVX1-ONLY-NEXT:    vpshufb %xmm1, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vmovdqa {{.*#+}} xmm8 = [128,128,128,3,128,128,128,128,4,128,128,128,128,5,128,128]
; AVX1-ONLY-NEXT:    vpshufb %xmm8, %xmm0, %xmm0
; AVX1-ONLY-NEXT:    vpor %xmm0, %xmm2, %xmm0
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm2 = xmm7[0],xmm9[0],xmm7[1],xmm9[1],xmm7[2],xmm9[2],xmm7[3],xmm9[3],xmm7[4],xmm9[4],xmm7[5],xmm9[5],xmm7[6],xmm9[6],xmm7[7],xmm9[7]
; AVX1-ONLY-NEXT:    vpshufb %xmm11, %xmm2, %xmm4
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm2 = xmm2[u,u,0,1,u,u,u,2,3,u,u,u,4,5,u,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm4, %ymm2, %ymm2
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm4 = xmm5[0],xmm3[0],xmm5[1],xmm3[1],xmm5[2],xmm3[2],xmm5[3],xmm3[3],xmm5[4],xmm3[4],xmm5[5],xmm3[5],xmm5[6],xmm3[6],xmm5[7],xmm3[7]
; AVX1-ONLY-NEXT:    vpshufb %xmm6, %xmm4, %xmm4
; AVX1-ONLY-NEXT:    vpunpcklbw {{.*#+}} xmm6 = xmm3[0],xmm5[0],xmm3[1],xmm5[1],xmm3[2],xmm5[2],xmm3[3],xmm5[3],xmm3[4],xmm5[4],xmm3[5],xmm5[5],xmm3[6],xmm5[6],xmm3[7],xmm5[7]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm6[0,1,u,u,u,2,3,u,u,u,4,5,u,u,u,6]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm4, %ymm6, %ymm4
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm6 = [255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; AVX1-ONLY-NEXT:    vandnps %ymm2, %ymm6, %ymm2
; AVX1-ONLY-NEXT:    vandps %ymm6, %ymm4, %ymm4
; AVX1-ONLY-NEXT:    vorps %ymm2, %ymm4, %ymm4
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm4, %xmm2
; AVX1-ONLY-NEXT:    vpshufb %xmm1, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vpshufb %xmm8, %xmm13, %xmm6
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm2, %xmm2
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm4 = xmm4[0,1,2,3],zero,xmm4[5,6,7,8],zero,xmm4[10,11,12,13],zero,xmm4[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = zero,zero,zero,zero,xmm13[0],zero,zero,zero,zero,xmm13[1],zero,zero,zero,zero,xmm13[2],zero
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm4, %xmm4
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = xmm5[u,u,u],zero,xmm5[7,u,u,u],zero,xmm5[8,u,u,u],zero,xmm5[9,u]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm3 = xmm3[u,u,u,7],zero,xmm3[u,u,u,8],zero,xmm3[u,u,u,9],zero,xmm3[u]
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = xmm10[u,u,4,5,u,u,u,6,7,u,u,u,8,9,u,u]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm5, %ymm3, %ymm3
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = zero,xmm9[6,u,u,u],zero,xmm9[7,u,u,u],zero,xmm9[8,u,u,u],zero
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm7[6],zero,xmm7[u,u,u,7],zero,xmm7[u,u,u,8],zero,xmm7[u,u,u,9]
; AVX1-ONLY-NEXT:    vpor %xmm5, %xmm6, %xmm5
; AVX1-ONLY-NEXT:    vpunpckhbw {{.*#+}} xmm6 = xmm9[8],xmm7[8],xmm9[9],xmm7[9],xmm9[10],xmm7[10],xmm9[11],xmm7[11],xmm9[12],xmm7[12],xmm9[13],xmm7[13],xmm9[14],xmm7[14],xmm9[15],xmm7[15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = xmm6[2,u,u,u,5,4,u,u,u,7,6,u,u,u,9,8]
; AVX1-ONLY-NEXT:    vinsertf128 $1, %xmm6, %ymm5, %ymm5
; AVX1-ONLY-NEXT:    vmovaps {{.*#+}} ymm1 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255]
; AVX1-ONLY-NEXT:    vandnps %ymm3, %ymm1, %ymm3
; AVX1-ONLY-NEXT:    vandps %ymm1, %ymm5, %ymm5
; AVX1-ONLY-NEXT:    vorps %ymm3, %ymm5, %ymm3
; AVX1-ONLY-NEXT:    vextractf128 $1, %ymm3, %xmm5
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm5 = xmm5[0],zero,xmm5[2,3,4,5],zero,xmm5[7,8,9,10],zero,xmm5[12,13,14,15]
; AVX1-ONLY-NEXT:    vpshufb {{.*#+}} xmm6 = zero,xmm13[9],zero,zero,zero,zero,xmm13[10],zero,zero,zero,zero,xmm13[11],zero,zero,zero,zero
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm5, %xmm5
; AVX1-ONLY-NEXT:    vpshufb %xmm14, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vpshufb %xmm15, %xmm13, %xmm6
; AVX1-ONLY-NEXT:    vpor %xmm6, %xmm3, %xmm3
; AVX1-ONLY-NEXT:    vmovdqa %xmm3, 32(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm5, 48(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm4, (%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm2, 16(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm0, 96(%r9)
; AVX1-ONLY-NEXT:    vmovdqa %xmm12, 112(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 64(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 80(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 128(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 144(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 160(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 176(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 224(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 240(%r9)
; AVX1-ONLY-NEXT:    vmovaps (%rsp), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 192(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 208(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 288(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 304(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 256(%r9)
; AVX1-ONLY-NEXT:    vmovaps {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX1-ONLY-NEXT:    vmovaps %xmm0, 272(%r9)
; AVX1-ONLY-NEXT:    addq $104, %rsp
; AVX1-ONLY-NEXT:    vzeroupper
; AVX1-ONLY-NEXT:    retq
;
; AVX2-SLOW-LABEL: store_i8_stride5_vf64:
; AVX2-SLOW:       # %bb.0:
; AVX2-SLOW-NEXT:    subq $312, %rsp # imm = 0x138
; AVX2-SLOW-NEXT:    vmovdqa 32(%rdi), %ymm10
; AVX2-SLOW-NEXT:    vmovdqa (%rcx), %xmm1
; AVX2-SLOW-NEXT:    vmovdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa 32(%rcx), %xmm11
; AVX2-SLOW-NEXT:    vmovdqa %xmm11, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} xmm0 = <128,6,128,8,u,128,7,128,9,128,11,u,128,10,128,12>
; AVX2-SLOW-NEXT:    vpshufb %xmm0, %xmm1, %xmm1
; AVX2-SLOW-NEXT:    vmovdqa (%rdx), %xmm3
; AVX2-SLOW-NEXT:    vmovdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa 32(%rdx), %xmm12
; AVX2-SLOW-NEXT:    vmovdqa %xmm12, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} xmm2 = <6,128,8,128,u,7,128,9,128,11,128,u,10,128,12,128>
; AVX2-SLOW-NEXT:    vpshufb %xmm2, %xmm3, %xmm3
; AVX2-SLOW-NEXT:    vpor %xmm1, %xmm3, %xmm1
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,1]
; AVX2-SLOW-NEXT:    vmovdqa (%rdi), %xmm4
; AVX2-SLOW-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} xmm3 = <8,128,u,7,128,9,128,u,128,u,10,128,12,128,u,11>
; AVX2-SLOW-NEXT:    vpshufb %xmm3, %xmm4, %xmm4
; AVX2-SLOW-NEXT:    vmovdqa (%rsi), %xmm14
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} xmm5 = <128,8,u,128,7,128,9,u,11,u,128,10,128,12,u,128>
; AVX2-SLOW-NEXT:    vpshufb %xmm5, %xmm14, %xmm8
; AVX2-SLOW-NEXT:    vpor %xmm4, %xmm8, %xmm4
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,1,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm8 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm8, %ymm1, %ymm4, %ymm1
; AVX2-SLOW-NEXT:    vmovdqa (%r8), %xmm4
; AVX2-SLOW-NEXT:    vmovdqa %xmm4, (%rsp) # 16-byte Spill
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm4 = xmm4[1,1,2,2]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,1,1,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm9 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm9, %ymm1, %ymm4, %ymm1
; AVX2-SLOW-NEXT:    vmovdqu %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa 32(%rdi), %xmm4
; AVX2-SLOW-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-SLOW-NEXT:    vpshufb %xmm0, %xmm11, %xmm0
; AVX2-SLOW-NEXT:    vpshufb %xmm2, %xmm12, %xmm1
; AVX2-SLOW-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX2-SLOW-NEXT:    vmovdqa 32(%rsi), %xmm2
; AVX2-SLOW-NEXT:    vmovdqa %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-SLOW-NEXT:    vpshufb %xmm3, %xmm4, %xmm1
; AVX2-SLOW-NEXT:    vpshufb %xmm5, %xmm2, %xmm2
; AVX2-SLOW-NEXT:    vpor %xmm1, %xmm2, %xmm1
; AVX2-SLOW-NEXT:    vmovdqa 32(%rsi), %ymm2
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,1,1]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,1]
; AVX2-SLOW-NEXT:    vpblendvb %ymm8, %ymm0, %ymm1, %ymm0
; AVX2-SLOW-NEXT:    vmovdqa 32(%r8), %xmm1
; AVX2-SLOW-NEXT:    vmovdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[1,1,2,2]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,1,1,1]
; AVX2-SLOW-NEXT:    vpblendvb %ymm9, %ymm0, %ymm1, %ymm0
; AVX2-SLOW-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm15 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,21,128,19,20,128,22,128,24,128,22,23,128,25,128,23]
; AVX2-SLOW-NEXT:    vpshufb %ymm15, %ymm10, %ymm1
; AVX2-SLOW-NEXT:    vmovdqa %ymm10, %ymm11
; AVX2-SLOW-NEXT:    vmovdqu %ymm10, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX2-SLOW-NEXT:    vbroadcasti128 {{.*#+}} ymm5 = [19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128]
; AVX2-SLOW-NEXT:    # ymm5 = mem[0,1,0,1]
; AVX2-SLOW-NEXT:    vpshufb %ymm5, %ymm2, %ymm3
; AVX2-SLOW-NEXT:    vmovdqu %ymm2, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[2,2,3,3]
; AVX2-SLOW-NEXT:    vpor %ymm1, %ymm3, %ymm1
; AVX2-SLOW-NEXT:    vmovdqa 32(%rcx), %ymm7
; AVX2-SLOW-NEXT:    vbroadcasti128 {{.*#+}} ymm3 = [128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128]
; AVX2-SLOW-NEXT:    # ymm3 = mem[0,1,0,1]
; AVX2-SLOW-NEXT:    vpshufb %ymm3, %ymm7, %ymm4
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm9 = ymm4[2,2,3,3]
; AVX2-SLOW-NEXT:    vmovdqa 32(%rdx), %ymm13
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm4 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,25]
; AVX2-SLOW-NEXT:    vpshufb %ymm4, %ymm13, %ymm12
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm12 = ymm12[2,2,3,3]
; AVX2-SLOW-NEXT:    vpor %ymm9, %ymm12, %ymm9
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm12 = <255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0>
; AVX2-SLOW-NEXT:    vpblendvb %ymm12, %ymm1, %ymm9, %ymm9
; AVX2-SLOW-NEXT:    vmovdqa (%rdi), %ymm6
; AVX2-SLOW-NEXT:    vpshufb %ymm15, %ymm6, %ymm1
; AVX2-SLOW-NEXT:    vmovdqu %ymm6, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX2-SLOW-NEXT:    vmovdqa (%rsi), %ymm15
; AVX2-SLOW-NEXT:    vpshufb %ymm5, %ymm15, %ymm5
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX2-SLOW-NEXT:    vpor %ymm1, %ymm5, %ymm5
; AVX2-SLOW-NEXT:    vmovdqa (%rcx), %ymm10
; AVX2-SLOW-NEXT:    vpshufb %ymm3, %ymm10, %ymm3
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm3[2,2,3,3]
; AVX2-SLOW-NEXT:    vmovdqa (%rdx), %ymm8
; AVX2-SLOW-NEXT:    vpshufb %ymm4, %ymm8, %ymm4
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,2,3,3]
; AVX2-SLOW-NEXT:    vpor %ymm0, %ymm4, %ymm0
; AVX2-SLOW-NEXT:    vpblendvb %ymm12, %ymm5, %ymm0, %ymm0
; AVX2-SLOW-NEXT:    vmovdqa 32(%r8), %ymm12
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm4 = ymm12[0,2,1,1,4,6,5,5]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,3,3,2]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm5 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm5, %ymm9, %ymm4, %ymm1
; AVX2-SLOW-NEXT:    vmovdqu %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa (%r8), %ymm9
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm4 = ymm9[0,2,1,1,4,6,5,5]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,3,3,2]
; AVX2-SLOW-NEXT:    vpblendvb %ymm5, %ymm0, %ymm4, %ymm0
; AVX2-SLOW-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm3 = <3,3,3,u,4,4,4,4>
; AVX2-SLOW-NEXT:    vpermd %ymm11, %ymm3, %ymm4
; AVX2-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm5 = [0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14]
; AVX2-SLOW-NEXT:    vpshufb %ymm5, %ymm2, %ymm0
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm1 = <u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm1, %ymm4, %ymm0, %ymm0
; AVX2-SLOW-NEXT:    vpermd %ymm6, %ymm3, %ymm2
; AVX2-SLOW-NEXT:    vpshufb %ymm5, %ymm15, %ymm4
; AVX2-SLOW-NEXT:    vpblendvb %ymm1, %ymm2, %ymm4, %ymm1
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm2 = [128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128]
; AVX2-SLOW-NEXT:    vmovdqa %ymm7, %ymm3
; AVX2-SLOW-NEXT:    vpshufb %ymm2, %ymm7, %ymm4
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm5 = [128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128,128]
; AVX2-SLOW-NEXT:    vpshufb %ymm5, %ymm13, %ymm11
; AVX2-SLOW-NEXT:    vpor %ymm4, %ymm11, %ymm4
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm11 = <u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm11, %ymm0, %ymm4, %ymm0
; AVX2-SLOW-NEXT:    vpshufb %ymm2, %ymm10, %ymm2
; AVX2-SLOW-NEXT:    vpshufb %ymm5, %ymm8, %ymm4
; AVX2-SLOW-NEXT:    vpor %ymm2, %ymm4, %ymm2
; AVX2-SLOW-NEXT:    vpblendvb %ymm11, %ymm1, %ymm2, %ymm1
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm2 = <3,3,3,3,u,4,4,4>
; AVX2-SLOW-NEXT:    vpermd %ymm12, %ymm2, %ymm4
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm5 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm5, %ymm0, %ymm4, %ymm0
; AVX2-SLOW-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vpermd %ymm9, %ymm2, %ymm0
; AVX2-SLOW-NEXT:    vpblendvb %ymm5, %ymm1, %ymm0, %ymm0
; AVX2-SLOW-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-SLOW-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX2-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm0 = xmm0[0],xmm14[0],xmm0[1],xmm14[1],xmm0[2],xmm14[2],xmm0[3],xmm14[3],xmm0[4],xmm14[4],xmm0[5],xmm14[5],xmm0[6],xmm14[6],xmm0[7],xmm14[7]
; AVX2-SLOW-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
; AVX2-SLOW-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm1 # 16-byte Folded Reload
; AVX2-SLOW-NEXT:    # xmm1 = xmm1[0],mem[0],xmm1[1],mem[1],xmm1[2],mem[2],xmm1[3],mem[3],xmm1[4],mem[4],xmm1[5],mem[5],xmm1[6],mem[6],xmm1[7],mem[7]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} xmm14 = <0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13>
; AVX2-SLOW-NEXT:    vpshufb %xmm14, %xmm0, %xmm0
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,1,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} xmm7 = <2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8>
; AVX2-SLOW-NEXT:    vpshufb %xmm7, %xmm1, %xmm1
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm11 = <255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255>
; AVX2-SLOW-NEXT:    vpblendvb %ymm11, %ymm0, %ymm1, %ymm6
; AVX2-SLOW-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm0 # 16-byte Reload
; AVX2-SLOW-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm0, %xmm0 # 16-byte Folded Reload
; AVX2-SLOW-NEXT:    # xmm0 = xmm0[0],mem[0],xmm0[1],mem[1],xmm0[2],mem[2],xmm0[3],mem[3],xmm0[4],mem[4],xmm0[5],mem[5],xmm0[6],mem[6],xmm0[7],mem[7]
; AVX2-SLOW-NEXT:    vpshufb %xmm14, %xmm0, %xmm0
; AVX2-SLOW-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Reload
; AVX2-SLOW-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm1, %xmm1 # 16-byte Folded Reload
; AVX2-SLOW-NEXT:    # xmm1 = xmm1[0],mem[0],xmm1[1],mem[1],xmm1[2],mem[2],xmm1[3],mem[3],xmm1[4],mem[4],xmm1[5],mem[5],xmm1[6],mem[6],xmm1[7],mem[7]
; AVX2-SLOW-NEXT:    vpshufb %xmm7, %xmm1, %xmm1
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,1,1]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,1]
; AVX2-SLOW-NEXT:    vpblendvb %ymm11, %ymm0, %ymm1, %ymm0
; AVX2-SLOW-NEXT:    vpshufd $80, (%rsp), %xmm1 # 16-byte Folded Reload
; AVX2-SLOW-NEXT:    # xmm1 = mem[0,0,1,1]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,0,1]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm2 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-SLOW-NEXT:    vpblendvb %ymm2, %ymm6, %ymm1, %ymm6
