From 02744294cd79706f195351560ce830bf023f9a0b Mon Sep 17 00:00:00 2001 From: Michael Zuckerman Date: Sat, 30 Sep 2017 14:55:03 +0000 Subject: [PATCH] Code refactoring for the interleaved code Change-Id: I7831c9febad8e14278a5bc87584a0053dc837be1 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@314596 91177308-0d34-0410-b5e6-96231b3b80d8 --- lib/Target/X86/X86InterleavedAccess.cpp | 46 ++++----- test/CodeGen/X86/x86-interleaved-access.ll | 114 ++++++++++----------- .../InterleavedAccess/X86/interleavedStore.ll | 28 ++--- 3 files changed, 89 insertions(+), 99 deletions(-) diff --git a/lib/Target/X86/X86InterleavedAccess.cpp b/lib/Target/X86/X86InterleavedAccess.cpp index e93ec274ee9..b0c3bb28f72 100644 --- a/lib/Target/X86/X86InterleavedAccess.cpp +++ b/lib/Target/X86/X86InterleavedAccess.cpp @@ -279,6 +279,7 @@ void X86InterleavedAccessGroup::interleave8bitStride4( TransposedMatrix.resize(4); SmallVector MaskHigh; SmallVector MaskLow; + SmallVector LowHighMask[2]; SmallVector MaskHighTemp1; SmallVector MaskLowTemp1; SmallVector MaskHighWord; @@ -289,50 +290,40 @@ void X86InterleavedAccessGroup::interleave8bitStride4( // MaskHighTemp and MaskLowTemp built in the vpunpckhbw and vpunpcklbw X86 // shuffle pattern. - createUnpackShuffleMask(VT, MaskHigh, false, false); createUnpackShuffleMask(VT, MaskLow, true, false); + createUnpackShuffleMask(VT, MaskHigh, false, false); // MaskHighTemp1 and MaskLowTemp1 built in the vpunpckhdw and vpunpckldw X86 // shuffle pattern. createUnpackShuffleMask(HalfVT, MaskLowTemp1, true, false); createUnpackShuffleMask(HalfVT, MaskHighTemp1, false, false); - scaleShuffleMask(2, MaskHighTemp1, MaskHighWord); - scaleShuffleMask(2, MaskLowTemp1, MaskLowWord); + scaleShuffleMask(2, MaskLowTemp1, LowHighMask[0]); + scaleShuffleMask(2, MaskHighTemp1, LowHighMask[1]); // IntrVec1Low = c0 m0 c1 m1 ... c7 m7 | c16 m16 c17 m17 ... c23 m23 // IntrVec1High = c8 m8 c9 m9 ... c15 m15 | c24 m24 c25 m25 ... c31 m31 // IntrVec2Low = y0 k0 y1 k1 ... y7 k7 | y16 k16 y17 k17 ... y23 k23 // IntrVec2High = y8 k8 y9 k9 ... y15 k15 | y24 k24 y25 k25 ... y31 k31 + Value *IntrVec[4]; - Value *IntrVec1Low = - Builder.CreateShuffleVector(Matrix[0], Matrix[1], MaskLow); - Value *IntrVec1High = - Builder.CreateShuffleVector(Matrix[0], Matrix[1], MaskHigh); - Value *IntrVec2Low = - Builder.CreateShuffleVector(Matrix[2], Matrix[3], MaskLow); - Value *IntrVec2High = - Builder.CreateShuffleVector(Matrix[2], Matrix[3], MaskHigh); + IntrVec[0] = Builder.CreateShuffleVector(Matrix[0], Matrix[1], MaskLow); + IntrVec[1] = Builder.CreateShuffleVector(Matrix[0], Matrix[1], MaskHigh); + IntrVec[2] = Builder.CreateShuffleVector(Matrix[2], Matrix[3], MaskLow); + IntrVec[3] = Builder.CreateShuffleVector(Matrix[2], Matrix[3], MaskHigh); // cmyk4 cmyk5 cmyk6 cmyk7 | cmyk20 cmyk21 cmyk22 cmyk23 // cmyk12 cmyk13 cmyk14 cmyk15 | cmyk28 cmyk29 cmyk30 cmyk31 // cmyk0 cmyk1 cmyk2 cmyk3 | cmyk16 cmyk17 cmyk18 cmyk19 // cmyk8 cmyk9 cmyk10 cmyk11 | cmyk24 cmyk25 cmyk26 cmyk27 - Value *High = - Builder.CreateShuffleVector(IntrVec1Low, IntrVec2Low, MaskHighWord); - Value *High1 = - Builder.CreateShuffleVector(IntrVec1High, IntrVec2High, MaskHighWord); - Value *Low = - Builder.CreateShuffleVector(IntrVec1Low, IntrVec2Low, MaskLowWord); - Value *Low1 = - Builder.CreateShuffleVector(IntrVec1High, IntrVec2High, MaskLowWord); + Value *VecOut[4]; + for (int i = 0; i < 4; i++) + VecOut[i] = Builder.CreateShuffleVector(IntrVec[i / 2], IntrVec[i / 2 + 2], + LowHighMask[i % 2]); if (VT == MVT::v16i8) { - TransposedMatrix[0] = Low; - TransposedMatrix[1] = High; - TransposedMatrix[2] = Low1; - TransposedMatrix[3] = High1; + std::copy(VecOut, VecOut + 4, TransposedMatrix.begin()); return; } @@ -343,14 +334,13 @@ void X86InterleavedAccessGroup::interleave8bitStride4( // ConcatHigh and ConcatLow built in the vperm2i128 and vinserti128 X86 // shuffle pattern. - SmallVector ConcatHigh12, ConcatHigh13; createConcatShuffleMask(numberOfElement, ConcatLow, true); createConcatShuffleMask(numberOfElement, ConcatHigh, false); - TransposedMatrix[0] = Builder.CreateShuffleVector(Low, High, ConcatLow); - TransposedMatrix[1] = Builder.CreateShuffleVector(Low1, High1, ConcatLow); - TransposedMatrix[2] = Builder.CreateShuffleVector(Low, High, ConcatHigh); - TransposedMatrix[3] = Builder.CreateShuffleVector(Low1, High1, ConcatHigh); + TransposedMatrix[0] = Builder.CreateShuffleVector(VecOut[0], VecOut[1], ConcatLow); + TransposedMatrix[1] = Builder.CreateShuffleVector(VecOut[2], VecOut[3], ConcatLow); + TransposedMatrix[2] = Builder.CreateShuffleVector(VecOut[0], VecOut[1], ConcatHigh); + TransposedMatrix[3] = Builder.CreateShuffleVector(VecOut[2], VecOut[3], ConcatHigh); } // createShuffleStride returns shuffle mask of size N. diff --git a/test/CodeGen/X86/x86-interleaved-access.ll b/test/CodeGen/X86/x86-interleaved-access.ll index c703337a2c1..d4116d4153b 100644 --- a/test/CodeGen/X86/x86-interleaved-access.ll +++ b/test/CodeGen/X86/x86-interleaved-access.ll @@ -289,38 +289,38 @@ define void @store_factori64_4(<16 x i64>* %ptr, <4 x i64> %v0, <4 x i64> %v1, < define void @interleaved_store_vf32_i8_stride4(<32 x i8> %x1, <32 x i8> %x2, <32 x i8> %x3, <32 x i8> %x4, <128 x i8>* %p) { ; AVX1-LABEL: interleaved_store_vf32_i8_stride4: ; AVX1: # BB#0: -; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm9 = 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] +; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm4 = 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] ; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm5 ; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm6 ; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm7 = xmm6[0],xmm5[0],xmm6[1],xmm5[1],xmm6[2],xmm5[2],xmm6[3],xmm5[3],xmm6[4],xmm5[4],xmm6[5],xmm5[5],xmm6[6],xmm5[6],xmm6[7],xmm5[7] ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm8 = 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-NEXT: vpunpckhbw {{.