; AVX2-SLOW-NEXT:    vpshufd $80, {{[-0-9]+}}(%r{{[sb]}}p), %xmm1 # 16-byte Folded Reload
; AVX2-SLOW-NEXT:    # xmm1 = mem[0,0,1,1]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,0,1]
; AVX2-SLOW-NEXT:    vpblendvb %ymm2, %ymm0, %ymm1, %ymm7
; AVX2-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm0 = [9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12]
; AVX2-SLOW-NEXT:    vpshufb %ymm0, %ymm3, %ymm1
; AVX2-SLOW-NEXT:    vpshufhw {{.*#+}} ymm2 = ymm13[0,1,2,3,5,6,7,6,8,9,10,11,13,14,15,14]
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm2 = ymm2[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vbroadcasti128 {{.*#+}} ymm4 = [255,0,255,0,0,255,0,255,0,255,0,0,255,0,255,0,255,0,255,0,0,255,0,255,0,255,0,0,255,0,255,0]
; AVX2-SLOW-NEXT:    # ymm4 = mem[0,1,0,1]
; AVX2-SLOW-NEXT:    vpblendvb %ymm4, %ymm1, %ymm2, %ymm1
; AVX2-SLOW-NEXT:    vpshufb %ymm0, %ymm10, %ymm0
; AVX2-SLOW-NEXT:    vpshufhw {{.*#+}} ymm2 = ymm8[0,1,2,3,5,6,7,6,8,9,10,11,13,14,15,14]
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm2 = ymm2[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vpblendvb %ymm4, %ymm0, %ymm2, %ymm0
; AVX2-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm2 = [11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14]
; AVX2-SLOW-NEXT:    vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm3 # 32-byte Reload
; AVX2-SLOW-NEXT:    vpshufb %ymm2, %ymm3, %ymm3
; AVX2-SLOW-NEXT:    vpshufhw $230, {{[-0-9]+}}(%r{{[sb]}}p), %ymm4 # 32-byte Folded Reload
; AVX2-SLOW-NEXT:    # ymm4 = mem[0,1,2,3,6,5,6,7,8,9,10,11,14,13,14,15]
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm4 = ymm4[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vbroadcasti128 {{.*#+}} ymm5 = [255,0,0,255,0,255,0,0,0,0,255,0,255,0,0,255,255,0,0,255,0,255,0,0,0,0,255,0,255,0,0,255]
; AVX2-SLOW-NEXT:    # ymm5 = mem[0,1,0,1]
; AVX2-SLOW-NEXT:    vpblendvb %ymm5, %ymm3, %ymm4, %ymm3
; AVX2-SLOW-NEXT:    vpshufb %ymm2, %ymm15, %ymm2
; AVX2-SLOW-NEXT:    vpshufhw $230, {{[-0-9]+}}(%r{{[sb]}}p), %ymm4 # 32-byte Folded Reload
; AVX2-SLOW-NEXT:    # ymm4 = mem[0,1,2,3,6,5,6,7,8,9,10,11,14,13,14,15]
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm4 = ymm4[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vpblendvb %ymm5, %ymm2, %ymm4, %ymm2
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[2,2,3,3]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm4 = <255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u>
; AVX2-SLOW-NEXT:    vpblendvb %ymm4, %ymm1, %ymm3, %ymm1
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,2,3,3]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX2-SLOW-NEXT:    vpblendvb %ymm4, %ymm0, %ymm2, %ymm0
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm2 = ymm12[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX2-SLOW-NEXT:    vmovdqa {{.*#+}} ymm3 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; AVX2-SLOW-NEXT:    vpblendvb %ymm3, %ymm1, %ymm2, %ymm1
; AVX2-SLOW-NEXT:    vpshufd {{.*#+}} ymm2 = ymm9[2,2,3,3,6,6,7,7]
; AVX2-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX2-SLOW-NEXT:    vpblendvb %ymm3, %ymm0, %ymm2, %ymm0
; AVX2-SLOW-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm2 # 32-byte Reload
; AVX2-SLOW-NEXT:    vmovaps %ymm2, 64(%r9)
; AVX2-SLOW-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm2 # 32-byte Reload
; AVX2-SLOW-NEXT:    vmovaps %ymm2, 224(%r9)
; AVX2-SLOW-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm2 # 32-byte Reload
; AVX2-SLOW-NEXT:    vmovaps %ymm2, 96(%r9)
; AVX2-SLOW-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm2 # 32-byte Reload
; AVX2-SLOW-NEXT:    vmovaps %ymm2, 256(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm0, 128(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm7, 160(%r9)
; AVX2-SLOW-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-SLOW-NEXT:    vmovaps %ymm0, 192(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm1, 288(%r9)
; AVX2-SLOW-NEXT:    vmovdqa %ymm6, (%r9)
; AVX2-SLOW-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-SLOW-NEXT:    vmovaps %ymm0, 32(%r9)
; AVX2-SLOW-NEXT:    addq $312, %rsp # imm = 0x138
; AVX2-SLOW-NEXT:    vzeroupper
; AVX2-SLOW-NEXT:    retq
;
; AVX2-FAST-LABEL: store_i8_stride5_vf64:
; AVX2-FAST:       # %bb.0:
; AVX2-FAST-NEXT:    subq $168, %rsp
; AVX2-FAST-NEXT:    vmovdqa 32(%rdx), %ymm13
; AVX2-FAST-NEXT:    vmovdqa 32(%rcx), %ymm10
; AVX2-FAST-NEXT:    vmovdqa (%r8), %ymm11
; AVX2-FAST-NEXT:    vmovdqa (%rcx), %xmm1
; AVX2-FAST-NEXT:    vmovdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-NEXT:    vmovdqa 32(%rcx), %xmm6
; AVX2-FAST-NEXT:    vmovdqa %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm0 = <128,6,128,8,u,128,7,128,9,128,11,u,128,10,128,12>
; AVX2-FAST-NEXT:    vpshufb %xmm0, %xmm1, %xmm1
; AVX2-FAST-NEXT:    vmovdqa (%rdx), %xmm3
; AVX2-FAST-NEXT:    vmovdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-NEXT:    vmovdqa 32(%rdx), %xmm8
; AVX2-FAST-NEXT:    vmovdqa %xmm8, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm2 = <6,128,8,128,u,7,128,9,128,11,128,u,10,128,12,128>
; AVX2-FAST-NEXT:    vpshufb %xmm2, %xmm3, %xmm3
; AVX2-FAST-NEXT:    vpor %xmm1, %xmm3, %xmm1
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,1]
; AVX2-FAST-NEXT:    vmovdqa (%rdi), %xmm4
; AVX2-FAST-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm3 = <8,128,u,7,128,9,128,u,128,u,10,128,12,128,u,11>
; AVX2-FAST-NEXT:    vpshufb %xmm3, %xmm4, %xmm4
; AVX2-FAST-NEXT:    vmovdqa (%rsi), %xmm7
; AVX2-FAST-NEXT:    vmovdqa %xmm7, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm5 = <128,8,u,128,7,128,9,u,11,u,128,10,128,12,u,128>
; AVX2-FAST-NEXT:    vpshufb %xmm5, %xmm7, %xmm7
; AVX2-FAST-NEXT:    vpor %xmm4, %xmm7, %xmm4
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,1,1]
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm7 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255>
; AVX2-FAST-NEXT:    vpblendvb %ymm7, %ymm1, %ymm4, %ymm1
; AVX2-FAST-NEXT:    vmovdqa 32(%rdi), %xmm4
; AVX2-FAST-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-NEXT:    vpshufb %xmm0, %xmm6, %xmm0
; AVX2-FAST-NEXT:    vpshufb %xmm2, %xmm8, %xmm2
; AVX2-FAST-NEXT:    vpor %xmm0, %xmm2, %xmm0
; AVX2-FAST-NEXT:    vmovdqa 32(%rsi), %xmm6
; AVX2-FAST-NEXT:    vmovdqa %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-NEXT:    vpshufb %xmm3, %xmm4, %xmm2
; AVX2-FAST-NEXT:    vpshufb %xmm5, %xmm6, %xmm3
; AVX2-FAST-NEXT:    vpor %xmm2, %xmm3, %xmm2
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm3 = [1,1,2,2,2,2,2,2]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,1,1]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[0,0,1,1]
; AVX2-FAST-NEXT:    vpblendvb %ymm7, %ymm0, %ymm2, %ymm0
; AVX2-FAST-NEXT:    vpermd %ymm11, %ymm3, %ymm2
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm4, %ymm1, %ymm2, %ymm1
; AVX2-FAST-NEXT:    vmovdqu %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-NEXT:    vmovdqa 32(%r8), %ymm12
; AVX2-FAST-NEXT:    vpermd %ymm12, %ymm3, %ymm1
; AVX2-FAST-NEXT:    vpblendvb %ymm4, %ymm0, %ymm1, %ymm0
; AVX2-FAST-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm0 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,27,128,29,26,128,28,128,30,128,28,29,128,31,128,29]
; AVX2-FAST-NEXT:    vpshufb %ymm0, %ymm13, %ymm1
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX2-FAST-NEXT:    vbroadcasti128 {{.*#+}} ymm8 = [25,128,27,128,128,26,128,28,128,30,128,128,29,128,31,128,25,128,27,128,128,26,128,28,128,30,128,128,29,128,31,128]
; AVX2-FAST-NEXT:    # ymm8 = mem[0,1,0,1]
; AVX2-FAST-NEXT:    vpshufb %ymm8, %ymm10, %ymm2
; AVX2-FAST-NEXT:    vmovdqa %ymm10, %ymm14
; AVX2-FAST-NEXT:    vmovdqu %ymm10, (%rsp) # 32-byte Spill
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm1, %ymm2, %ymm1
; AVX2-FAST-NEXT:    vmovdqa 32(%rdi), %ymm3
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm9 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,29,26,128,28,128,26,27,28,29,128,31,128,29,30,128]
; AVX2-FAST-NEXT:    vpshufb %ymm9, %ymm3, %ymm2
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm4 = ymm2[2,2,3,3]
; AVX2-FAST-NEXT:    vmovdqa 32(%rsi), %ymm2
; AVX2-FAST-NEXT:    vbroadcasti128 {{.*#+}} ymm15 = [27,128,128,26,128,28,128,128,128,128,29,128,31,128,128,30,27,128,128,26,128,28,128,128,128,128,29,128,31,128,128,30]
; AVX2-FAST-NEXT:    # ymm15 = mem[0,1,0,1]
; AVX2-FAST-NEXT:    vpshufb %ymm15, %ymm2, %ymm5
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm4, %ymm5, %ymm4
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm7 = <255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u>
; AVX2-FAST-NEXT:    vpblendvb %ymm7, %ymm1, %ymm4, %ymm4
; AVX2-FAST-NEXT:    vmovdqa (%rdx), %ymm10
; AVX2-FAST-NEXT:    vpshufb %ymm0, %ymm10, %ymm0
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,2,3,3]
; AVX2-FAST-NEXT:    vmovdqa (%rcx), %ymm5
; AVX2-FAST-NEXT:    vpshufb %ymm8, %ymm5, %ymm1
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm0, %ymm1, %ymm8
; AVX2-FAST-NEXT:    vmovdqa (%rdi), %ymm1
; AVX2-FAST-NEXT:    vpshufb %ymm9, %ymm1, %ymm0
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm9 = ymm0[2,2,3,3]
; AVX2-FAST-NEXT:    vmovdqa (%rsi), %ymm0
; AVX2-FAST-NEXT:    vpshufb %ymm15, %ymm0, %ymm15
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm15 = ymm15[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm9, %ymm15, %ymm9
; AVX2-FAST-NEXT:    vpblendvb %ymm7, %ymm8, %ymm9, %ymm7
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = [6,6,6,6,7,7,7,7]
; AVX2-FAST-NEXT:    vpermd %ymm12, %ymm8, %ymm9
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm15 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; AVX2-FAST-NEXT:    vpblendvb %ymm15, %ymm4, %ymm9, %ymm4
; AVX2-FAST-NEXT:    vmovdqu %ymm4, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-NEXT:    vpermd %ymm11, %ymm8, %ymm4
; AVX2-FAST-NEXT:    vpblendvb %ymm15, %ymm7, %ymm4, %ymm4
; AVX2-FAST-NEXT:    vmovdqu %ymm4, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm6 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,21,128,19,20,128,22,128,24,128,22,23,128,25,128,23]
; AVX2-FAST-NEXT:    vpshufb %ymm6, %ymm3, %ymm7
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX2-FAST-NEXT:    vbroadcasti128 {{.*#+}} ymm8 = [19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128]
; AVX2-FAST-NEXT:    # ymm8 = mem[0,1,0,1]
; AVX2-FAST-NEXT:    vpshufb %ymm8, %ymm2, %ymm9
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm7, %ymm9, %ymm7
; AVX2-FAST-NEXT:    vbroadcasti128 {{.*#+}} ymm9 = [128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128]
; AVX2-FAST-NEXT:    # ymm9 = mem[0,1,0,1]
; AVX2-FAST-NEXT:    vpshufb %ymm9, %ymm14, %ymm15
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm15 = ymm15[2,2,3,3]
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,25]
; AVX2-FAST-NEXT:    vpshufb %ymm4, %ymm13, %ymm14
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm14 = ymm14[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm15, %ymm14, %ymm14
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm15 = <255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0>
; AVX2-FAST-NEXT:    vpblendvb %ymm15, %ymm7, %ymm14, %ymm7
; AVX2-FAST-NEXT:    vpshufb %ymm6, %ymm1, %ymm6
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[2,2,3,3]
; AVX2-FAST-NEXT:    vpshufb %ymm8, %ymm0, %ymm8
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm6, %ymm8, %ymm6
; AVX2-FAST-NEXT:    vpshufb %ymm9, %ymm5, %ymm8
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-NEXT:    vpshufb %ymm4, %ymm10, %ymm4
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,2,3,3]
; AVX2-FAST-NEXT:    vpor %ymm4, %ymm8, %ymm4
; AVX2-FAST-NEXT:    vpblendvb %ymm15, %ymm6, %ymm4, %ymm4
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm6 = [4,6,5,5,5,5,4,6]
; AVX2-FAST-NEXT:    vpermd %ymm12, %ymm6, %ymm8
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm14 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm14, %ymm7, %ymm8, %ymm9
; AVX2-FAST-NEXT:    vpermd %ymm11, %ymm6, %ymm6
; AVX2-FAST-NEXT:    vpblendvb %ymm14, %ymm4, %ymm6, %ymm7
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = <3,3,3,u,4,4,4,4>
; AVX2-FAST-NEXT:    vpermd %ymm3, %ymm4, %ymm3
; AVX2-FAST-NEXT:    vpbroadcastq {{.*#+}} ymm6 = [0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14]
; AVX2-FAST-NEXT:    vpshufb %ymm6, %ymm2, %ymm2
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm8 = <u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255>
; AVX2-FAST-NEXT:    vpblendvb %ymm8, %ymm3, %ymm2, %ymm2
; AVX2-FAST-NEXT:    vpermd %ymm1, %ymm4, %ymm1
; AVX2-FAST-NEXT:    vpshufb %ymm6, %ymm0, %ymm0
; AVX2-FAST-NEXT:    vpblendvb %ymm8, %ymm1, %ymm0, %ymm0
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm1 = [128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128]
; AVX2-FAST-NEXT:    vmovdqu (%rsp), %ymm3 # 32-byte Reload
; AVX2-FAST-NEXT:    vpshufb %ymm1, %ymm3, %ymm3
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = [128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128,128]
; AVX2-FAST-NEXT:    vpshufb %ymm4, %ymm13, %ymm6
; AVX2-FAST-NEXT:    vpor %ymm3, %ymm6, %ymm3
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm6 = <u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255>
; AVX2-FAST-NEXT:    vpblendvb %ymm6, %ymm2, %ymm3, %ymm2
; AVX2-FAST-NEXT:    vpshufb %ymm1, %ymm5, %ymm1
; AVX2-FAST-NEXT:    vpshufb %ymm4, %ymm10, %ymm3
; AVX2-FAST-NEXT:    vpor %ymm1, %ymm3, %ymm1
; AVX2-FAST-NEXT:    vpblendvb %ymm6, %ymm0, %ymm1, %ymm1
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm3 = <3,3,3,3,u,4,4,4>
; AVX2-FAST-NEXT:    vpermd %ymm12, %ymm3, %ymm0
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm4, %ymm2, %ymm0, %ymm0
; AVX2-FAST-NEXT:    vpermd %ymm11, %ymm3, %ymm2
; AVX2-FAST-NEXT:    vpblendvb %ymm4, %ymm1, %ymm2, %ymm1
; AVX2-FAST-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 # 16-byte Reload
; AVX2-FAST-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm2, %xmm2 # 16-byte Folded Reload
; AVX2-FAST-NEXT:    # xmm2 = xmm2[0],mem[0],xmm2[1],mem[1],xmm2[2],mem[2],xmm2[3],mem[3],xmm2[4],mem[4],xmm2[5],mem[5],xmm2[6],mem[6],xmm2[7],mem[7]
; AVX2-FAST-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Reload
; AVX2-FAST-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm3, %xmm3 # 16-byte Folded Reload
; AVX2-FAST-NEXT:    # xmm3 = xmm3[0],mem[0],xmm3[1],mem[1],xmm3[2],mem[2],xmm3[3],mem[3],xmm3[4],mem[4],xmm3[5],mem[5],xmm3[6],mem[6],xmm3[7],mem[7]
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm4 = <0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13>
; AVX2-FAST-NEXT:    vpshufb %xmm4, %xmm2, %xmm2
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[0,0,1,1]
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} xmm5 = <2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8>
; AVX2-FAST-NEXT:    vpshufb %xmm5, %xmm3, %xmm3
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,1,1]
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm6 = <255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255>
; AVX2-FAST-NEXT:    vpblendvb %ymm6, %ymm2, %ymm3, %ymm2
; AVX2-FAST-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Reload
; AVX2-FAST-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm3, %xmm3 # 16-byte Folded Reload
; AVX2-FAST-NEXT:    # xmm3 = xmm3[0],mem[0],xmm3[1],mem[1],xmm3[2],mem[2],xmm3[3],mem[3],xmm3[4],mem[4],xmm3[5],mem[5],xmm3[6],mem[6],xmm3[7],mem[7]
; AVX2-FAST-NEXT:    vpshufb %xmm4, %xmm3, %xmm3
; AVX2-FAST-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm4 # 16-byte Reload
; AVX2-FAST-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm4, %xmm4 # 16-byte Folded Reload
; AVX2-FAST-NEXT:    # xmm4 = xmm4[0],mem[0],xmm4[1],mem[1],xmm4[2],mem[2],xmm4[3],mem[3],xmm4[4],mem[4],xmm4[5],mem[5],xmm4[6],mem[6],xmm4[7],mem[7]
; AVX2-FAST-NEXT:    vpshufb %xmm5, %xmm4, %xmm4
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,1,1]
; AVX2-FAST-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,1,1]
; AVX2-FAST-NEXT:    vpblendvb %ymm6, %ymm3, %ymm4, %ymm3
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm4 = [0,0,0,0,0,0,1,1]
; AVX2-FAST-NEXT:    vpermd %ymm11, %ymm4, %ymm5
; AVX2-FAST-NEXT:    vmovdqa {{.*#+}} ymm6 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-FAST-NEXT:    vpblendvb %ymm6, %ymm2, %ymm5, %ymm2
; AVX2-FAST-NEXT:    vpermd %ymm12, %ymm4, %ymm4
; AVX2-FAST-NEXT:    vpblendvb %ymm6, %ymm3, %ymm4, %ymm3
; AVX2-FAST-NEXT:    vmovdqa %ymm1, 64(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm0, 224(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm7, 96(%r9)
; AVX2-FAST-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-FAST-NEXT:    vmovaps %ymm0, 128(%r9)
; AVX2-FAST-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-FAST-NEXT:    vmovaps %ymm0, 288(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm9, 256(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm3, 160(%r9)
; AVX2-FAST-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-FAST-NEXT:    vmovaps %ymm0, 192(%r9)
; AVX2-FAST-NEXT:    vmovdqa %ymm2, (%r9)
; AVX2-FAST-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-FAST-NEXT:    vmovaps %ymm0, 32(%r9)
; AVX2-FAST-NEXT:    addq $168, %rsp
; AVX2-FAST-NEXT:    vzeroupper
; AVX2-FAST-NEXT:    retq
;
; AVX2-FAST-PERLANE-LABEL: store_i8_stride5_vf64:
; AVX2-FAST-PERLANE:       # %bb.0:
; AVX2-FAST-PERLANE-NEXT:    subq $200, %rsp
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%rdx), %ymm11
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rcx), %xmm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm1, (%rsp) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%rcx), %xmm8
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm8, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} xmm0 = <128,6,128,8,u,128,7,128,9,128,11,u,128,10,128,12>
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm0, %xmm1, %xmm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdx), %xmm3
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm3, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%rdx), %xmm9
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm9, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} xmm2 = <6,128,8,128,u,7,128,9,128,11,128,u,10,128,12,128>
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm2, %xmm3, %xmm3
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm1, %xmm3, %xmm1
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdi), %xmm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} xmm3 = <8,128,u,7,128,9,128,u,128,u,10,128,12,128,u,11>
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm3, %xmm4, %xmm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rsi), %xmm6
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm6, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} xmm5 = <128,8,u,128,7,128,9,u,11,u,128,10,128,12,u,128>
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm5, %xmm6, %xmm6
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm4, %xmm6, %xmm4
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm6 = <255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm6, %ymm1, %ymm4, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%r8), %xmm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} xmm4 = xmm4[1,1,2,2]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,1,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm7 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm7, %ymm1, %ymm4, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqu %ymm1, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%rdi), %xmm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm4, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm0, %xmm8, %xmm0
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm2, %xmm9, %xmm1
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%rsi), %xmm2
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm2, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm3, %xmm4, %xmm1
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm5, %xmm2, %xmm2
; AVX2-FAST-PERLANE-NEXT:    vpor %xmm1, %xmm2, %xmm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%rcx), %ymm3
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm6, %ymm0, %ymm1, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%r8), %xmm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %xmm1, {{[-0-9]+}}(%r{{[sb]}}p) # 16-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} xmm1 = xmm1[1,1,2,2]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[0,1,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm7, %ymm0, %ymm1, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm0 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,27,128,29,26,128,28,128,30,128,28,29,128,31,128,29]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm0, %ymm11, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vbroadcasti128 {{.*#+}} ymm8 = [25,128,27,128,128,26,128,28,128,30,128,128,29,128,31,128,25,128,27,128,128,26,128,28,128,30,128,128,29,128,31,128]
; AVX2-FAST-PERLANE-NEXT:    # ymm8 = mem[0,1,0,1]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm8, %ymm3, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm3, %ymm12
; AVX2-FAST-PERLANE-NEXT:    vmovdqu %ymm3, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm1, %ymm2, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%rdi), %ymm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm5 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,29,26,128,28,128,26,27,28,29,128,31,128,29,30,128]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm5, %ymm4, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm6 = ymm2[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%rsi), %ymm2
; AVX2-FAST-PERLANE-NEXT:    vbroadcasti128 {{.*#+}} ymm9 = [27,128,128,26,128,28,128,128,128,128,29,128,31,128,128,30,27,128,128,26,128,28,128,128,128,128,29,128,31,128,128,30]
; AVX2-FAST-PERLANE-NEXT:    # ymm9 = mem[0,1,0,1]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm9, %ymm2, %ymm7
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm6, %ymm7, %ymm6
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm10 = <255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u,0,0,255,255,u>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm10, %ymm1, %ymm6, %ymm3
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdx), %ymm13
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm0, %ymm13, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rcx), %ymm7
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm8, %ymm7, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm0, %ymm1, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rdi), %ymm1
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm5, %ymm1, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm5 = ymm0[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rsi), %ymm0
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm9, %ymm0, %ymm9
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm5, %ymm9, %ymm5
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm10, %ymm8, %ymm5, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vmovdqa 32(%r8), %ymm5
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} ymm9 = ymm5[2,2,3,3,6,6,7,7]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm10 = [255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm10, %ymm3, %ymm9, %ymm3
; AVX2-FAST-PERLANE-NEXT:    vmovdqu %ymm3, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%r8), %ymm3
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} ymm9 = ymm3[2,2,3,3,6,6,7,7]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm10, %ymm8, %ymm9, %ymm6
; AVX2-FAST-PERLANE-NEXT:    vmovdqu %ymm6, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm8 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,21,128,19,20,128,22,128,24,128,22,23,128,25,128,23]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm8, %ymm4, %ymm9
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vbroadcasti128 {{.*#+}} ymm10 = [19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128]
; AVX2-FAST-PERLANE-NEXT:    # ymm10 = mem[0,1,0,1]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm10, %ymm2, %ymm15
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm15 = ymm15[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm9, %ymm15, %ymm9
; AVX2-FAST-PERLANE-NEXT:    vbroadcasti128 {{.*#+}} ymm15 = [128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128]
; AVX2-FAST-PERLANE-NEXT:    # ymm15 = mem[0,1,0,1]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm15, %ymm12, %ymm14
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm14 = ymm14[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm6 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,25]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm6, %ymm11, %ymm12
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm12 = ymm12[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm14, %ymm12, %ymm12
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm14 = <255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm14, %ymm9, %ymm12, %ymm9
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm8, %ymm1, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm10, %ymm0, %ymm10
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm8, %ymm10, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm15, %ymm7, %ymm10
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm6, %ymm13, %ymm6
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[2,2,3,3]
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm6, %ymm10, %ymm6
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm14, %ymm8, %ymm6, %ymm6
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} ymm8 = ymm5[0,2,1,1,4,6,5,5]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,3,3,2]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm12 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm12, %ymm9, %ymm8, %ymm10
; AVX2-FAST-PERLANE-NEXT:    vpshufd {{.*#+}} ymm8 = ymm3[0,2,1,1,4,6,5,5]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,3,3,2]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm12, %ymm6, %ymm8, %ymm9
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm6 = <3,3,3,u,4,4,4,4>
; AVX2-FAST-PERLANE-NEXT:    vpermd %ymm4, %ymm6, %ymm4
; AVX2-FAST-PERLANE-NEXT:    vpbroadcastq {{.*#+}} ymm8 = [0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm8, %ymm2, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm12 = <u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255,0,u,u,u,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm12, %ymm4, %ymm2, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vpermd %ymm1, %ymm6, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm8, %ymm0, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm12, %ymm1, %ymm0, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm1 = [128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128]
; AVX2-FAST-PERLANE-NEXT:    vmovdqu {{[-0-9]+}}(%r{{[sb]}}p), %ymm4 # 32-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm1, %ymm4, %ymm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm6 = [128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128,128]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm6, %ymm11, %ymm8
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm4, %ymm8, %ymm4
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm8 = <u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm8, %ymm2, %ymm4, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm1, %ymm7, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vpshufb %ymm6, %ymm13, %ymm4
; AVX2-FAST-PERLANE-NEXT:    vpor %ymm1, %ymm4, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm8, %ymm0, %ymm1, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm4 = <3,3,3,3,u,4,4,4>
; AVX2-FAST-PERLANE-NEXT:    vpermd %ymm5, %ymm4, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm5 = [0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm5, %ymm2, %ymm0, %ymm0
; AVX2-FAST-PERLANE-NEXT:    vpermd %ymm3, %ymm4, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm5, %ymm1, %ymm2, %ymm1
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm2 # 16-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm2, %xmm2 # 16-byte Folded Reload
; AVX2-FAST-PERLANE-NEXT:    # xmm2 = xmm2[0],mem[0],xmm2[1],mem[1],xmm2[2],mem[2],xmm2[3],mem[3],xmm2[4],mem[4],xmm2[5],mem[5],xmm2[6],mem[6],xmm2[7],mem[7]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa (%rsp), %xmm3 # 16-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm3, %xmm3 # 16-byte Folded Reload
; AVX2-FAST-PERLANE-NEXT:    # xmm3 = xmm3[0],mem[0],xmm3[1],mem[1],xmm3[2],mem[2],xmm3[3],mem[3],xmm3[4],mem[4],xmm3[5],mem[5],xmm3[6],mem[6],xmm3[7],mem[7]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} xmm4 = <0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13>
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm4, %xmm2, %xmm2
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} xmm5 = <2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8>
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm5, %xmm3, %xmm3
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm6 = <255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255,0,0,u,255,255>
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm6, %ymm2, %ymm3, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm3 # 16-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm3, %xmm3 # 16-byte Folded Reload
; AVX2-FAST-PERLANE-NEXT:    # xmm3 = xmm3[0],mem[0],xmm3[1],mem[1],xmm3[2],mem[2],xmm3[3],mem[3],xmm3[4],mem[4],xmm3[5],mem[5],xmm3[6],mem[6],xmm3[7],mem[7]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm4, %xmm3, %xmm3
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{[-0-9]+}}(%r{{[sb]}}p), %xmm4 # 16-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vpunpcklbw {{[-0-9]+}}(%r{{[sb]}}p), %xmm4, %xmm4 # 16-byte Folded Reload
; AVX2-FAST-PERLANE-NEXT:    # xmm4 = xmm4[0],mem[0],xmm4[1],mem[1],xmm4[2],mem[2],xmm4[3],mem[3],xmm4[4],mem[4],xmm4[5],mem[5],xmm4[6],mem[6],xmm4[7],mem[7]
; AVX2-FAST-PERLANE-NEXT:    vpshufb %xmm5, %xmm4, %xmm4
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm6, %ymm3, %ymm4, %ymm3
; AVX2-FAST-PERLANE-NEXT:    vpshufd $80, {{[-0-9]+}}(%r{{[sb]}}p), %xmm4 # 16-byte Folded Reload
; AVX2-FAST-PERLANE-NEXT:    # xmm4 = mem[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,0,1]
; AVX2-FAST-PERLANE-NEXT:    vmovdqa {{.*#+}} ymm5 = [255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm5, %ymm2, %ymm4, %ymm2