*#+}} xmm1 = xmm6[8],xmm5[8],xmm6[9],xmm5[9],xmm6[10],xmm5[10],xmm6[11],xmm5[11],xmm6[12],xmm5[12],xmm6[13],xmm5[13],xmm6[14],xmm5[14],xmm6[15],xmm5[15] +; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm9 = xmm6[8],xmm5[8],xmm6[9],xmm5[9],xmm6[10],xmm5[10],xmm6[11],xmm5[11],xmm6[12],xmm5[12],xmm6[13],xmm5[13],xmm6[14],xmm5[14],xmm6[15],xmm5[15] ; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm5 = 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-NEXT: vextractf128 $1, %ymm3, %xmm6 ; AVX1-NEXT: vextractf128 $1, %ymm2, %xmm0 -; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm4 = xmm0[0],xmm6[0],xmm0[1],xmm6[1],xmm0[2],xmm6[2],xmm0[3],xmm6[3],xmm0[4],xmm6[4],xmm0[5],xmm6[5],xmm0[6],xmm6[6],xmm0[7],xmm6[7] +; AVX1-NEXT: vpunpcklbw {{.*#+}} xmm1 = xmm0[0],xmm6[0],xmm0[1],xmm6[1],xmm0[2],xmm6[2],xmm0[3],xmm6[3],xmm0[4],xmm6[4],xmm0[5],xmm6[5],xmm0[6],xmm6[6],xmm0[7],xmm6[7] ; AVX1-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] ; AVX1-NEXT: vpunpckhbw {{.*#+}} xmm0 = xmm0[8],xmm6[8],xmm0[9],xmm6[9],xmm0[10],xmm6[10],xmm0[11],xmm6[11],xmm0[12],xmm6[12],xmm0[13],xmm6[13],xmm0[14],xmm6[14],xmm0[15],xmm6[15] -; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm3 = xmm7[4],xmm4[4],xmm7[5],xmm4[5],xmm7[6],xmm4[6],xmm7[7],xmm4[7] -; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm6 = xmm9[4],xmm5[4],xmm9[5],xmm5[5],xmm9[6],xmm5[6],xmm9[7],xmm5[7] -; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm6, %ymm10 -; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm3 = xmm1[4],xmm0[4],xmm1[5],xmm0[5],xmm1[6],xmm0[6],xmm1[7],xmm0[7] -; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm11 = xmm8[4],xmm2[4],xmm8[5],xmm2[5],xmm8[6],xmm2[6],xmm8[7],xmm2[7] -; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm11, %ymm3 -; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm4 = xmm7[0],xmm4[0],xmm7[1],xmm4[1],xmm7[2],xmm4[2],xmm7[3],xmm4[3] -; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm5 = xmm9[0],xmm5[0],xmm9[1],xmm5[1],xmm9[2],xmm5[2],xmm9[3],xmm5[3] -; AVX1-NEXT: vinsertf128 $1, %xmm4, %ymm5, %ymm4 -; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] -; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm8[0],xmm2[0],xmm8[1],xmm2[1],xmm8[2],xmm2[2],xmm8[3],xmm2[3] -; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0 -; AVX1-NEXT: vinsertf128 $1, %xmm6, %ymm4, %ymm1 -; AVX1-NEXT: vinsertf128 $1, %xmm11, %ymm0, %ymm2 -; AVX1-NEXT: vperm2f128 {{.*#+}} ymm4 = ymm4[2,3],ymm10[2,3] -; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm0[2,3],ymm3[2,3] +; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm3 = xmm7[0],xmm1[0],xmm7[1],xmm1[1],xmm7[2],xmm1[2],xmm7[3],xmm1[3] +; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm6 = xmm4[0],xmm5[0],xmm4[1],xmm5[1],xmm4[2],xmm5[2],xmm4[3],xmm5[3] +; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm6, %ymm3 +; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm7[4],xmm1[4],xmm7[5],xmm1[5],xmm7[6],xmm1[6],xmm7[7],xmm1[7] +; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm4 = xmm4[4],xmm5[4],xmm4[5],xmm5[5],xmm4[6],xmm5[6],xmm4[7],xmm5[7] +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm4, %ymm1 +; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm5 = xmm9[0],xmm0[0],xmm9[1],xmm0[1],xmm9[2],xmm0[2],xmm9[3],xmm0[3] +; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm6 = xmm8[0],xmm2[0],xmm8[1],xmm2[1],xmm8[2],xmm2[2],xmm8[3],xmm2[3] +; AVX1-NEXT: vinsertf128 $1, %xmm5, %ymm6, %ymm5 +; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm9[4],xmm0[4],xmm9[5],xmm0[5],xmm9[6],xmm0[6],xmm9[7],xmm0[7] +; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm2 = xmm8[4],xmm2[4],xmm8[5],xmm2[5],xmm8[6],xmm2[6],xmm8[7],xmm2[7] +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm2, %ymm0 +; AVX1-NEXT: vinsertf128 $1, %xmm4, %ymm3, %ymm4 +; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm5, %ymm2 +; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm3[2,3],ymm1[2,3] +; AVX1-NEXT: vperm2f128 {{.*#+}} ymm0 = ymm5[2,3],ymm0[2,3] ; AVX1-NEXT: vmovaps %ymm0, 96(%rdi) -; AVX1-NEXT: vmovaps %ymm4, 64(%rdi) +; AVX1-NEXT: vmovaps %ymm1, 64(%rdi) ; AVX1-NEXT: vmovaps %ymm2, 32(%rdi) -; AVX1-NEXT: vmovaps %ymm1, (%rdi) +; AVX1-NEXT: vmovaps %ymm4, (%rdi) ; AVX1-NEXT: vzeroupper ; AVX1-NEXT: retq ; @@ -330,17 +330,17 @@ define void @interleaved_store_vf32_i8_stride4(<32 x i8> %x1, <32 x i8> %x2, <32 ; AVX2-NEXT: vpunpckhbw {{.*#+}} ymm0 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31] ; AVX2-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm2[0],ymm3[0],ymm2[1],ymm3[1],ymm2[2],ymm3[2],ymm2[3],ymm3[3],ymm2[4],ymm3[4],ymm2[5],ymm3[5],ymm2[6],ymm3[6],ymm2[7],ymm3[7],ymm2[16],ymm3[16],ymm2[17],ymm3[17],ymm2[18],ymm3[18],ymm2[19],ymm3[19],ymm2[20],ymm3[20],ymm2[21],ymm3[21],ymm2[22],ymm3[22],ymm2[23],ymm3[23] ; AVX2-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm2[8],ymm3[8],ymm2[9],ymm3[9],ymm2[10],ymm3[10],ymm2[11],ymm3[11],ymm2[12],ymm3[12],ymm2[13],ymm3[13],ymm2[14],ymm3[14],ymm2[15],ymm3[15],ymm2[24],ymm3[24],ymm2[25],ymm3[25],ymm2[26],ymm3[26],ymm2[27],ymm3[27],ymm2[28],ymm3[28],ymm2[29],ymm3[29],ymm2[30],ymm3[30],ymm2[31],ymm3[31] -; AVX2-NEXT: vpunpckhwd {{.*#+}} ymm3 = ymm4[4],ymm1[4],ymm4[5],ymm1[5],ymm4[6],ymm1[6],ymm4[7],ymm1[7],ymm4[12],ymm1[12],ymm4[13],ymm1[13],ymm4[14],ymm1[14],ymm4[15],ymm1[15] -; AVX2-NEXT: vpunpckhwd {{.*#+}} ymm5 = ymm0[4],ymm2[4],ymm0[5],ymm2[5],ymm0[6],ymm2[6],ymm0[7],ymm2[7],ymm0[12],ymm2[12],ymm0[13],ymm2[13],ymm0[14],ymm2[14],ymm0[15],ymm2[15] -; AVX2-NEXT: vpunpcklwd {{.*#+}} ymm1 = ymm4[0],ymm1[0],ymm4[1],ymm1[1],ymm4[2],ymm1[2],ymm4[3],ymm1[3],ymm4[8],ymm1[8],ymm4[9],ymm1[9],ymm4[10],ymm1[10],ymm4[11],ymm1[11] -; AVX2-NEXT: vpunpcklwd {{.*#+}} ymm0 = ymm0[0],ymm2[0],ymm0[1],ymm2[1],ymm0[2],ymm2[2],ymm0[3],ymm2[3],ymm0[8],ymm2[8],ymm0[9],ymm2[9],ymm0[10],ymm2[10],ymm0[11],ymm2[11] -; AVX2-NEXT: vinserti128 $1, %xmm3, %ymm1, %ymm2 -; AVX2-NEXT: vinserti128 $1, %xmm5, %ymm0, %ymm4 -; AVX2-NEXT: vperm2i128 {{.*#+}} ymm1 = ymm1[2,3],ymm3[2,3] -; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm5[2,3] +; AVX2-NEXT: vpunpcklwd {{.*#+}} ymm3 = ymm4[0],ymm1[0],ymm4[1],ymm1[1],ymm4[2],ymm1[2],ymm4[3],ymm1[3],ymm4[8],ymm1[8],ymm4[9],ymm1[9],ymm4[10],ymm1[10],ymm4[11],ymm1[11] +; AVX2-NEXT: vpunpckhwd {{.*#+}} ymm1 = ymm4[4],ymm1[4],ymm4[5],ymm1[5],ymm4[6],ymm1[6],ymm4[7],ymm1[7],ymm4[12],ymm1[12],ymm4[13],ymm1[13],ymm4[14],ymm1[14],ymm4[15],ymm1[15] +; AVX2-NEXT: vpunpcklwd {{.