; AVX2-FAST-PERLANE-NEXT:    vpshufd $80, {{[-0-9]+}}(%r{{[sb]}}p), %xmm4 # 16-byte Folded Reload
; AVX2-FAST-PERLANE-NEXT:    # xmm4 = mem[0,0,1,1]
; AVX2-FAST-PERLANE-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,0,1]
; AVX2-FAST-PERLANE-NEXT:    vpblendvb %ymm5, %ymm3, %ymm4, %ymm3
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm1, 64(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm0, 224(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm9, 96(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vmovaps %ymm0, 128(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vmovaps %ymm0, 288(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm10, 256(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm3, 160(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vmovaps %ymm0, 192(%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovdqa %ymm2, (%r9)
; AVX2-FAST-PERLANE-NEXT:    vmovups {{[-0-9]+}}(%r{{[sb]}}p), %ymm0 # 32-byte Reload
; AVX2-FAST-PERLANE-NEXT:    vmovaps %ymm0, 32(%r9)
; AVX2-FAST-PERLANE-NEXT:    addq $200, %rsp
; AVX2-FAST-PERLANE-NEXT:    vzeroupper
; AVX2-FAST-PERLANE-NEXT:    retq
;
; AVX512F-SLOW-LABEL: store_i8_stride5_vf64:
; AVX512F-SLOW:       # %bb.0:
; AVX512F-SLOW-NEXT:    vmovdqa 32(%rsi), %ymm3
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm15 = [128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128,128,128]
; AVX512F-SLOW-NEXT:    vpshufb %ymm15, %ymm3, %ymm0
; AVX512F-SLOW-NEXT:    vmovdqa 32(%rdi), %ymm2
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm7 = <12,13,128,15,12,13,14,128,12,13,14,15,128,u,u,u,16,128,18,19,16,17,128,19,16,17,18,128,16,17,18,19>
; AVX512F-SLOW-NEXT:    vpshufb %ymm7, %ymm2, %ymm1
; AVX512F-SLOW-NEXT:    vpor %ymm0, %ymm1, %ymm0
; AVX512F-SLOW-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX512F-SLOW-NEXT:    vmovdqa 32(%rdi), %xmm1
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} xmm8 = <8,128,u,7,128,9,128,u,128,u,10,128,12,128,u,11>
; AVX512F-SLOW-NEXT:    vpshufb %xmm8, %xmm1, %xmm0
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm1, %xmm29
; AVX512F-SLOW-NEXT:    vmovdqa 32(%rsi), %xmm4
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} xmm14 = <128,8,u,128,7,128,9,u,11,u,128,10,128,12,u,128>
; AVX512F-SLOW-NEXT:    vpshufb %xmm14, %xmm4, %xmm1
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm4, %xmm30
; AVX512F-SLOW-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX512F-SLOW-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX512F-SLOW-NEXT:    vmovdqa 32(%rcx), %ymm9
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm0 = [128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128]
; AVX512F-SLOW-NEXT:    vpshufb %ymm0, %ymm9, %ymm4
; AVX512F-SLOW-NEXT:    vmovdqa 32(%rdx), %ymm11
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm1 = <u,u,12,13,128,u,u,u,14,128,u,u,14,15,128,u,u,u,16,128,u,u,16,17,128,u,u,u,18,128,u,u>
; AVX512F-SLOW-NEXT:    vpshufb %ymm1, %ymm11, %ymm10
; AVX512F-SLOW-NEXT:    vpor %ymm4, %ymm10, %ymm4
; AVX512F-SLOW-NEXT:    vmovdqu %ymm4, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX512F-SLOW-NEXT:    vmovdqa 32(%rcx), %xmm13
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} xmm5 = <128,6,128,8,u,128,7,128,9,128,11,u,128,10,128,12>
; AVX512F-SLOW-NEXT:    vpshufb %xmm5, %xmm13, %xmm4
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm5, %xmm25
; AVX512F-SLOW-NEXT:    vmovdqa 32(%rdx), %xmm10
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} xmm5 = <6,128,8,128,u,7,128,9,128,11,128,u,10,128,12,128>
; AVX512F-SLOW-NEXT:    vpshufb %xmm5, %xmm10, %xmm12
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm5, %xmm26
; AVX512F-SLOW-NEXT:    vporq %xmm4, %xmm12, %xmm31
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm4 = ymm3[11,u,u,10,u,12,u,u,u,u,13,u,15,u,u,14,27,u,u,26,u,28,u,u,u,u,29,u,31,u,u,30]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm3 = ymm3[3,u,5,u,u,4,u,6,u,8,u,u,7,u,9,u,19,u,21,u,u,20,u,22,u,24,u,u,23,u,25,u]
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm4, %zmm3, %zmm22
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm3 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm2[26],zero,ymm2[28],zero,zero,ymm2[27],zero,ymm2[29],zero,ymm2[31],zero,zero,ymm2[30],zero
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm2 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm2[21],zero,zero,ymm2[20],zero,ymm2[22],zero,ymm2[24],zero,zero,ymm2[23],zero,ymm2[25],zero,zero
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm3, %zmm2, %zmm23
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm2 = ymm11[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm11[27],zero,zero,ymm11[26],zero,ymm11[28],zero,ymm11[30],zero,zero,ymm11[29],zero,ymm11[31],zero,zero
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm3 = ymm9[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm9[19],zero,ymm9[21],zero,zero,ymm9[20],zero,ymm9[22],zero,ymm9[24],zero,zero,ymm9[23],zero
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm2, %zmm3, %zmm24
; AVX512F-SLOW-NEXT:    vmovdqa (%rcx), %ymm12
; AVX512F-SLOW-NEXT:    vpshufb %ymm0, %ymm12, %ymm0
; AVX512F-SLOW-NEXT:    vmovdqa (%rdx), %ymm6
; AVX512F-SLOW-NEXT:    vpshufb %ymm1, %ymm6, %ymm1
; AVX512F-SLOW-NEXT:    vporq %ymm0, %ymm1, %ymm20
; AVX512F-SLOW-NEXT:    vmovdqa (%rsi), %ymm5
; AVX512F-SLOW-NEXT:    vpshufb %ymm15, %ymm5, %ymm2
; AVX512F-SLOW-NEXT:    vmovdqa (%rdi), %ymm4
; AVX512F-SLOW-NEXT:    vpshufb %ymm7, %ymm4, %ymm3
; AVX512F-SLOW-NEXT:    vporq %ymm2, %ymm3, %ymm21
; AVX512F-SLOW-NEXT:    vmovdqa (%rdi), %xmm0
; AVX512F-SLOW-NEXT:    vpshufb %xmm8, %xmm0, %xmm2
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm0, %xmm17
; AVX512F-SLOW-NEXT:    vmovdqa (%rsi), %xmm8
; AVX512F-SLOW-NEXT:    vpshufb %xmm14, %xmm8, %xmm3
; AVX512F-SLOW-NEXT:    vporq %xmm2, %xmm3, %xmm27
; AVX512F-SLOW-NEXT:    vmovdqa (%rcx), %xmm1
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm25, %xmm0
; AVX512F-SLOW-NEXT:    vpshufb %xmm0, %xmm1, %xmm0
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm1, %xmm19
; AVX512F-SLOW-NEXT:    vmovdqa (%rdx), %xmm2
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm26, %xmm1
; AVX512F-SLOW-NEXT:    vpshufb %xmm1, %xmm2, %xmm15
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm2, %xmm16
; AVX512F-SLOW-NEXT:    vporq %xmm0, %xmm15, %xmm28
; AVX512F-SLOW-NEXT:    vpshufd {{.*#+}} xmm0 = mem[1,1,2,2]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[0,1,1,1]
; AVX512F-SLOW-NEXT:    vmovdqa64 {{.*#+}} ymm25 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX512F-SLOW-NEXT:    vpandnq %ymm0, %ymm25, %ymm0
; AVX512F-SLOW-NEXT:    vmovdqa 32(%r8), %ymm15
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm1 = [12,128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128]
; AVX512F-SLOW-NEXT:    vpshufb %ymm1, %ymm15, %ymm14
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm14, %zmm0, %zmm26
; AVX512F-SLOW-NEXT:    vmovdqa (%r8), %ymm0
; AVX512F-SLOW-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
; AVX512F-SLOW-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[0,2,1,1,4,6,5,5]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,3,3,2]
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm14 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX512F-SLOW-NEXT:    vpandn %ymm0, %ymm14, %ymm0
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm0, %zmm1, %zmm25
; AVX512F-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm0 = [9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12]
; AVX512F-SLOW-NEXT:    vpshufb %ymm0, %ymm9, %ymm9
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm14 = ymm12[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm12[19],zero,ymm12[21],zero,zero,ymm12[20],zero,ymm12[22],zero,ymm12[24],zero,zero,ymm12[23],zero
; AVX512F-SLOW-NEXT:    vpshufb %ymm0, %ymm12, %ymm12
; AVX512F-SLOW-NEXT:    vbroadcasti128 {{.*#+}} ymm0 = [18,19,128,21,128,21,20,128,22,128,24,128,22,23,128,25,18,19,128,21,128,21,20,128,22,128,24,128,22,23,128,25]
; AVX512F-SLOW-NEXT:    # ymm0 = mem[0,1,0,1]
; AVX512F-SLOW-NEXT:    vpshufb %ymm0, %ymm11, %ymm11
; AVX512F-SLOW-NEXT:    vpshufb %ymm0, %ymm6, %ymm2
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm6 = ymm6[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm6[27],zero,zero,ymm6[26],zero,ymm6[28],zero,ymm6[30],zero,zero,ymm6[29],zero,ymm6[31],zero,zero
; AVX512F-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm10 = xmm13[0],xmm10[0],xmm13[1],xmm10[1],xmm13[2],xmm10[2],xmm13[3],xmm10[3],xmm13[4],xmm10[4],xmm13[5],xmm10[5],xmm13[6],xmm10[6],xmm13[7],xmm10[7]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm13 = ymm5[3,u,5,u,u,4,u,6,u,8,u,u,7,u,9,u,19,u,21,u,u,20,u,22,u,24,u,u,23,u,25,u]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm7 = ymm5[11,u,u,10,u,12,u,u,u,u,13,u,15,u,u,14,27,u,u,26,u,28,u,u,u,u,29,u,31,u,u,30]
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm0 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm4[21],zero,zero,ymm4[20],zero,ymm4[22],zero,ymm4[24],zero,zero,ymm4[23],zero,ymm4[25],zero,zero
; AVX512F-SLOW-NEXT:    vpshufb {{.*#+}} ymm5 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm4[26],zero,ymm4[28],zero,zero,ymm4[27],zero,ymm4[29],zero,ymm4[31],zero,zero,ymm4[30],zero
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm29, %xmm1
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm30, %xmm3
; AVX512F-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm3 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3],xmm1[4],xmm3[4],xmm1[5],xmm3[5],xmm1[6],xmm3[6],xmm1[7],xmm3[7]
; AVX512F-SLOW-NEXT:    vmovdqa64 (%r8), %zmm29
; AVX512F-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm1 = <4,u,5,5,5,5,u,6,6,6,6,u,7,7,7,7>
; AVX512F-SLOW-NEXT:    vpermd %zmm15, %zmm1, %zmm4
; AVX512F-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm1 = <6,6,6,u,7,7,7,7,u,16,16,16,16,u,17,17>
; AVX512F-SLOW-NEXT:    vpermi2d %zmm15, %zmm29, %zmm1
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm30 = ymm9[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm18 = ymm11[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm14 = ymm14[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm13 = ymm13[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm12 = ymm12[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[2,2,3,3]
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm17, %xmm9
; AVX512F-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm8 = xmm9[0],xmm8[0],xmm9[1],xmm8[1],xmm9[2],xmm8[2],xmm9[3],xmm8[3],xmm9[4],xmm8[4],xmm9[5],xmm8[5],xmm9[6],xmm8[6],xmm9[7],xmm8[7]
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} xmm15 = <0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13>
; AVX512F-SLOW-NEXT:    vpshufb %xmm15, %xmm3, %xmm3
; AVX512F-SLOW-NEXT:    vpshufb %xmm15, %xmm8, %xmm8
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} xmm15 = <2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8>
; AVX512F-SLOW-NEXT:    vpshufb %xmm15, %xmm10, %xmm10
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[0,0,1,1]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,1,1]
; AVX512F-SLOW-NEXT:    vinserti32x4 $2, %xmm27, %zmm8, %zmm8
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm19, %xmm9
; AVX512F-SLOW-NEXT:    vmovdqa64 %xmm16, %xmm11
; AVX512F-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm9 = xmm9[0],xmm11[0],xmm9[1],xmm11[1],xmm9[2],xmm11[2],xmm9[3],xmm11[3],xmm9[4],xmm11[4],xmm9[5],xmm11[5],xmm9[6],xmm11[6],xmm9[7],xmm11[7]
; AVX512F-SLOW-NEXT:    vpshufb %xmm15, %xmm9, %xmm9
; AVX512F-SLOW-NEXT:    vinserti32x4 $2, %xmm28, %zmm9, %zmm9
; AVX512F-SLOW-NEXT:    vpermq $80, {{[-0-9]+}}(%r{{[sb]}}p), %ymm15 # 32-byte Folded Reload
; AVX512F-SLOW-NEXT:    # ymm15 = mem[0,0,1,1]
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, {{[-0-9]+}}(%r{{[sb]}}p), %zmm15, %zmm15 # 32-byte Folded Reload
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} ymm16 = ymm31[0,0,1,1]
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, {{[-0-9]+}}(%r{{[sb]}}p), %zmm16, %zmm16 # 32-byte Folded Reload
; AVX512F-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm17 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0]
; AVX512F-SLOW-NEXT:    vpternlogq $226, %zmm15, %zmm17, %zmm16
; AVX512F-SLOW-NEXT:    vpor %ymm2, %ymm14, %ymm2
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm2, %zmm20, %zmm2
; AVX512F-SLOW-NEXT:    vmovdqa {{.*#+}} ymm14 = [18374966859431608575,18374966859431608575,18446463693966278400,18446463693966278400]
; AVX512F-SLOW-NEXT:    vpternlogq $248, %ymm14, %ymm13, %ymm0
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm0, %zmm21, %zmm0
; AVX512F-SLOW-NEXT:    vpternlogq $226, %zmm2, %zmm17, %zmm0
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} zmm2 = zmm22[2,2,3,3,6,6,7,7]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} zmm13 = zmm23[2,2,3,3,6,6,7,7]
; AVX512F-SLOW-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm13
; AVX512F-SLOW-NEXT:    vpternlogq $248, %ymm14, %ymm12, %ymm6
; AVX512F-SLOW-NEXT:    vpandq %ymm14, %ymm30, %ymm2
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm2, %zmm18, %zmm2
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} zmm11 = zmm24[2,2,3,3,6,6,7,7]
; AVX512F-SLOW-NEXT:    vporq %zmm11, %zmm2, %zmm2
; AVX512F-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm11 = [0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; AVX512F-SLOW-NEXT:    vpternlogq $226, %zmm13, %zmm11, %zmm2
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm10, %zmm6, %zmm6
; AVX512F-SLOW-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm7, %ymm5
; AVX512F-SLOW-NEXT:    vinserti64x4 $1, %ymm3, %zmm5, %zmm3
; AVX512F-SLOW-NEXT:    vpternlogq $226, %zmm6, %zmm11, %zmm3
; AVX512F-SLOW-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm16, %zmm26
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm4
; AVX512F-SLOW-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm25
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm1
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} zmm0 = zmm8[0,0,1,1,4,4,5,5]
; AVX512F-SLOW-NEXT:    vpermq {{.*#+}} zmm2 = zmm9[0,0,1,1,4,4,5,5]
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm0, %zmm2
; AVX512F-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm0 = <u,0,0,0,0,u,1,1,1,1,u,2,2,2,2,u>
; AVX512F-SLOW-NEXT:    vpermd %zmm29, %zmm0, %zmm0
; AVX512F-SLOW-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm0
; AVX512F-SLOW-NEXT:    vmovdqa64 %zmm25, 64(%r9)
; AVX512F-SLOW-NEXT:    vmovdqa64 %zmm0, (%r9)
; AVX512F-SLOW-NEXT:    vmovdqa64 %zmm1, 128(%r9)
; AVX512F-SLOW-NEXT:    vmovdqa64 %zmm4, 256(%r9)
; AVX512F-SLOW-NEXT:    vmovdqa64 %zmm26, 192(%r9)
; AVX512F-SLOW-NEXT:    vzeroupper
; AVX512F-SLOW-NEXT:    retq
;
; AVX512F-FAST-LABEL: store_i8_stride5_vf64:
; AVX512F-FAST:       # %bb.0:
; AVX512F-FAST-NEXT:    vmovdqa 32(%rsi), %ymm6
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm13 = [128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128,128,128]
; AVX512F-FAST-NEXT:    vpshufb %ymm13, %ymm6, %ymm0
; AVX512F-FAST-NEXT:    vmovdqa 32(%rdi), %ymm2
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm7 = <12,13,128,15,12,13,14,128,12,13,14,15,128,u,u,u,16,128,18,19,16,17,128,19,16,17,18,128,16,17,18,19>
; AVX512F-FAST-NEXT:    vpshufb %ymm7, %ymm2, %ymm1
; AVX512F-FAST-NEXT:    vpor %ymm0, %ymm1, %ymm0
; AVX512F-FAST-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX512F-FAST-NEXT:    vmovdqa 32(%rdi), %xmm1
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} xmm15 = <8,128,u,7,128,9,128,u,128,u,10,128,12,128,u,11>
; AVX512F-FAST-NEXT:    vpshufb %xmm15, %xmm1, %xmm0
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm1, %xmm18
; AVX512F-FAST-NEXT:    vmovdqa 32(%rsi), %xmm3
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} xmm4 = <128,8,u,128,7,128,9,u,11,u,128,10,128,12,u,128>
; AVX512F-FAST-NEXT:    vpshufb %xmm4, %xmm3, %xmm1
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm4, %xmm25
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm3, %xmm17
; AVX512F-FAST-NEXT:    vpor %xmm0, %xmm1, %xmm0
; AVX512F-FAST-NEXT:    vmovdqu %ymm0, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX512F-FAST-NEXT:    vmovdqa 32(%rcx), %ymm9
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm0 = [128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128]
; AVX512F-FAST-NEXT:    vpshufb %ymm0, %ymm9, %ymm3
; AVX512F-FAST-NEXT:    vmovdqa 32(%rdx), %ymm8
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm1 = <u,u,12,13,128,u,u,u,14,128,u,u,14,15,128,u,u,u,16,128,u,u,16,17,128,u,u,u,18,128,u,u>
; AVX512F-FAST-NEXT:    vpshufb %ymm1, %ymm8, %ymm4
; AVX512F-FAST-NEXT:    vpor %ymm3, %ymm4, %ymm3
; AVX512F-FAST-NEXT:    vmovdqu %ymm3, {{[-0-9]+}}(%r{{[sb]}}p) # 32-byte Spill
; AVX512F-FAST-NEXT:    vmovdqa 32(%rcx), %xmm10
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} xmm4 = <128,6,128,8,u,128,7,128,9,128,11,u,128,10,128,12>
; AVX512F-FAST-NEXT:    vpshufb %xmm4, %xmm10, %xmm3
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm4, %xmm26
; AVX512F-FAST-NEXT:    vmovdqa 32(%rdx), %xmm11
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} xmm14 = <6,128,8,128,u,7,128,9,128,11,128,u,10,128,12,128>
; AVX512F-FAST-NEXT:    vpshufb %xmm14, %xmm11, %xmm4
; AVX512F-FAST-NEXT:    vporq %xmm3, %xmm4, %xmm19
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm3 = ymm6[11,u,u,10,u,12,u,u,u,u,13,u,15,u,u,14,27,u,u,26,u,28,u,u,u,u,29,u,31,u,u,30]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm4 = ymm6[3,u,5,u,u,4,u,6,u,8,u,u,7,u,9,u,19,u,21,u,u,20,u,22,u,24,u,u,23,u,25,u]
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm3, %zmm4, %zmm22
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm3 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm2[26],zero,ymm2[28],zero,zero,ymm2[27],zero,ymm2[29],zero,ymm2[31],zero,zero,ymm2[30],zero
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = ymm2[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm2[21],zero,zero,ymm2[20],zero,ymm2[22],zero,ymm2[24],zero,zero,ymm2[23],zero,ymm2[25],zero,zero
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm3, %zmm2, %zmm23
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = ymm8[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm8[27],zero,zero,ymm8[26],zero,ymm8[28],zero,ymm8[30],zero,zero,ymm8[29],zero,ymm8[31],zero,zero
; AVX512F-FAST-NEXT:    vbroadcasti128 {{.*#+}} ymm4 = [128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128]
; AVX512F-FAST-NEXT:    # ymm4 = mem[0,1,0,1]
; AVX512F-FAST-NEXT:    vpshufb %ymm4, %ymm9, %ymm3
; AVX512F-FAST-NEXT:    vmovdqa64 %ymm4, %ymm30
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm2, %zmm3, %zmm24
; AVX512F-FAST-NEXT:    vmovdqa (%rcx), %ymm12
; AVX512F-FAST-NEXT:    vpshufb %ymm0, %ymm12, %ymm0
; AVX512F-FAST-NEXT:    vmovdqa (%rdx), %ymm6
; AVX512F-FAST-NEXT:    vpshufb %ymm1, %ymm6, %ymm1
; AVX512F-FAST-NEXT:    vporq %ymm0, %ymm1, %ymm20
; AVX512F-FAST-NEXT:    vmovdqa (%rsi), %ymm5
; AVX512F-FAST-NEXT:    vpshufb %ymm13, %ymm5, %ymm0
; AVX512F-FAST-NEXT:    vmovdqa (%rdi), %ymm4
; AVX512F-FAST-NEXT:    vpshufb %ymm7, %ymm4, %ymm1
; AVX512F-FAST-NEXT:    vporq %ymm0, %ymm1, %ymm21
; AVX512F-FAST-NEXT:    vmovdqa (%rdi), %xmm1
; AVX512F-FAST-NEXT:    vpshufb %xmm15, %xmm1, %xmm0
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm1, %xmm16
; AVX512F-FAST-NEXT:    vmovdqa (%rsi), %xmm3
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm25, %xmm1
; AVX512F-FAST-NEXT:    vpshufb %xmm1, %xmm3, %xmm2
; AVX512F-FAST-NEXT:    vporq %xmm0, %xmm2, %xmm28
; AVX512F-FAST-NEXT:    vmovdqa (%rcx), %xmm7
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm26, %xmm0
; AVX512F-FAST-NEXT:    vpshufb %xmm0, %xmm7, %xmm0
; AVX512F-FAST-NEXT:    vmovdqa (%rdx), %xmm2
; AVX512F-FAST-NEXT:    vpshufb %xmm14, %xmm2, %xmm14
; AVX512F-FAST-NEXT:    vporq %xmm0, %xmm14, %xmm29
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm14 = [1,1,2,2,2,2,2,2]
; AVX512F-FAST-NEXT:    vmovdqa 32(%r8), %ymm0
; AVX512F-FAST-NEXT:    vpermd %ymm0, %ymm14, %ymm14
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} ymm25 = [255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255]
; AVX512F-FAST-NEXT:    vpandnq %ymm14, %ymm25, %ymm14
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm1 = [12,128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128]
; AVX512F-FAST-NEXT:    vpshufb %ymm1, %ymm0, %ymm15
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm15, %zmm14, %zmm27
; AVX512F-FAST-NEXT:    vmovdqa64 (%r8), %zmm25
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm31 = <4,u,5,5,5,5,u,6,30,30,30,u,31,31,31,31>
; AVX512F-FAST-NEXT:    vpermi2d %zmm25, %zmm0, %zmm31
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm15 = <4,u,5,5,5,5,u,6>
; AVX512F-FAST-NEXT:    vmovdqa (%r8), %ymm0
; AVX512F-FAST-NEXT:    vpermd %ymm0, %ymm15, %ymm15
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} ymm26 = [255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255,255,0,255,255,255]
; AVX512F-FAST-NEXT:    vpandnq %ymm15, %ymm26, %ymm15
; AVX512F-FAST-NEXT:    vpshufb %ymm1, %ymm0, %ymm1
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm15, %zmm1, %zmm26
; AVX512F-FAST-NEXT:    vpbroadcastq {{.*#+}} ymm1 = [9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12]
; AVX512F-FAST-NEXT:    vpshufb %ymm1, %ymm9, %ymm9
; AVX512F-FAST-NEXT:    vmovdqa64 %ymm30, %ymm13
; AVX512F-FAST-NEXT:    vpshufb %ymm13, %ymm12, %ymm15
; AVX512F-FAST-NEXT:    vpshufb %ymm1, %ymm12, %ymm1
; AVX512F-FAST-NEXT:    vbroadcasti128 {{.*#+}} ymm12 = [18,19,128,21,128,21,20,128,22,128,24,128,22,23,128,25,18,19,128,21,128,21,20,128,22,128,24,128,22,23,128,25]
; AVX512F-FAST-NEXT:    # ymm12 = mem[0,1,0,1]
; AVX512F-FAST-NEXT:    vpshufb %ymm12, %ymm8, %ymm8
; AVX512F-FAST-NEXT:    vpshufb %ymm12, %ymm6, %ymm12
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm6 = ymm6[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm6[27],zero,zero,ymm6[26],zero,ymm6[28],zero,ymm6[30],zero,zero,ymm6[29],zero,ymm6[31],zero,zero
; AVX512F-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm10 = xmm10[0],xmm11[0],xmm10[1],xmm11[1],xmm10[2],xmm11[2],xmm10[3],xmm11[3],xmm10[4],xmm11[4],xmm10[5],xmm11[5],xmm10[6],xmm11[6],xmm10[7],xmm11[7]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm30 = ymm9[2,2,3,3]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm11 = ymm5[3,u,5,u,u,4,u,6,u,8,u,u,7,u,9,u,19,u,21,u,u,20,u,22,u,24,u,u,23,u,25,u]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm5 = ymm5[11,u,u,10,u,12,u,u,u,u,13,u,15,u,u,14,27,u,u,26,u,28,u,u,u,u,29,u,31,u,u,30]
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm9 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm4[21],zero,zero,ymm4[20],zero,ymm4[22],zero,ymm4[24],zero,zero,ymm4[23],zero,ymm4[25],zero,zero
; AVX512F-FAST-NEXT:    vpshufb {{.*#+}} ymm4 = ymm4[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm4[26],zero,ymm4[28],zero,zero,ymm4[27],zero,ymm4[29],zero,ymm4[31],zero,zero,ymm4[30],zero
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm18, %xmm14
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm17, %xmm13
; AVX512F-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm14 = xmm14[0],xmm13[0],xmm14[1],xmm13[1],xmm14[2],xmm13[2],xmm14[3],xmm13[3],xmm14[4],xmm13[4],xmm14[5],xmm13[5],xmm14[6],xmm13[6],xmm14[7],xmm13[7]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm15 = ymm15[2,2,3,3]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm12 = ymm12[2,2,3,3]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm11 = ymm11[2,2,3,3]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm9 = ymm9[2,2,3,3]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm1 = ymm1[2,2,3,3]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[2,2,3,3]
; AVX512F-FAST-NEXT:    vmovdqa64 %xmm16, %xmm13
; AVX512F-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm3 = xmm13[0],xmm3[0],xmm13[1],xmm3[1],xmm13[2],xmm3[2],xmm13[3],xmm3[3],xmm13[4],xmm3[4],xmm13[5],xmm3[5],xmm13[6],xmm3[6],xmm13[7],xmm3[7]
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} xmm13 = <0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13>
; AVX512F-FAST-NEXT:    vpshufb %xmm13, %xmm14, %xmm14
; AVX512F-FAST-NEXT:    vpshufb %xmm13, %xmm3, %xmm3
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} xmm13 = <2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8>
; AVX512F-FAST-NEXT:    vpshufb %xmm13, %xmm10, %xmm10
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm10 = ymm10[0,0,1,1]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm5 = ymm5[2,2,3,3]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,2,3,3]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm14 = ymm14[0,0,1,1]
; AVX512F-FAST-NEXT:    vinserti32x4 $2, %xmm28, %zmm3, %zmm3
; AVX512F-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm2 = xmm7[0],xmm2[0],xmm7[1],xmm2[1],xmm7[2],xmm2[2],xmm7[3],xmm2[3],xmm7[4],xmm2[4],xmm7[5],xmm2[5],xmm7[6],xmm2[6],xmm7[7],xmm2[7]
; AVX512F-FAST-NEXT:    vpshufb %xmm13, %xmm2, %xmm2
; AVX512F-FAST-NEXT:    vinserti32x4 $2, %xmm29, %zmm2, %zmm2
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm0, %zmm25, %zmm0
; AVX512F-FAST-NEXT:    vpermq $80, {{[-0-9]+}}(%r{{[sb]}}p), %ymm7 # 32-byte Folded Reload
; AVX512F-FAST-NEXT:    # ymm7 = mem[0,0,1,1]
; AVX512F-FAST-NEXT:    vinserti64x4 $1, {{[-0-9]+}}(%r{{[sb]}}p), %zmm7, %zmm7 # 32-byte Folded Reload
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} ymm13 = ymm19[0,0,1,1]
; AVX512F-FAST-NEXT:    vinserti64x4 $1, {{[-0-9]+}}(%r{{[sb]}}p), %zmm13, %zmm13 # 32-byte Folded Reload
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm16 = [255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0]
; AVX512F-FAST-NEXT:    vpternlogq $226, %zmm7, %zmm16, %zmm13
; AVX512F-FAST-NEXT:    vpor %ymm15, %ymm12, %ymm7
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm7, %zmm20, %zmm7
; AVX512F-FAST-NEXT:    vmovdqa {{.*#+}} ymm12 = [18374966859431608575,18374966859431608575,18446463693966278400,18446463693966278400]
; AVX512F-FAST-NEXT:    vpternlogq $248, %ymm12, %ymm11, %ymm9
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm9, %zmm21, %zmm9
; AVX512F-FAST-NEXT:    vpternlogq $226, %zmm7, %zmm16, %zmm9
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} zmm7 = zmm22[2,2,3,3,6,6,7,7]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} zmm11 = zmm23[2,2,3,3,6,6,7,7]
; AVX512F-FAST-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm7, %zmm11
; AVX512F-FAST-NEXT:    vpternlogq $248, %ymm12, %ymm1, %ymm6
; AVX512F-FAST-NEXT:    vpandq %ymm12, %ymm30, %ymm1
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm1, %zmm8, %zmm1
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} zmm7 = zmm24[2,2,3,3,6,6,7,7]
; AVX512F-FAST-NEXT:    vporq %zmm7, %zmm1, %zmm1
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm7 = [0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255,0,0,255,255,255]
; AVX512F-FAST-NEXT:    vpternlogq $226, %zmm11, %zmm7, %zmm1
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm10, %zmm6, %zmm6
; AVX512F-FAST-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm5, %ymm4
; AVX512F-FAST-NEXT:    vinserti64x4 $1, %ymm14, %zmm4, %zmm4
; AVX512F-FAST-NEXT:    vpternlogq $226, %zmm6, %zmm7, %zmm4
; AVX512F-FAST-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm13, %zmm27
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm1, %zmm31
; AVX512F-FAST-NEXT:    vpternlogq $248, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm9, %zmm26
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm1 = <6,6,6,u,7,7,7,7,u,8,8,8,8,u,9,9>
; AVX512F-FAST-NEXT:    vpermd %zmm25, %zmm1, %zmm1
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm4, %zmm1
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} zmm3 = zmm3[0,0,1,1,4,4,5,5]
; AVX512F-FAST-NEXT:    vpermq {{.*#+}} zmm2 = zmm2[0,0,1,1,4,4,5,5]
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm3, %zmm2
; AVX512F-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm3 = <u,0,0,0,0,u,1,1,9,9,u,10,10,10,10,u>
; AVX512F-FAST-NEXT:    vpermd %zmm0, %zmm3, %zmm0
; AVX512F-FAST-NEXT:    vpternlogq $216, {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %zmm2, %zmm0
; AVX512F-FAST-NEXT:    vmovdqa64 %zmm26, 64(%r9)
; AVX512F-FAST-NEXT:    vmovdqa64 %zmm0, (%r9)
; AVX512F-FAST-NEXT:    vmovdqa64 %zmm1, 128(%r9)
; AVX512F-FAST-NEXT:    vmovdqa64 %zmm31, 256(%r9)
; AVX512F-FAST-NEXT:    vmovdqa64 %zmm27, 192(%r9)
; AVX512F-FAST-NEXT:    vzeroupper
; AVX512F-FAST-NEXT:    retq
;
; AVX512BW-ONLY-SLOW-LABEL: store_i8_stride5_vf64:
; AVX512BW-ONLY-SLOW:       # %bb.0:
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 (%r8), %zmm3
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa (%rcx), %ymm0
; AVX512BW-ONLY-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm8 = [9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm8, %ymm0, %ymm2
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa (%rdx), %ymm1
; AVX512BW-ONLY-SLOW-NEXT:    vpshufhw {{.*#+}} ymm4 = ymm1[0,1,2,3,5,6,7,6,8,9,10,11,13,14,15,14]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufd {{.*#+}} ymm4 = ymm4[2,2,3,3,6,6,7,7]
; AVX512BW-ONLY-SLOW-NEXT:    movl $693250386, %eax # imm = 0x29522952
; AVX512BW-ONLY-SLOW-NEXT:    kmovd %eax, %k1
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %ymm4, %ymm2 {%k1}
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa 32(%rdx), %xmm6
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa 32(%rcx), %xmm12
; AVX512BW-ONLY-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm4 = xmm12[0],xmm6[0],xmm12[1],xmm6[1],xmm12[2],xmm6[2],xmm12[3],xmm6[3],xmm12[4],xmm6[4],xmm12[5],xmm6[5],xmm12[6],xmm6[6],xmm12[7],xmm6[7]
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa {{.*#+}} xmm7 = <2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8>
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm7, %xmm4, %xmm4
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,1,1]
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm4, %zmm2, %zmm10
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa (%rsi), %ymm4
; AVX512BW-ONLY-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm15 = [11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm15, %ymm4, %ymm2
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa (%rdi), %ymm5
; AVX512BW-ONLY-SLOW-NEXT:    vpshufhw {{.*#+}} ymm9 = ymm5[0,1,2,3,6,5,6,7,8,9,10,11,14,13,14,15]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufd {{.*#+}} ymm9 = ymm9[2,2,3,3,6,6,7,7]
; AVX512BW-ONLY-SLOW-NEXT:    movl $1251232404, %eax # imm = 0x4A944A94
; AVX512BW-ONLY-SLOW-NEXT:    kmovd %eax, %k5