*#+}} ymm4 = ymm0[0],ymm2[0],ymm0[1],ymm2[1],ymm0[2],ymm2[2],ymm0[3],ymm2[3],ymm0[8],ymm2[8],ymm0[9],ymm2[9],ymm0[10],ymm2[10],ymm0[11],ymm2[11] +; AVX2-NEXT: vpunpckhwd {{.*#+}} ymm0 = ymm0[4],ymm2[4],ymm0[5],ymm2[5],ymm0[6],ymm2[6],ymm0[7],ymm2[7],ymm0[12],ymm2[12],ymm0[13],ymm2[13],ymm0[14],ymm2[14],ymm0[15],ymm2[15] +; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm3, %ymm2 +; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm4, %ymm5 +; AVX2-NEXT: vperm2i128 {{.*#+}} ymm1 = ymm3[2,3],ymm1[2,3] +; AVX2-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm4[2,3],ymm0[2,3] ; AVX2-NEXT: vmovdqa %ymm0, 96(%rdi) ; AVX2-NEXT: vmovdqa %ymm1, 64(%rdi) -; AVX2-NEXT: vmovdqa %ymm4, 32(%rdi) +; AVX2-NEXT: vmovdqa %ymm5, 32(%rdi) ; AVX2-NEXT: vmovdqa %ymm2, (%rdi) ; AVX2-NEXT: vzeroupper ; AVX2-NEXT: retq @@ -351,15 +351,15 @@ define void @interleaved_store_vf32_i8_stride4(<32 x i8> %x1, <32 x i8> %x2, <32 ; AVX512-NEXT: vpunpckhbw {{.*#+}} ymm0 = ymm0[8],ymm1[8],ymm0[9],ymm1[9],ymm0[10],ymm1[10],ymm0[11],ymm1[11],ymm0[12],ymm1[12],ymm0[13],ymm1[13],ymm0[14],ymm1[14],ymm0[15],ymm1[15],ymm0[24],ymm1[24],ymm0[25],ymm1[25],ymm0[26],ymm1[26],ymm0[27],ymm1[27],ymm0[28],ymm1[28],ymm0[29],ymm1[29],ymm0[30],ymm1[30],ymm0[31],ymm1[31] ; AVX512-NEXT: vpunpcklbw {{.*#+}} ymm1 = ymm2[0],ymm3[0],ymm2[1],ymm3[1],ymm2[2],ymm3[2],ymm2[3],ymm3[3],ymm2[4],ymm3[4],ymm2[5],ymm3[5],ymm2[6],ymm3[6],ymm2[7],ymm3[7],ymm2[16],ymm3[16],ymm2[17],ymm3[17],ymm2[18],ymm3[18],ymm2[19],ymm3[19],ymm2[20],ymm3[20],ymm2[21],ymm3[21],ymm2[22],ymm3[22],ymm2[23],ymm3[23] ; AVX512-NEXT: vpunpckhbw {{.*#+}} ymm2 = ymm2[8],ymm3[8],ymm2[9],ymm3[9],ymm2[10],ymm3[10],ymm2[11],ymm3[11],ymm2[12],ymm3[12],ymm2[13],ymm3[13],ymm2[14],ymm3[14],ymm2[15],ymm3[15],ymm2[24],ymm3[24],ymm2[25],ymm3[25],ymm2[26],ymm3[26],ymm2[27],ymm3[27],ymm2[28],ymm3[28],ymm2[29],ymm3[29],ymm2[30],ymm3[30],ymm2[31],ymm3[31] -; AVX512-NEXT: vpunpckhwd {{.*#+}} ymm3 = ymm4[4],ymm1[4],ymm4[5],ymm1[5],ymm4[6],ymm1[6],ymm4[7],ymm1[7],ymm4[12],ymm1[12],ymm4[13],ymm1[13],ymm4[14],ymm1[14],ymm4[15],ymm1[15] -; AVX512-NEXT: vpunpckhwd {{.*#+}} ymm5 = ymm0[4],ymm2[4],ymm0[5],ymm2[5],ymm0[6],ymm2[6],ymm0[7],ymm2[7],ymm0[12],ymm2[12],ymm0[13],ymm2[13],ymm0[14],ymm2[14],ymm0[15],ymm2[15] -; AVX512-NEXT: vpunpcklwd {{.*#+}} ymm1 = ymm4[0],ymm1[0],ymm4[1],ymm1[1],ymm4[2],ymm1[2],ymm4[3],ymm1[3],ymm4[8],ymm1[8],ymm4[9],ymm1[9],ymm4[10],ymm1[10],ymm4[11],ymm1[11] -; AVX512-NEXT: vpunpcklwd {{.*#+}} ymm0 = ymm0[0],ymm2[0],ymm0[1],ymm2[1],ymm0[2],ymm2[2],ymm0[3],ymm2[3],ymm0[8],ymm2[8],ymm0[9],ymm2[9],ymm0[10],ymm2[10],ymm0[11],ymm2[11] -; AVX512-NEXT: vinserti128 $1, %xmm3, %ymm1, %ymm2 -; AVX512-NEXT: vinserti128 $1, %xmm5, %ymm0, %ymm4 -; AVX512-NEXT: vperm2i128 {{.*#+}} ymm1 = ymm1[2,3],ymm3[2,3] -; AVX512-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm0[2,3],ymm5[2,3] -; AVX512-NEXT: vinserti64x4 $1, %ymm4, %zmm2, %zmm2 +; AVX512-NEXT: vpunpcklwd {{.*#+}} ymm3 = ymm4[0],ymm1[0],ymm4[1],ymm1[1],ymm4[2],ymm1[2],ymm4[3],ymm1[3],ymm4[8],ymm1[8],ymm4[9],ymm1[9],ymm4[10],ymm1[10],ymm4[11],ymm1[11] +; AVX512-NEXT: vpunpckhwd {{.*#+}} ymm1 = ymm4[4],ymm1[4],ymm4[5],ymm1[5],ymm4[6],ymm1[6],ymm4[7],ymm1[7],ymm4[12],ymm1[12],ymm4[13],ymm1[13],ymm4[14],ymm1[14],ymm4[15],ymm1[15] +; AVX512-NEXT: vpunpcklwd {{.*#+}} ymm4 = ymm0[0],ymm2[0],ymm0[1],ymm2[1],ymm0[2],ymm2[2],ymm0[3],ymm2[3],ymm0[8],ymm2[8],ymm0[9],ymm2[9],ymm0[10],ymm2[10],ymm0[11],ymm2[11] +; AVX512-NEXT: vpunpckhwd {{.*#+}} ymm0 = ymm0[4],ymm2[4],ymm0[5],ymm2[5],ymm0[6],ymm2[6],ymm0[7],ymm2[7],ymm0[12],ymm2[12],ymm0[13],ymm2[13],ymm0[14],ymm2[14],ymm0[15],ymm2[15] +; AVX512-NEXT: vinserti128 $1, %xmm1, %ymm3, %ymm2 +; AVX512-NEXT: vinserti128 $1, %xmm0, %ymm4, %ymm5 +; AVX512-NEXT: vperm2i128 {{.*#+}} ymm1 = ymm3[2,3],ymm1[2,3] +; AVX512-NEXT: vperm2i128 {{.*#+}} ymm0 = ymm4[2,3],ymm0[2,3] +; AVX512-NEXT: vinserti64x4 $1, %ymm5, %zmm2, %zmm2 ; AVX512-NEXT: vinserti64x4 $1, %ymm0, %zmm1, %zmm0 ; AVX512-NEXT: vmovdqa32 %zmm0, 64(%rdi) ; AVX512-NEXT: vmovdqa32 %zmm2, (%rdi) @@ -379,12 +379,12 @@ define void @interleaved_store_vf16_i8_stride4(<16 x i8> %x1, <16 x i8> %x2, <16 ; AVX1-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-NEXT: vpunpcklbw {{.*#+}} xmm1 = 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-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] -; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm3 = xmm4[4],xmm1[4],xmm4[5],xmm1[5],xmm4[6],xmm1[6],xmm4[7],xmm1[7] -; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm5 = xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] -; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm4[0],xmm1[0],xmm4[1],xmm1[1],xmm4[2],xmm1[2],xmm4[3],xmm1[3] -; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3] -; AVX1-NEXT: vinsertf128 $1, %xmm3, %ymm1, %ymm1 -; AVX1-NEXT: vinsertf128 $1, %xmm5, %ymm0, %ymm0 +; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm3 = xmm4[0],xmm1[0],xmm4[1],xmm1[1],xmm4[2],xmm1[2],xmm4[3],xmm1[3] +; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm4[4],xmm1[4],xmm4[5],xmm1[5],xmm4[6],xmm1[6],xmm4[7],xmm1[7] +; AVX1-NEXT: vpunpcklwd {{.*#+}} xmm4 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3] +; AVX1-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] +; AVX1-NEXT: vinsertf128 $1, %xmm1, %ymm3, %ymm1 +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm4, %ymm0 ; AVX1-NEXT: vmovaps %ymm0, 32(%rdi) ; AVX1-NEXT: vmovaps %ymm1, (%rdi) ; AVX1-NEXT: vzeroupper @@ -396,12 +396,12 @@ define void @interleaved_store_vf16_i8_stride4(<16 x i8> %x1, <16 x i8> %x2, <16 ; AVX2-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-NEXT: vpunpcklbw {{.*#+}} xmm1 = 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] ; AVX2-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-NEXT: vpunpckhwd {{.*#+}} xmm3 = xmm4[4],xmm1[4],xmm4[5],xmm1[5],xmm4[6],xmm1[6],xmm4[7],xmm1[7] -; AVX2-NEXT: vpunpckhwd {{.*#+}} xmm5 = xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] -; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm4[0],xmm1[0],xmm4[1],xmm1[1],xmm4[2],xmm1[2],xmm4[3],xmm1[3] -; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3] -; AVX2-NEXT: vinserti128 $1, %xmm3, %ymm1, %ymm1 -; AVX2-NEXT: vinserti128 $1, %xmm5, %ymm0, %ymm0 +; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm3 = xmm4[0],xmm1[0],xmm4[1],xmm1[1],xmm4[2],xmm1[2],xmm4[3],xmm1[3] +; AVX2-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm4[4],xmm1[4],xmm4[5],xmm1[5],xmm4[6],xmm1[6],xmm4[7],xmm1[7] +; AVX2-NEXT: vpunpcklwd {{.