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %ymm9, %ymm2 {%k5}
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa 32(%rsi), %xmm13
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa 32(%rdi), %xmm14
; AVX512BW-ONLY-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm11 = xmm14[0],xmm13[0],xmm14[1],xmm13[1],xmm14[2],xmm13[2],xmm14[3],xmm13[3],xmm14[4],xmm13[4],xmm14[5],xmm13[5],xmm14[6],xmm13[6],xmm14[7],xmm13[7]
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa {{.*#+}} xmm9 = <0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13>
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm9, %xmm11, %xmm11
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm11 = ymm11[0,0,1,1]
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm11, %zmm2, %zmm2
; AVX512BW-ONLY-SLOW-NEXT:    movabsq $1785168781326730801, %rax # imm = 0x18C6318C6318C631
; AVX512BW-ONLY-SLOW-NEXT:    kmovq %rax, %k4
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm10, %zmm2 {%k4}
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 32(%r8), %ymm16
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm10 = [6,6,6,6,7,7,7,7,16,16,16,16,16,16,17,17]
; AVX512BW-ONLY-SLOW-NEXT:    vpermi2d %zmm16, %zmm3, %zmm10
; AVX512BW-ONLY-SLOW-NEXT:    movabsq $2380225041768974402, %rax # imm = 0x2108421084210842
; AVX512BW-ONLY-SLOW-NEXT:    kmovq %rax, %k2
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm10, %zmm2 {%k2}
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 32(%rdx), %ymm23
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa {{.*#+}} ymm10 = [128,128,12,13,128,128,128,128,14,128,128,128,14,15,128,128,128,128,16,128,128,128,16,17,128,128,128,128,18,128,128,128]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm10, %ymm23, %ymm17
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 32(%rcx), %ymm24
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa {{.*#+}} ymm11 = [128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm11, %ymm24, %ymm18
; AVX512BW-ONLY-SLOW-NEXT:    vporq %ymm17, %ymm18, %ymm17
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} xmm20 = <128,6,128,8,u,128,7,128,9,128,11,u,128,10,128,12>
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm20, %xmm12, %xmm12
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} xmm22 = <6,128,8,128,u,7,128,9,128,11,128,u,10,128,12,128>
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm22, %xmm6, %xmm6
; AVX512BW-ONLY-SLOW-NEXT:    vpor %xmm6, %xmm12, %xmm6
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,0,1,1]
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm17, %zmm6, %zmm6
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} xmm19 = <8,128,u,7,128,9,128,u,128,u,10,128,12,128,u,11>
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm19, %xmm14, %xmm12
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} xmm21 = <128,8,u,128,7,128,9,u,11,u,128,10,128,12,u,128>
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm21, %xmm13, %xmm13
; AVX512BW-ONLY-SLOW-NEXT:    vpor %xmm12, %xmm13, %xmm12
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm13 = ymm12[0,0,1,1]
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 32(%rdi), %ymm25
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa {{.*#+}} ymm12 = <3,3,3,u,4,4,4,4>
; AVX512BW-ONLY-SLOW-NEXT:    vpermd %ymm25, %ymm12, %ymm17
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 32(%rsi), %ymm26
; AVX512BW-ONLY-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm14 = [0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14]
; AVX512BW-ONLY-SLOW-NEXT:    movl $138547332, %eax # imm = 0x8421084
; AVX512BW-ONLY-SLOW-NEXT:    kmovd %eax, %k3
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm14, %ymm26, %ymm17 {%k3}
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm17, %zmm13, %zmm13
; AVX512BW-ONLY-SLOW-NEXT:    movabsq $-8330787646191410408, %rax # imm = 0x8C6318C6318C6318
; AVX512BW-ONLY-SLOW-NEXT:    kmovq %rax, %k2
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm13, %zmm6 {%k2}
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa {{.*#+}} ymm13 = <3,3,3,3,u,4,4,4>
; AVX512BW-ONLY-SLOW-NEXT:    vpermd %ymm16, %ymm13, %ymm17
; AVX512BW-ONLY-SLOW-NEXT:    vpshufd {{.*#+}} xmm18 = mem[1,1,2,2]
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm18 = ymm18[0,1,1,1]
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm17, %zmm18, %zmm17
; AVX512BW-ONLY-SLOW-NEXT:    movabsq $4760450083537948804, %rax # imm = 0x4210842108421084
; AVX512BW-ONLY-SLOW-NEXT:    kmovq %rax, %k6
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm17, %zmm6 {%k6}
; AVX512BW-ONLY-SLOW-NEXT:    vbroadcasti32x4 {{.*#+}} ymm17 = [19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128]
; AVX512BW-ONLY-SLOW-NEXT:    # ymm17 = mem[0,1,2,3,0,1,2,3]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm17, %ymm26, %ymm18
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm27 = ymm18[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} ymm18 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128,128]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm18, %ymm25, %ymm28
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm28 = ymm28[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vporq %ymm27, %ymm28, %ymm27
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm15, %ymm26, %ymm15
; AVX512BW-ONLY-SLOW-NEXT:    vpshufhw {{.*#+}} ymm25 = ymm25[0,1,2,3,6,5,6,7,8,9,10,11,14,13,14,15]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufd {{.*#+}} ymm25 = ymm25[2,2,3,3,6,6,7,7]
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %ymm25, %ymm15 {%k5}
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm15 = ymm15[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm15, %zmm27, %zmm15
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} ymm25 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,18,19,128,21,128,21,20,128,22,128,24,128,22,23,128,25]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm25, %ymm23, %ymm26
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm26 = ymm26[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vbroadcasti32x4 {{.*#+}} ymm27 = [128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128]
; AVX512BW-ONLY-SLOW-NEXT:    # ymm27 = mem[0,1,2,3,0,1,2,3]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm27, %ymm24, %ymm28
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm28 = ymm28[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vporq %ymm26, %ymm28, %ymm26
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm8, %ymm24, %ymm8
; AVX512BW-ONLY-SLOW-NEXT:    vpshufhw {{.*#+}} ymm23 = ymm23[0,1,2,3,5,6,7,6,8,9,10,11,13,14,15,14]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufd {{.*#+}} ymm23 = ymm23[2,2,3,3,6,6,7,7]
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %ymm23, %ymm8 {%k1}
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm8, %zmm26, %zmm8
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm15, %zmm8 {%k4}
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm15 = [4,6,5,5,5,5,4,6,6,6,6,6,7,7,7,7]
; AVX512BW-ONLY-SLOW-NEXT:    vpermd %zmm16, %zmm15, %zmm15
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 (%rdx), %xmm16
; AVX512BW-ONLY-SLOW-NEXT:    movabsq $-8925843906633654008, %rax # imm = 0x8421084210842108
; AVX512BW-ONLY-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm15, %zmm8 {%k1}
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa (%rcx), %xmm15
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm20, %xmm15, %xmm20
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm22, %xmm16, %xmm22
; AVX512BW-ONLY-SLOW-NEXT:    vporq %xmm20, %xmm22, %xmm20
; AVX512BW-ONLY-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm15 = xmm15[0],xmm16[0],xmm15[1],xmm16[1],xmm15[2],xmm16[2],xmm15[3],xmm16[3],xmm15[4],xmm16[4],xmm15[5],xmm16[5],xmm15[6],xmm16[6],xmm15[7],xmm16[7]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm7, %xmm15, %xmm7
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa (%rsi), %xmm15
; AVX512BW-ONLY-SLOW-NEXT:    vinserti32x4 $2, %xmm20, %zmm7, %zmm7
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 (%rdi), %xmm16
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm19, %xmm16, %xmm19
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm21, %xmm15, %xmm20
; AVX512BW-ONLY-SLOW-NEXT:    vporq %xmm19, %xmm20, %xmm19
; AVX512BW-ONLY-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm15 = xmm16[0],xmm15[0],xmm16[1],xmm15[1],xmm16[2],xmm15[2],xmm16[3],xmm15[3],xmm16[4],xmm15[4],xmm16[5],xmm15[5],xmm16[6],xmm15[6],xmm16[7],xmm15[7]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %xmm9, %xmm15, %xmm9
; AVX512BW-ONLY-SLOW-NEXT:    vinserti32x4 $2, %xmm19, %zmm9, %zmm9
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} zmm7 = zmm7[0,0,1,1,4,4,5,5]
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} zmm9 = zmm9[0,0,1,1,4,4,5,5]
; AVX512BW-ONLY-SLOW-NEXT:    movabsq $-4165393823095705204, %rax # imm = 0xC6318C6318C6318C
; AVX512BW-ONLY-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm7, %zmm9 {%k1}
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm7 = [0,0,0,0,0,0,1,1,1,1,2,2,2,2,2,2]
; AVX512BW-ONLY-SLOW-NEXT:    vpermd %zmm3, %zmm7, %zmm3
; AVX512BW-ONLY-SLOW-NEXT:    movabsq $595056260442243600, %rax # imm = 0x842108421084210
; AVX512BW-ONLY-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm3, %zmm9 {%k1}
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm25, %ymm1, %ymm3
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm27, %ymm0, %ymm7
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vpor %ymm3, %ymm7, %ymm3
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm10, %ymm1, %ymm1
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm11, %ymm0, %ymm0
; AVX512BW-ONLY-SLOW-NEXT:    vpor %ymm1, %ymm0, %ymm0
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm3, %zmm0, %zmm0
; AVX512BW-ONLY-SLOW-NEXT:    vpermd %ymm5, %ymm12, %ymm1
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm14, %ymm4, %ymm1 {%k3}
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm17, %ymm4, %ymm3
; AVX512BW-ONLY-SLOW-NEXT:    vpshufb %ymm18, %ymm5, %ymm4
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,2,3,3]
; AVX512BW-ONLY-SLOW-NEXT:    vpor %ymm3, %ymm4, %ymm3
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm3, %zmm1, %zmm1
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm0, %zmm1 {%k2}
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa (%r8), %ymm0
; AVX512BW-ONLY-SLOW-NEXT:    vpermd %ymm0, %ymm13, %ymm3
; AVX512BW-ONLY-SLOW-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[0,2,1,1,4,6,5,5]
; AVX512BW-ONLY-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,3,3,2]
; AVX512BW-ONLY-SLOW-NEXT:    vinserti64x4 $1, %ymm0, %zmm3, %zmm0
; AVX512BW-ONLY-SLOW-NEXT:    movabsq $1190112520884487201, %rax # imm = 0x1084210842108421
; AVX512BW-ONLY-SLOW-NEXT:    kmovq %rax, %k1
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqu8 %zmm0, %zmm1 {%k1}
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 %zmm1, 64(%r9)
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 %zmm9, (%r9)
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 %zmm8, 256(%r9)
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 %zmm6, 192(%r9)
; AVX512BW-ONLY-SLOW-NEXT:    vmovdqa64 %zmm2, 128(%r9)
; AVX512BW-ONLY-SLOW-NEXT:    vzeroupper
; AVX512BW-ONLY-SLOW-NEXT:    retq
;
; AVX512BW-FAST-LABEL: store_i8_stride5_vf64:
; AVX512BW-FAST:       # %bb.0:
; AVX512BW-FAST-NEXT:    vmovdqa64 (%r8), %zmm5
; AVX512BW-FAST-NEXT:    vmovdqa64 32(%rdx), %ymm19
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} ymm1 = [128,128,12,13,128,128,128,128,14,128,128,128,14,15,128,128,128,128,16,128,128,128,16,17,128,128,128,128,18,128,128,128]
; AVX512BW-FAST-NEXT:    vpshufb %ymm1, %ymm19, %ymm0
; AVX512BW-FAST-NEXT:    vmovdqa 32(%rcx), %ymm13
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} ymm2 = [128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128]
; AVX512BW-FAST-NEXT:    vpshufb %ymm2, %ymm13, %ymm3
; AVX512BW-FAST-NEXT:    vpor %ymm0, %ymm3, %ymm0
; AVX512BW-FAST-NEXT:    vmovdqa (%rcx), %xmm6
; AVX512BW-FAST-NEXT:    vmovdqa64 32(%rcx), %xmm16
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} xmm8 = <128,6,128,8,u,128,7,128,9,128,11,u,128,10,128,12>
; AVX512BW-FAST-NEXT:    vpshufb %xmm8, %xmm16, %xmm3
; AVX512BW-FAST-NEXT:    vmovdqa (%rdx), %xmm7
; AVX512BW-FAST-NEXT:    vmovdqa64 32(%rdx), %xmm18
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} xmm10 = <6,128,8,128,u,7,128,9,128,11,128,u,10,128,12,128>
; AVX512BW-FAST-NEXT:    vpshufb %xmm10, %xmm18, %xmm4
; AVX512BW-FAST-NEXT:    vpor %xmm3, %xmm4, %xmm3
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[0,0,1,1]
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm0, %zmm3, %zmm0
; AVX512BW-FAST-NEXT:    vmovdqa (%rdi), %xmm9
; AVX512BW-FAST-NEXT:    vmovdqa 32(%rdi), %xmm15
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} xmm12 = <8,128,u,7,128,9,128,u,128,u,10,128,12,128,u,11>
; AVX512BW-FAST-NEXT:    vpshufb %xmm12, %xmm15, %xmm3
; AVX512BW-FAST-NEXT:    vmovdqa (%rsi), %xmm11
; AVX512BW-FAST-NEXT:    vmovdqa64 32(%rsi), %xmm17
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} xmm14 = <128,8,u,128,7,128,9,u,11,u,128,10,128,12,u,128>
; AVX512BW-FAST-NEXT:    vpshufb %xmm14, %xmm17, %xmm4
; AVX512BW-FAST-NEXT:    vpor %xmm3, %xmm4, %xmm3
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm20 = ymm3[0,0,1,1]
; AVX512BW-FAST-NEXT:    vmovdqa64 32(%rdi), %ymm21
; AVX512BW-FAST-NEXT:    vmovdqa {{.*#+}} ymm3 = <3,3,3,u,4,4,4,4>
; AVX512BW-FAST-NEXT:    vpermd %ymm21, %ymm3, %ymm22
; AVX512BW-FAST-NEXT:    vmovdqa64 32(%rsi), %ymm23
; AVX512BW-FAST-NEXT:    vpbroadcastq {{.*#+}} ymm4 = [0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14]
; AVX512BW-FAST-NEXT:    movl $138547332, %eax # imm = 0x8421084
; AVX512BW-FAST-NEXT:    kmovd %eax, %k1
; AVX512BW-FAST-NEXT:    vpshufb %ymm4, %ymm23, %ymm22 {%k1}
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm22, %zmm20, %zmm20
; AVX512BW-FAST-NEXT:    movabsq $-8330787646191410408, %rax # imm = 0x8C6318C6318C6318
; AVX512BW-FAST-NEXT:    kmovq %rax, %k2
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm20, %zmm0 {%k2}
; AVX512BW-FAST-NEXT:    vmovdqa64 32(%r8), %ymm20
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm22 = <1,1,2,2,2,2,2,2,27,27,27,27,u,28,28,28>
; AVX512BW-FAST-NEXT:    vpermi2d %zmm5, %zmm20, %zmm22
; AVX512BW-FAST-NEXT:    movabsq $4760450083537948804, %rax # imm = 0x4210842108421084
; AVX512BW-FAST-NEXT:    kmovq %rax, %k3
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm22, %zmm0 {%k3}
; AVX512BW-FAST-NEXT:    vshufi64x2 {{.*#+}} zmm22 = zmm23[0,1,2,3],mem[4,5,6,7]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} zmm22 = zmm22[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19],zero,zmm22[21],zero,zero,zmm22[20],zero,zmm22[22],zero,zmm22[24],zero,zero,zmm22[23],zero,zmm22[25],zero,zmm22[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,59],zero,zero,zmm22[58],zero,zmm22[60],zero,zero,zero,zero,zmm22[61],zero,zmm22[63],zero,zero,zmm22[62]
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} zmm22 = zmm22[2,2,3,3,6,6,7,7]