*#+}} xmm4 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3] +; AVX2-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] +; AVX2-NEXT: vinserti128 $1, %xmm1, %ymm3, %ymm1 +; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm4, %ymm0 ; AVX2-NEXT: vmovdqa %ymm0, 32(%rdi) ; AVX2-NEXT: vmovdqa %ymm1, (%rdi) ; AVX2-NEXT: vzeroupper @@ -413,12 +413,12 @@ define void @interleaved_store_vf16_i8_stride4(<16 x i8> %x1, <16 x i8> %x2, <16 ; AVX512-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] ; AVX512-NEXT: vpunpcklbw {{.*#+}} xmm1 = 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] ; AVX512-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] -; AVX512-NEXT: vpunpckhwd {{.*#+}} xmm3 = xmm4[4],xmm1[4],xmm4[5],xmm1[5],xmm4[6],xmm1[6],xmm4[7],xmm1[7] -; AVX512-NEXT: vpunpckhwd {{.*#+}} xmm5 = xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] -; AVX512-NEXT: vpunpcklwd {{.*#+}} xmm1 = xmm4[0],xmm1[0],xmm4[1],xmm1[1],xmm4[2],xmm1[2],xmm4[3],xmm1[3] -; AVX512-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3] -; AVX512-NEXT: vinserti128 $1, %xmm3, %ymm1, %ymm1 -; AVX512-NEXT: vinserti128 $1, %xmm5, %ymm0, %ymm0 +; AVX512-NEXT: vpunpcklwd {{.*#+}} xmm3 = xmm4[0],xmm1[0],xmm4[1],xmm1[1],xmm4[2],xmm1[2],xmm4[3],xmm1[3] +; AVX512-NEXT: vpunpckhwd {{.*#+}} xmm1 = xmm4[4],xmm1[4],xmm4[5],xmm1[5],xmm4[6],xmm1[6],xmm4[7],xmm1[7] +; AVX512-NEXT: vpunpcklwd {{.*#+}} xmm4 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3] +; AVX512-NEXT: vpunpckhwd {{.*#+}} xmm0 = xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] +; AVX512-NEXT: vinserti128 $1, %xmm1, %ymm3, %ymm1 +; AVX512-NEXT: vinserti128 $1, %xmm0, %ymm4, %ymm0 ; AVX512-NEXT: vinserti64x4 $1, %ymm0, %zmm1, %zmm0 ; AVX512-NEXT: vmovdqa32 %zmm0, (%rdi) ; AVX512-NEXT: vzeroupper diff --git a/test/Transforms/InterleavedAccess/X86/interleavedStore.ll b/test/Transforms/InterleavedAccess/X86/interleavedStore.ll index 0cb4f7f93e2..1492647f0ae 100644 --- a/test/Transforms/InterleavedAccess/X86/interleavedStore.ll +++ b/test/Transforms/InterleavedAccess/X86/interleavedStore.ll @@ -13,14 +13,14 @@ define void @interleaved_store_vf32_i8_stride4(<32 x i8> %x1, <32 x i8> %x2, <32 ; CHECK-NEXT: [[TMP6:%.*]] = shufflevector <32 x i8> [[TMP1]], <32 x i8> [[TMP2]], <32 x i32> ; CHECK-NEXT: [[TMP7:%.*]] = shufflevector <32 x i8> [[TMP3]], <32 x i8> [[TMP4]], <32 x i32> ; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <32 x i8> [[TMP3]], <32 x i8> [[TMP4]], <32 x i32> -; CHECK-NEXT: [[TMP9:%.*]] = shufflevector <32 x i8> [[TMP5]], <32 x i8> [[TMP7]], <32 x i32> -; CHECK-NEXT: [[TMP10:%.*]] = shufflevector <32 x i8> [[TMP6]], <32 x i8> [[TMP8]], <32 x i32> -; CHECK-NEXT: [[TMP11:%.*]] = shufflevector <32 x i8> [[TMP5]], <32 x i8> [[TMP7]], <32 x i32> -; CHECK-NEXT: [[TMP12:%.*]] = shufflevector <32 x i8> [[TMP6]], <32 x i8> [[TMP8]], <32 x i32> -; CHECK-NEXT: [[TMP13:%.*]] = shufflevector <32 x i8> [[TMP11]], <32 x i8> [[TMP9]], <32 x i32> -; CHECK-NEXT: [[TMP14:%.*]] = shufflevector <32 x i8> [[TMP12]], <32 x i8> [[TMP10]], <32 x i32> -; CHECK-NEXT: [[TMP15:%.*]] = shufflevector <32 x i8> [[TMP11]], <32 x i8> [[TMP9]], <32 x i32> -; CHECK-NEXT: [[TMP16:%.*]] = shufflevector <32 x i8> [[TMP12]], <32 x i8> [[TMP10]], <32 x i32> +; CHECK-NEXT: [[TMP9:%.*]] = shufflevector <32 x i8> [[TMP5]], <32 x i8> [[TMP7]], <32 x i32> +; CHECK-NEXT: [[TMP10:%.*]] = shufflevector <32 x i8> [[TMP5]], <32 x i8> [[TMP7]], <32 x i32> +; CHECK-NEXT: [[TMP11:%.*]] = shufflevector <32 x i8> [[TMP6]], <32 x i8> [[TMP8]], <32 x i32> +; CHECK-NEXT: [[TMP12:%.*]] = shufflevector <32 x i8> [[TMP6]], <32 x i8> [[TMP8]], <32 x i32> +; CHECK-NEXT: [[TMP13:%.*]] = shufflevector <32 x i8> [[TMP9]], <32 x i8> [[TMP10]], <32 x i32> +; CHECK-NEXT: [[TMP14:%.*]] = shufflevector <32 x i8> [[TMP11]], <32 x i8> [[TMP12]], <32 x i32> +; CHECK-NEXT: [[TMP15:%.*]] = shufflevector <32 x i8> [[TMP9]], <32 x i8> [[TMP10]], <32 x i32> +; CHECK-NEXT: [[TMP16:%.*]] = shufflevector <32 x i8> [[TMP11]], <32 x i8> [[TMP12]], <32 x i32> ; CHECK-NEXT: [[TMP17:%.*]] = shufflevector <32 x i8> [[TMP13]], <32 x i8> [[TMP14]], <64 x i32> ; CHECK-NEXT: [[TMP18:%.*]] = shufflevector <32 x i8> [[TMP15]], <32 x i8> [[TMP16]], <64 x i32> ; CHECK-NEXT: [[TMP19:%.*]] = shufflevector <64 x i8> [[TMP17]], <64 x i8> [[TMP18]], <128 x i32> @@ -46,12 +46,12 @@ define void @interleaved_store_vf16_i8_stride4(<16 x i8> %x1, <16 x i8> %x2, <16 ; CHECK-NEXT: [[TMP6:%.*]] = shufflevector <16 x i8> [[TMP1]], <16 x i8> [[TMP2]], <16 x i32> ; CHECK-NEXT: [[TMP7:%.*]] = shufflevector <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i32> ; CHECK-NEXT: [[TMP8:%.*]] = shufflevector <16 x i8> [[TMP3]], <16 x i8> [[TMP4]], <16 x i32> -; CHECK-NEXT: [[TMP9:%.*]] = shufflevector <16 x i8> [[TMP5]], <16 x i8> [[TMP7]], <16 x i32> -; CHECK-NEXT: [[TMP10:%.*]] = shufflevector <16 x i8> [[TMP6]], <16 x i8> [[TMP8]], <16 x i32> -; CHECK-NEXT: [[TMP11:%.*]] = shufflevector <16 x i8> [[TMP5]], <16 x i8> [[TMP7]], <16 x i32> -; CHECK-NEXT: [[TMP12:%.*]] = shufflevector <16 x i8> [[TMP6]], <16 x i8> [[TMP8]], <16 x i32> -; CHECK-NEXT: [[TMP13:%.*]] = shufflevector <16 x i8> [[TMP11]], <16 x i8> [[TMP9]], <32 x i32> -; CHECK-NEXT: [[TMP14:%.*]] = shufflevector <16 x i8> [[TMP12]], <16 x i8> [[TMP10]], <32 x i32> +; CHECK-NEXT: [[TMP9:%.*]] = shufflevector <16 x i8> [[TMP5]], <16 x i8> [[TMP7]], <16 x i32> +; CHECK-NEXT: [[TMP10:%.*]] = shufflevector <16 x i8> [[TMP5]], <16 x i8> [[TMP7]], <16 x i32> +; CHECK-NEXT: [[TMP11:%.*]] = shufflevector <16 x i8> [[TMP6]], <16 x i8> [[TMP8]], <16 x i32> +; CHECK-NEXT: [[TMP12:%.*]] = shufflevector <16 x i8> [[TMP6]], <16 x i8> [[TMP8]], <16 x i32> +; CHECK-NEXT: [[TMP13:%.*]] = shufflevector <16 x i8> [[TMP9]], <16 x i8> [[TMP10]], <32 x i32> +; CHECK-NEXT: [[TMP14:%.*]] = shufflevector <16 x i8> [[TMP11]], <16 x i8> [[TMP12]], <32 x i32> ; CHECK-NEXT: [[TMP15:%.*]] = shufflevector <32 x i8> [[TMP13]], <32 x i8> [[TMP14]], <64 x i32> ; CHECK-NEXT: store <64 x i8> [[TMP15]], <64 x i8>* [[P:%.*]] ; CHECK-NEXT: ret void -- 2.11.0