; AVX512BW-FAST-NEXT:    vshufi64x2 {{.*#+}} zmm21 = zmm21[0,1,2,3],mem[4,5,6,7]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} zmm21 = zmm21[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zmm21[21],zero,zero,zmm21[20],zero,zmm21[22],zero,zmm21[24],zero,zero,zmm21[23],zero,zmm21[25],zero,zero,zmm21[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,zmm21[58],zero,zmm21[60],zero,zero,zmm21[59],zero,zmm21[61],zero,zmm21[63],zero,zero,zmm21[62],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} zmm21 = zmm21[2,2,3,3,6,6,7,7]
; AVX512BW-FAST-NEXT:    vporq %zmm22, %zmm21, %zmm21
; AVX512BW-FAST-NEXT:    vshufi64x2 {{.*#+}} zmm19 = zmm19[0,1,2,3],mem[4,5,6,7]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} zmm19 = zmm19[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,18,19],zero,zmm19[21],zero,zmm19[21,20],zero,zmm19[22],zero,zmm19[24],zero,zmm19[22,23],zero,zmm19[25,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,57],zero,zmm19[59],zero,zero,zmm19[58],zero,zmm19[60],zero,zmm19[62],zero,zero,zmm19[61],zero,zmm19[63],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} zmm19 = zmm19[2,2,3,3,6,6,7,7]
; AVX512BW-FAST-NEXT:    vshufi64x2 {{.*#+}} zmm13 = zmm13[0,1,2,3],mem[4,5,6,7]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} zmm13 = zmm13[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,zmm13[19],zero,zmm13[21],zero,zero,zmm13[20],zero,zmm13[22],zero,zmm13[24],zero,zero,zmm13[23],zero,zmm13[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zmm13[59],zero,zero,zmm13[58],zero,zmm13[60],zero,zmm13[62],zero,zero,zmm13[61],zero,zmm13[63],zero,zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} zmm13 = zmm13[2,2,3,3,6,6,7,7]
; AVX512BW-FAST-NEXT:    vporq %zmm19, %zmm13, %zmm13
; AVX512BW-FAST-NEXT:    movabsq $1785168781326730801, %rax # imm = 0x18C6318C6318C631
; AVX512BW-FAST-NEXT:    kmovq %rax, %k3
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm21, %zmm13 {%k3}
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm19 = [4,6,5,5,5,5,4,6,30,30,30,30,31,31,31,31]
; AVX512BW-FAST-NEXT:    vpermi2d %zmm5, %zmm20, %zmm19
; AVX512BW-FAST-NEXT:    movabsq $-8925843906633654008, %rax # imm = 0x8421084210842108
; AVX512BW-FAST-NEXT:    kmovq %rax, %k4
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm19, %zmm13 {%k4}
; AVX512BW-FAST-NEXT:    vmovdqa64 (%rcx), %ymm19
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm20 = ymm19[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,25],zero,ymm19[27],zero,zero,ymm19[26],zero,ymm19[28],zero,ymm19[30],zero,zero,ymm19[29],zero,ymm19[31],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm20 = ymm20[2,2,3,3]
; AVX512BW-FAST-NEXT:    vmovdqa64 (%rdx), %ymm21
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm22 = ymm21[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm21[27],zero,zero,ymm21[26],zero,ymm21[28],zero,ymm21[30],zero,zero,ymm21[29],zero,ymm21[31],zero,zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm22 = ymm22[2,2,3,3]
; AVX512BW-FAST-NEXT:    vporq %ymm20, %ymm22, %ymm20
; AVX512BW-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm16 = xmm16[0],xmm18[0],xmm16[1],xmm18[1],xmm16[2],xmm18[2],xmm16[3],xmm18[3],xmm16[4],xmm18[4],xmm16[5],xmm18[5],xmm16[6],xmm18[6],xmm16[7],xmm18[7]
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} xmm18 = <2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8>
; AVX512BW-FAST-NEXT:    vpshufb %xmm18, %xmm16, %xmm16
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm16 = ymm16[0,0,1,1]
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm16, %zmm20, %zmm20
; AVX512BW-FAST-NEXT:    vmovdqa64 (%rsi), %ymm16
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm22 = ymm16[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,27],zero,zero,ymm16[26],zero,ymm16[28],zero,zero,zero,zero,ymm16[29],zero,ymm16[31],zero,zero,ymm16[30]
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm22 = ymm22[2,2,3,3]
; AVX512BW-FAST-NEXT:    vmovdqa64 (%rdi), %ymm23
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm24 = ymm23[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm23[26],zero,ymm23[28],zero,zero,ymm23[27],zero,ymm23[29],zero,ymm23[31],zero,zero,ymm23[30],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm24 = ymm24[2,2,3,3]
; AVX512BW-FAST-NEXT:    vporq %ymm22, %ymm24, %ymm22
; AVX512BW-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm15 = xmm15[0],xmm17[0],xmm15[1],xmm17[1],xmm15[2],xmm17[2],xmm15[3],xmm17[3],xmm15[4],xmm17[4],xmm15[5],xmm17[5],xmm15[6],xmm17[6],xmm15[7],xmm17[7]
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} xmm17 = <0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13>
; AVX512BW-FAST-NEXT:    vpshufb %xmm17, %xmm15, %xmm15
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm15 = ymm15[0,0,1,1]
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm15, %zmm22, %zmm15
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm20, %zmm15 {%k3}
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm20 = [6,6,6,6,7,7,7,7,8,8,8,8,8,8,9,9]
; AVX512BW-FAST-NEXT:    vpermd %zmm5, %zmm20, %zmm5
; AVX512BW-FAST-NEXT:    movabsq $2380225041768974402, %rax # imm = 0x2108421084210842
; AVX512BW-FAST-NEXT:    kmovq %rax, %k3
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm5, %zmm15 {%k3}
; AVX512BW-FAST-NEXT:    vpshufb %xmm8, %xmm6, %xmm5
; AVX512BW-FAST-NEXT:    vpshufb %xmm10, %xmm7, %xmm8
; AVX512BW-FAST-NEXT:    vpor %xmm5, %xmm8, %xmm5
; AVX512BW-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm6 = xmm6[0],xmm7[0],xmm6[1],xmm7[1],xmm6[2],xmm7[2],xmm6[3],xmm7[3],xmm6[4],xmm7[4],xmm6[5],xmm7[5],xmm6[6],xmm7[6],xmm6[7],xmm7[7]
; AVX512BW-FAST-NEXT:    vpshufb %xmm18, %xmm6, %xmm6
; AVX512BW-FAST-NEXT:    vinserti32x4 $2, %xmm5, %zmm6, %zmm5
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} zmm5 = zmm5[0,0,1,1,4,4,5,5]
; AVX512BW-FAST-NEXT:    vpshufb %xmm12, %xmm9, %xmm6
; AVX512BW-FAST-NEXT:    vpshufb %xmm14, %xmm11, %xmm7
; AVX512BW-FAST-NEXT:    vpor %xmm6, %xmm7, %xmm6
; AVX512BW-FAST-NEXT:    vpunpcklbw {{.*#+}} xmm7 = xmm9[0],xmm11[0],xmm9[1],xmm11[1],xmm9[2],xmm11[2],xmm9[3],xmm11[3],xmm9[4],xmm11[4],xmm9[5],xmm11[5],xmm9[6],xmm11[6],xmm9[7],xmm11[7]
; AVX512BW-FAST-NEXT:    vpshufb %xmm17, %xmm7, %xmm7
; AVX512BW-FAST-NEXT:    vinserti32x4 $2, %xmm6, %zmm7, %zmm6
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} zmm6 = zmm6[0,0,1,1,4,4,5,5]
; AVX512BW-FAST-NEXT:    movabsq $-4165393823095705204, %rax # imm = 0xC6318C6318C6318C
; AVX512BW-FAST-NEXT:    kmovq %rax, %k3
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm5, %zmm6 {%k3}
; AVX512BW-FAST-NEXT:    vbroadcasti64x4 {{.*#+}} zmm5 = mem[0,1,2,3,0,1,2,3]
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm7 = [0,0,0,0,0,0,1,1,9,9,10,10,10,10,10,10]
; AVX512BW-FAST-NEXT:    vpermd %zmm5, %zmm7, %zmm7
; AVX512BW-FAST-NEXT:    movabsq $595056260442243600, %rax # imm = 0x842108421084210
; AVX512BW-FAST-NEXT:    kmovq %rax, %k3
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm7, %zmm6 {%k3}
; AVX512BW-FAST-NEXT:    vpshufb %ymm1, %ymm21, %ymm1
; AVX512BW-FAST-NEXT:    vpshufb %ymm2, %ymm19, %ymm2
; AVX512BW-FAST-NEXT:    vpor %ymm1, %ymm2, %ymm1
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = ymm21[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,18,19],zero,ymm21[21],zero,ymm21[21,20],zero,ymm21[22],zero,ymm21[24],zero,ymm21[22,23],zero,ymm21[25]
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm7 = ymm19[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,zero,ymm19[19],zero,ymm19[21],zero,zero,ymm19[20],zero,ymm19[22],zero,ymm19[24],zero,zero,ymm19[23],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpor %ymm2, %ymm7, %ymm2
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm2, %zmm1, %zmm1
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm2 = ymm16[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,19],zero,ymm16[21],zero,zero,ymm16[20],zero,ymm16[22],zero,ymm16[24],zero,zero,ymm16[23],zero,ymm16[25],zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpshufb {{.*#+}} ymm7 = ymm23[u,u,u,u,u,u,u,u,u,u,u,u,u,u,u,u],zero,ymm23[21],zero,zero,ymm23[20],zero,ymm23[22],zero,ymm23[24],zero,zero,ymm23[23],zero,ymm23[25],zero,zero
; AVX512BW-FAST-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512BW-FAST-NEXT:    vpor %ymm2, %ymm7, %ymm2
; AVX512BW-FAST-NEXT:    vpermd %ymm23, %ymm3, %ymm3
; AVX512BW-FAST-NEXT:    vpshufb %ymm4, %ymm16, %ymm3 {%k1}
; AVX512BW-FAST-NEXT:    vinserti64x4 $1, %ymm2, %zmm3, %zmm2
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm1, %zmm2 {%k2}
; AVX512BW-FAST-NEXT:    vmovdqa64 {{.*#+}} zmm1 = <3,3,3,3,u,4,4,4,12,14,13,13,13,13,12,14>
; AVX512BW-FAST-NEXT:    vpermd %zmm5, %zmm1, %zmm1
; AVX512BW-FAST-NEXT:    movabsq $1190112520884487201, %rax # imm = 0x1084210842108421
; AVX512BW-FAST-NEXT:    kmovq %rax, %k1
; AVX512BW-FAST-NEXT:    vmovdqu8 %zmm1, %zmm2 {%k1}
; AVX512BW-FAST-NEXT:    vmovdqa64 %zmm2, 64(%r9)
; AVX512BW-FAST-NEXT:    vmovdqa64 %zmm6, (%r9)
; AVX512BW-FAST-NEXT:    vmovdqa64 %zmm15, 128(%r9)
; AVX512BW-FAST-NEXT:    vmovdqa64 %zmm13, 256(%r9)
; AVX512BW-FAST-NEXT:    vmovdqa64 %zmm0, 192(%r9)
; AVX512BW-FAST-NEXT:    vzeroupper
; AVX512BW-FAST-NEXT:    retq
;
; AVX512DQBW-SLOW-LABEL: store_i8_stride5_vf64:
; AVX512DQBW-SLOW:       # %bb.0:
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 (%r8), %zmm3
; AVX512DQBW-SLOW-NEXT:    vmovdqa (%rcx), %ymm0
; AVX512DQBW-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm8 = [9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12,9,14,11,0,13,10,15,12]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm8, %ymm0, %ymm2
; AVX512DQBW-SLOW-NEXT:    vmovdqa (%rdx), %ymm1
; AVX512DQBW-SLOW-NEXT:    vpshufhw {{.*#+}} ymm4 = ymm1[0,1,2,3,5,6,7,6,8,9,10,11,13,14,15,14]
; AVX512DQBW-SLOW-NEXT:    vpshufd {{.*#+}} ymm4 = ymm4[2,2,3,3,6,6,7,7]
; AVX512DQBW-SLOW-NEXT:    movl $693250386, %eax # imm = 0x29522952
; AVX512DQBW-SLOW-NEXT:    kmovd %eax, %k1
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %ymm4, %ymm2 {%k1}
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vmovdqa 32(%rdx), %xmm6
; AVX512DQBW-SLOW-NEXT:    vmovdqa 32(%rcx), %xmm12
; AVX512DQBW-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm4 = xmm12[0],xmm6[0],xmm12[1],xmm6[1],xmm12[2],xmm6[2],xmm12[3],xmm6[3],xmm12[4],xmm6[4],xmm12[5],xmm6[5],xmm12[6],xmm6[6],xmm12[7],xmm6[7]
; AVX512DQBW-SLOW-NEXT:    vmovdqa {{.*#+}} xmm7 = <2,u,1,0,5,4,u,3,u,7,6,11,10,u,9,8>
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm7, %xmm4, %xmm4
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[0,0,1,1]
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm4, %zmm2, %zmm10
; AVX512DQBW-SLOW-NEXT:    vmovdqa (%rsi), %ymm4
; AVX512DQBW-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm15 = [11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14,11,0,13,10,15,12,0,14]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm15, %ymm4, %ymm2
; AVX512DQBW-SLOW-NEXT:    vmovdqa (%rdi), %ymm5
; AVX512DQBW-SLOW-NEXT:    vpshufhw {{.*#+}} ymm9 = ymm5[0,1,2,3,6,5,6,7,8,9,10,11,14,13,14,15]
; AVX512DQBW-SLOW-NEXT:    vpshufd {{.*#+}} ymm9 = ymm9[2,2,3,3,6,6,7,7]
; AVX512DQBW-SLOW-NEXT:    movl $1251232404, %eax # imm = 0x4A944A94
; AVX512DQBW-SLOW-NEXT:    kmovd %eax, %k5
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %ymm9, %ymm2 {%k5}
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm2 = ymm2[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vmovdqa 32(%rsi), %xmm13
; AVX512DQBW-SLOW-NEXT:    vmovdqa 32(%rdi), %xmm14
; AVX512DQBW-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm11 = xmm14[0],xmm13[0],xmm14[1],xmm13[1],xmm14[2],xmm13[2],xmm14[3],xmm13[3],xmm14[4],xmm13[4],xmm14[5],xmm13[5],xmm14[6],xmm13[6],xmm14[7],xmm13[7]
; AVX512DQBW-SLOW-NEXT:    vmovdqa {{.*#+}} xmm9 = <0,1,4,5,u,2,3,6,7,10,11,u,8,9,12,13>
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm9, %xmm11, %xmm11
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm11 = ymm11[0,0,1,1]
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm11, %zmm2, %zmm2
; AVX512DQBW-SLOW-NEXT:    movabsq $1785168781326730801, %rax # imm = 0x18C6318C6318C631
; AVX512DQBW-SLOW-NEXT:    kmovq %rax, %k4
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm10, %zmm2 {%k4}
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 32(%r8), %ymm16
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm10 = [6,6,6,6,7,7,7,7,16,16,16,16,16,16,17,17]
; AVX512DQBW-SLOW-NEXT:    vpermi2d %zmm16, %zmm3, %zmm10
; AVX512DQBW-SLOW-NEXT:    movabsq $2380225041768974402, %rax # imm = 0x2108421084210842
; AVX512DQBW-SLOW-NEXT:    kmovq %rax, %k2
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm10, %zmm2 {%k2}
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 32(%rdx), %ymm23
; AVX512DQBW-SLOW-NEXT:    vmovdqa {{.*#+}} ymm10 = [128,128,12,13,128,128,128,128,14,128,128,128,14,15,128,128,128,128,16,128,128,128,16,17,128,128,128,128,18,128,128,128]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm10, %ymm23, %ymm17
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 32(%rcx), %ymm24
; AVX512DQBW-SLOW-NEXT:    vmovdqa {{.*#+}} ymm11 = [128,128,128,128,13,128,128,128,128,14,128,128,128,128,15,128,128,128,128,16,128,128,128,128,17,128,128,128,128,18,128,128]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm11, %ymm24, %ymm18
; AVX512DQBW-SLOW-NEXT:    vporq %ymm17, %ymm18, %ymm17
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} xmm20 = <128,6,128,8,u,128,7,128,9,128,11,u,128,10,128,12>
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm20, %xmm12, %xmm12
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} xmm22 = <6,128,8,128,u,7,128,9,128,11,128,u,10,128,12,128>
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm22, %xmm6, %xmm6
; AVX512DQBW-SLOW-NEXT:    vpor %xmm6, %xmm12, %xmm6
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm6 = ymm6[0,0,1,1]
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm17, %zmm6, %zmm6
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} xmm19 = <8,128,u,7,128,9,128,u,128,u,10,128,12,128,u,11>
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm19, %xmm14, %xmm12
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} xmm21 = <128,8,u,128,7,128,9,u,11,u,128,10,128,12,u,128>
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm21, %xmm13, %xmm13
; AVX512DQBW-SLOW-NEXT:    vpor %xmm12, %xmm13, %xmm12
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm13 = ymm12[0,0,1,1]
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 32(%rdi), %ymm25
; AVX512DQBW-SLOW-NEXT:    vmovdqa {{.*#+}} ymm12 = <3,3,3,u,4,4,4,4>
; AVX512DQBW-SLOW-NEXT:    vpermd %ymm25, %ymm12, %ymm17
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 32(%rsi), %ymm26
; AVX512DQBW-SLOW-NEXT:    vpbroadcastq {{.*#+}} ymm14 = [0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14,0,0,13,2,15,0,1,14]
; AVX512DQBW-SLOW-NEXT:    movl $138547332, %eax # imm = 0x8421084
; AVX512DQBW-SLOW-NEXT:    kmovd %eax, %k3
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm14, %ymm26, %ymm17 {%k3}
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm17, %zmm13, %zmm13
; AVX512DQBW-SLOW-NEXT:    movabsq $-8330787646191410408, %rax # imm = 0x8C6318C6318C6318
; AVX512DQBW-SLOW-NEXT:    kmovq %rax, %k2
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm13, %zmm6 {%k2}
; AVX512DQBW-SLOW-NEXT:    vmovdqa {{.*#+}} ymm13 = <3,3,3,3,u,4,4,4>
; AVX512DQBW-SLOW-NEXT:    vpermd %ymm16, %ymm13, %ymm17
; AVX512DQBW-SLOW-NEXT:    vpshufd {{.*#+}} xmm18 = mem[1,1,2,2]
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm18 = ymm18[0,1,1,1]
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm17, %zmm18, %zmm17
; AVX512DQBW-SLOW-NEXT:    movabsq $4760450083537948804, %rax # imm = 0x4210842108421084
; AVX512DQBW-SLOW-NEXT:    kmovq %rax, %k6
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm17, %zmm6 {%k6}
; AVX512DQBW-SLOW-NEXT:    vbroadcasti64x2 {{.*#+}} ymm17 = [19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128]
; AVX512DQBW-SLOW-NEXT:    # ymm17 = mem[0,1,0,1]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm17, %ymm26, %ymm18
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm27 = ymm18[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} ymm18 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,21,128,128,20,128,22,128,24,128,128,23,128,25,128,128]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm18, %ymm25, %ymm28
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm28 = ymm28[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vporq %ymm27, %ymm28, %ymm27
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm15, %ymm26, %ymm15
; AVX512DQBW-SLOW-NEXT:    vpshufhw {{.*#+}} ymm25 = ymm25[0,1,2,3,6,5,6,7,8,9,10,11,14,13,14,15]
; AVX512DQBW-SLOW-NEXT:    vpshufd {{.*#+}} ymm25 = ymm25[2,2,3,3,6,6,7,7]
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %ymm25, %ymm15 {%k5}
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm15 = ymm15[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm15, %zmm27, %zmm15
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} ymm25 = [128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,128,18,19,128,21,128,21,20,128,22,128,24,128,22,23,128,25]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm25, %ymm23, %ymm26
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm26 = ymm26[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vbroadcasti64x2 {{.*#+}} ymm27 = [128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128,128,128,19,128,21,128,128,20,128,22,128,24,128,128,23,128]
; AVX512DQBW-SLOW-NEXT:    # ymm27 = mem[0,1,0,1]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm27, %ymm24, %ymm28
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm28 = ymm28[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vporq %ymm26, %ymm28, %ymm26
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm8, %ymm24, %ymm8
; AVX512DQBW-SLOW-NEXT:    vpshufhw {{.*#+}} ymm23 = ymm23[0,1,2,3,5,6,7,6,8,9,10,11,13,14,15,14]
; AVX512DQBW-SLOW-NEXT:    vpshufd {{.*#+}} ymm23 = ymm23[2,2,3,3,6,6,7,7]
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %ymm23, %ymm8 {%k1}
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm8 = ymm8[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm8, %zmm26, %zmm8
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm15, %zmm8 {%k4}
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm15 = [4,6,5,5,5,5,4,6,6,6,6,6,7,7,7,7]
; AVX512DQBW-SLOW-NEXT:    vpermd %zmm16, %zmm15, %zmm15
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 (%rdx), %xmm16
; AVX512DQBW-SLOW-NEXT:    movabsq $-8925843906633654008, %rax # imm = 0x8421084210842108
; AVX512DQBW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm15, %zmm8 {%k1}
; AVX512DQBW-SLOW-NEXT:    vmovdqa (%rcx), %xmm15
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm20, %xmm15, %xmm20
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm22, %xmm16, %xmm22
; AVX512DQBW-SLOW-NEXT:    vporq %xmm20, %xmm22, %xmm20
; AVX512DQBW-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm15 = xmm15[0],xmm16[0],xmm15[1],xmm16[1],xmm15[2],xmm16[2],xmm15[3],xmm16[3],xmm15[4],xmm16[4],xmm15[5],xmm16[5],xmm15[6],xmm16[6],xmm15[7],xmm16[7]
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm7, %xmm15, %xmm7
; AVX512DQBW-SLOW-NEXT:    vmovdqa (%rsi), %xmm15
; AVX512DQBW-SLOW-NEXT:    vinserti32x4 $2, %xmm20, %zmm7, %zmm7
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 (%rdi), %xmm16
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm19, %xmm16, %xmm19
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm21, %xmm15, %xmm20
; AVX512DQBW-SLOW-NEXT:    vporq %xmm19, %xmm20, %xmm19
; AVX512DQBW-SLOW-NEXT:    vpunpcklbw {{.*#+}} xmm15 = xmm16[0],xmm15[0],xmm16[1],xmm15[1],xmm16[2],xmm15[2],xmm16[3],xmm15[3],xmm16[4],xmm15[4],xmm16[5],xmm15[5],xmm16[6],xmm15[6],xmm16[7],xmm15[7]
; AVX512DQBW-SLOW-NEXT:    vpshufb %xmm9, %xmm15, %xmm9
; AVX512DQBW-SLOW-NEXT:    vinserti32x4 $2, %xmm19, %zmm9, %zmm9
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} zmm7 = zmm7[0,0,1,1,4,4,5,5]
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} zmm9 = zmm9[0,0,1,1,4,4,5,5]
; AVX512DQBW-SLOW-NEXT:    movabsq $-4165393823095705204, %rax # imm = 0xC6318C6318C6318C
; AVX512DQBW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm7, %zmm9 {%k1}
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 {{.*#+}} zmm7 = [0,0,0,0,0,0,1,1,1,1,2,2,2,2,2,2]
; AVX512DQBW-SLOW-NEXT:    vpermd %zmm3, %zmm7, %zmm3
; AVX512DQBW-SLOW-NEXT:    movabsq $595056260442243600, %rax # imm = 0x842108421084210
; AVX512DQBW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm3, %zmm9 {%k1}
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm25, %ymm1, %ymm3
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm27, %ymm0, %ymm7
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm7 = ymm7[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vpor %ymm3, %ymm7, %ymm3
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm10, %ymm1, %ymm1
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm11, %ymm0, %ymm0
; AVX512DQBW-SLOW-NEXT:    vpor %ymm1, %ymm0, %ymm0
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm3, %zmm0, %zmm0
; AVX512DQBW-SLOW-NEXT:    vpermd %ymm5, %ymm12, %ymm1
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm14, %ymm4, %ymm1 {%k3}
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm17, %ymm4, %ymm3
; AVX512DQBW-SLOW-NEXT:    vpshufb %ymm18, %ymm5, %ymm4
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm3 = ymm3[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm4 = ymm4[2,2,3,3]
; AVX512DQBW-SLOW-NEXT:    vpor %ymm3, %ymm4, %ymm3
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm3, %zmm1, %zmm1
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm0, %zmm1 {%k2}
; AVX512DQBW-SLOW-NEXT:    vmovdqa (%r8), %ymm0
; AVX512DQBW-SLOW-NEXT:    vpermd %ymm0, %ymm13, %ymm3
; AVX512DQBW-SLOW-NEXT:    vpshufd {{.*#+}} ymm0 = ymm0[0,2,1,1,4,6,5,5]
; AVX512DQBW-SLOW-NEXT:    vpermq {{.*#+}} ymm0 = ymm0[2,3,3,2]
; AVX512DQBW-SLOW-NEXT:    vinserti64x4 $1, %ymm0, %zmm3, %zmm0
; AVX512DQBW-SLOW-NEXT:    movabsq $1190112520884487201, %rax # imm = 0x1084210842108421
; AVX512DQBW-SLOW-NEXT:    kmovq %rax, %k1
; AVX512DQBW-SLOW-NEXT:    vmovdqu8 %zmm0, %zmm1 {%k1}
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 %zmm1, 64(%r9)
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 %zmm9, (%r9)
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 %zmm8, 256(%r9)
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 %zmm6, 192(%r9)
; AVX512DQBW-SLOW-NEXT:    vmovdqa64 %zmm2, 128(%r9)
; AVX512DQBW-SLOW-NEXT:    vzeroupper
; AVX512DQBW-SLOW-NEXT:    retq
  %in.vec0 = load <64 x i8>, ptr %in.vecptr0, align 64
  %in.vec1 = load <64 x i8>, ptr %in.vecptr1, align 64
  %in.vec2 = load <64 x i8>, ptr %in.vecptr2, align 64
  %in.vec3 = load <64 x i8>, ptr %in.vecptr3, align 64
  %in.vec4 = load <64 x i8>, ptr %in.vecptr4, align 64
  %1 = shufflevector <64 x i8> %in.vec0, <64 x i8> %in.vec1, <128 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
  %2 = shufflevector <64 x i8> %in.vec2, <64 x i8> %in.vec3, <128 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127>
  %3 = shufflevector <128 x i8> %1, <128 x i8> %2, <256 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127, i32 128, i32 129, i32 130, i32 131, i32 132, i32 133, i32 134, i32 135, i32 136, i32 137, i32 138, i32 139, i32 140, i32 141, i32 142, i32 143, i32 144, i32 145, i32 146, i32 147, i32 148, i32 149, i32 150, i32 151, i32 152, i32 153, i32 154, i32 155, i32 156, i32 157, i32 158, i32 159, i32 160, i32 161, i32 162, i32 163, i32 164, i32 165, i32 166, i32 167, i32 168, i32 169, i32 170, i32 171, i32 172, i32 173, i32 174, i32 175, i32 176, i32 177, i32 178, i32 179, i32 180, i32 181, i32 182, i32 183, i32 184, i32 185, i32 186, i32 187, i32 188, i32 189, i32 190, i32 191, i32 192, i32 193, i32 194, i32 195, i32 196, i32 197, i32 198, i32 199, i32 200, i32 201, i32 202, i32 203, i32 204, i32 205, i32 206, i32 207, i32 208, i32 209, i32 210, i32 211, i32 212, i32 213, i32 214, i32 215, i32 216, i32 217, i32 218, i32 219, i32 220, i32 221, i32 222, i32 223, i32 224, i32 225, i32 226, i32 227, i32 228, i32 229, i32 230, i32 231, i32 232, i32 233, i32 234, i32 235, i32 236, i32 237, i32 238, i32 239, i32 240, i32 241, i32 242, i32 243, i32 244, i32 245, i32 246, i32 247, i32 248, i32 249, i32 250, i32 251, i32 252, i32 253, i32 254, i32 255>
  %4 = shufflevector <64 x i8> %in.vec4, <64 x i8> poison, <256 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, 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 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef, i32 undef>
  %5 = shufflevector <256 x i8> %3, <256 x i8> %4, <320 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 31, i32 32, i32 33, i32 34, i32 35, i32 36, i32 37, i32 38, i32 39, i32 40, i32 41, i32 42, i32 43, i32 44, i32 45, i32 46, i32 47, i32 48, i32 49, i32 50, i32 51, i32 52, i32 53, i32 54, i32 55, i32 56, i32 57, i32 58, i32 59, i32 60, i32 61, i32 62, i32 63, i32 64, i32 65, i32 66, i32 67, i32 68, i32 69, i32 70, i32 71, i32 72, i32 73, i32 74, i32 75, i32 76, i32 77, i32 78, i32 79, i32 80, i32 81, i32 82, i32 83, i32 84, i32 85, i32 86, i32 87, i32 88, i32 89, i32 90, i32 91, i32 92, i32 93, i32 94, i32 95, i32 96, i32 97, i32 98, i32 99, i32 100, i32 101, i32 102, i32 103, i32 104, i32 105, i32 106, i32 107, i32 108, i32 109, i32 110, i32 111, i32 112, i32 113, i32 114, i32 115, i32 116, i32 117, i32 118, i32 119, i32 120, i32 121, i32 122, i32 123, i32 124, i32 125, i32 126, i32 127, i32 128, i32 129, i32 130, i32 131, i32 132, i32 133, i32 134, i32 135, i32 136, i32 137, i32 138, i32 139, i32 140, i32 141, i32 142, i32 143, i32 144, i32 145, i32 146, i32 147, i32 148, i32 149, i32 150, i32 151, i32 152, i32 153, i32 154, i32 155, i32 156, i32 157, i32 158, i32 159, i32 160, i32 161, i32 162, i32 163, i32 164, i32 165, i32 166, i32 167, i32 168, i32 169, i32 170, i32 171, i32 172, i32 173, i32 174, i32 175, i32 176, i32 177, i32 178, i32 179, i32 180, i32 181, i32 182, i32 183, i32 184, i32 185, i32 186, i32 187, i32 188, i32 189, i32 190, i32 191, i32 192, i32 193, i32 194, i32 195, i32 196, i32 197, i32 198, i32 199, i32 200, i32 201, i32 202, i32 203, i32 204, i32 205, i32 206, i32 207, i32 208, i32 209, i32 210, i32 211, i32 212, i32 213, i32 214, i32 215, i32 216, i32 217, i32 218, i32 219, i32 220, i32 221, i32 222, i32 223, i32 224, i32 225, i32 226, i32 227, i32 228, i32 229, i32 230, i32 231, i32 232, i32 233, i32 234, i32 235, i32 236, i32 237, i32 238, i32 239, i32 240, i32 241, i32 242, i32 243, i32 244, i32 245, i32 246, i32 247, i32 248, i32 249, i32 250, i32 251, i32 252, i32 253, i32 254, i32 255, i32 256, i32 257, i32 258, i32 259, i32 260, i32 261, i32 262, i32 263, i32 264, i32 265, i32 266, i32 267, i32 268, i32 269, i32 270, i32 271, i32 272, i32 273, i32 274, i32 275, i32 276, i32 277, i32 278, i32 279, i32 280, i32 281, i32 282, i32 283, i32 284, i32 285, i32 286, i32 287, i32 288, i32 289, i32 290, i32 291, i32 292, i32 293, i32 294, i32 295, i32 296, i32 297, i32 298, i32 299, i32 300, i32 301, i32 302, i32 303, i32 304, i32 305, i32 306, i32 307, i32 308, i32 309, i32 310, i32 311, i32 312, i32 313, i32 314, i32 315, i32 316, i32 317, i32 318, i32 319>
  %interleaved.vec = shufflevector <320 x i8> %5, <320 x i8> poison, <320 x i32> <i32 0, i32 64, i32 128, i32 192, i32 256, i32 1, i32 65, i32 129, i32 193, i32 257, i32 2, i32 66, i32 130, i32 194, i32 258, i32 3, i32 67, i32 131, i32 195, i32 259, i32 4, i32 68, i32 132, i32 196, i32 260, i32 5, i32 69, i32 133, i32 197, i32 261, i32 6, i32 70, i32 134, i32 198, i32 262, i32 7, i32 71, i32 135, i32 199, i32 263, i32 8, i32 72, i32 136, i32 200, i32 264, i32 9, i32 73, i32 137, i32 201, i32 265, i32 10, i32 74, i32 138, i32 202, i32 266, i32 11, i32 75, i32 139, i32 203, i32 267, i32 12, i32 76, i32 140, i32 204, i32 268, i32 13, i32 77, i32 141, i32 205, i32 269, i32 14, i32 78, i32 142, i32 206, i32 270, i32 15, i32 79, i32 143, i32 207, i32 271, i32 16, i32 80, i32 144, i32 208, i32 272, i32 17, i32 81, i32 145, i32 209, i32 273, i32 18, i32 82, i32 146, i32 210, i32 274, i32 19, i32 83, i32 147, i32 211, i32 275, i32 20, i32 84, i32 148, i32 212, i32 276, i32 21, i32 85, i32 149, i32 213, i32 277, i32 22, i32 86, i32 150, i32 214, i32 278, i32 23, i32 87, i32 151, i32 215, i32 279, i32 24, i32 88, i32 152, i32 216, i32 280, i32 25, i32 89, i32 153, i32 217, i32 281, i32 26, i32 90, i32 154, i32 218, i32 282, i32 27, i32 91, i32 155, i32 219, i32 283, i32 28, i32 92, i32 156, i32 220, i32 284, i32 29, i32 93, i32 157, i32 221, i32 285, i32 30, i32 94, i32 158, i32 222, i32 286, i32 31, i32 95, i32 159, i32 223, i32 287, i32 32, i32 96, i32 160, i32 224, i32 288, i32 33, i32 97, i32 161, i32 225, i32 289, i32 34, i32 98, i32 162, i32 226, i32 290, i32 35, i32 99, i32 163, i32 227, i32 291, i32 36, i32 100, i32 164, i32 228, i32 292, i32 37, i32 101, i32 165, i32 229, i32 293, i32 38, i32 102, i32 166, i32 230, i32 294, i32 39, i32 103, i32 167, i32 231, i32 295, i32 40, i32 104, i32 168, i32 232, i32 296, i32 41, i32 105, i32 169, i32 233, i32 297, i32 42, i32 106, i32 170, i32 234, i32 298, i32 43, i32 107, i32 171, i32 235, i32 299, i32 44, i32 108, i32 172, i32 236, i32 300, i32 45, i32 109, i32 173, i32 237, i32 301, i32 46, i32 110, i32 174, i32 238, i32 302, i32 47, i32 111, i32 175, i32 239, i32 303, i32 48, i32 112, i32 176, i32 240, i32 304, i32 49, i32 113, i32 177, i32 241, i32 305, i32 50, i32 114, i32 178, i32 242, i32 306, i32 51, i32 115, i32 179, i32 243, i32 307, i32 52, i32 116, i32 180, i32 244, i32 308, i32 53, i32 117, i32 181, i32 245, i32 309, i32 54, i32 118, i32 182, i32 246, i32 310, i32 55, i32 119, i32 183, i32 247, i32 311, i32 56, i32 120, i32 184, i32 248, i32 312, i32 57, i32 121, i32 185, i32 249, i32 313, i32 58, i32 122, i32 186, i32 250, i32 314, i32 59, i32 123, i32 187, i32 251, i32 315, i32 60, i32 124, i32 188, i32 252, i32 316, i32 61, i32 125, i32 189, i32 253, i32 317, i32 62, i32 126, i32 190, i32 254, i32 318, i32 63, i32 127, i32 191, i32 255, i32 319>
  store <320 x i8> %interleaved.vec, ptr %out.vec, align 64
  ret void
}
;; NOTE: These prefixes are unused and the list is autogenerated. Do not add tests below this line:
; AVX1: {{.*}}
; AVX2: {{.*}}
; AVX512: {{.*}}
; AVX512-FAST: {{.*}}
; AVX512-SLOW: {{.*}}
; AVX512BW-ONLY-FAST: {{.*}}
; AVX512DQ-FAST: {{.*}}
; AVX512DQ-SLOW: {{.*}}
; AVX512DQBW-FAST: {{.*}}
; AVX512F-ONLY-FAST: {{.*}}
; AVX512F-ONLY-SLOW: {{.*}}
; FALLBACK0: {{.*}}
; FALLBACK1: {{.*}}
; FALLBACK10: {{.*}}
; FALLBACK11: {{.*}}
; FALLBACK12: {{.*}}
; FALLBACK2: {{.*}}
; FALLBACK3: {{.*}}
; FALLBACK4: {{.*}}
; FALLBACK5: {{.*}}
; FALLBACK6: {{.*}}
; FALLBACK7: {{.*}}
; FALLBACK8: {{.*}}
; FALLBACK9: {{.*}}