From d2b14b296c3ef1f67dd1738dda4a520d80d1461f Mon Sep 17 00:00:00 2001 From: Chandler Carruth Date: Fri, 20 Feb 2015 03:05:47 +0000 Subject: [PATCH] [x86] Make the new vector shuffle legality test on by default, which reflects the fact that the x86 backend can in fact lower any shuffle you want it to with reasonably high code quality. My recent work on the new vector shuffle has made this regress *very* little. The diff in the test cases makes me very, very happy. llvm-svn: 229958 --- llvm/lib/Target/X86/X86ISelLowering.cpp | 2 +- llvm/test/CodeGen/X86/bswap-vector.ll | 287 ++++++++---------------- llvm/test/CodeGen/X86/mmx-arg-passing-x86-64.ll | 8 - llvm/test/CodeGen/X86/sse3.ll | 9 +- llvm/test/CodeGen/X86/sse41.ll | 74 +++--- llvm/test/CodeGen/X86/vector-blend.ll | 59 +++-- 6 files changed, 158 insertions(+), 281 deletions(-) diff --git a/llvm/lib/Target/X86/X86ISelLowering.cpp b/llvm/lib/Target/X86/X86ISelLowering.cpp index 261645c..bbc2e69 100644 --- a/llvm/lib/Target/X86/X86ISelLowering.cpp +++ b/llvm/lib/Target/X86/X86ISelLowering.cpp @@ -73,7 +73,7 @@ static cl::opt ExperimentalVectorShuffleLowering( cl::Hidden); static cl::opt ExperimentalVectorShuffleLegality( - "x86-experimental-vector-shuffle-legality", cl::init(false), + "x86-experimental-vector-shuffle-legality", cl::init(true), cl::desc("Enable experimental shuffle legality based on the experimental " "shuffle lowering. Should only be used with the experimental " "shuffle lowering."), diff --git a/llvm/test/CodeGen/X86/bswap-vector.ll b/llvm/test/CodeGen/X86/bswap-vector.ll index 0351b22..7d5f380 100644 --- a/llvm/test/CodeGen/X86/bswap-vector.ll +++ b/llvm/test/CodeGen/X86/bswap-vector.ll @@ -13,37 +13,15 @@ declare <2 x i64> @llvm.bswap.v2i64(<2 x i64>) define <8 x i16> @test1(<8 x i16> %v) { ; CHECK-NOSSSE3-LABEL: test1: ; CHECK-NOSSSE3: # BB#0: # %entry -; CHECK-NOSSSE3-NEXT: pextrw $7, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm1 -; CHECK-NOSSSE3-NEXT: pextrw $3, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1],xmm2[2],xmm1[2],xmm2[3],xmm1[3] -; CHECK-NOSSSE3-NEXT: pextrw $5, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: pextrw $1, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm1 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3] -; CHECK-NOSSSE3-NEXT: pextrw $6, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: pextrw $2, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3] -; CHECK-NOSSSE3-NEXT: pextrw $4, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: movd %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm0 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3] +; CHECK-NOSSSE3-NEXT: pxor %xmm1, %xmm1 +; CHECK-NOSSSE3-NEXT: movdqa %xmm0, %xmm2 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm2 = 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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[1,0,3,2,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,5,4,7,6] +; CHECK-NOSSSE3-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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,0,3,2,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,4,7,6] +; CHECK-NOSSSE3-NEXT: packuswb %xmm2, %xmm0 ; CHECK-NOSSSE3-NEXT: retq ; ; CHECK-SSSE3-LABEL: test1: @@ -68,25 +46,15 @@ entry: define <4 x i32> @test2(<4 x i32> %v) { ; CHECK-NOSSSE3-LABEL: test2: ; CHECK-NOSSSE3: # BB#0: # %entry -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[3,1,2,3] -; CHECK-NOSSSE3-NEXT: movd %xmm1, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm1 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,2,3] -; CHECK-NOSSSE3-NEXT: movd %xmm2, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1] -; CHECK-NOSSSE3-NEXT: movd %xmm0, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm1 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1] -; CHECK-NOSSSE3-NEXT: movd %xmm0, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm0 -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1] -; CHECK-NOSSSE3-NEXT: movdqa %xmm1, %xmm0 +; CHECK-NOSSSE3-NEXT: pxor %xmm1, %xmm1 +; CHECK-NOSSSE3-NEXT: movdqa %xmm0, %xmm2 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm2 = 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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: packuswb %xmm2, %xmm0 ; CHECK-NOSSSE3-NEXT: retq ; ; CHECK-SSSE3-LABEL: test2: @@ -111,15 +79,17 @@ entry: define <2 x i64> @test3(<2 x i64> %v) { ; CHECK-NOSSSE3-LABEL: test3: ; CHECK-NOSSSE3: # BB#0: # %entry -; CHECK-NOSSSE3-NEXT: movd %xmm0, %rax -; CHECK-NOSSSE3-NEXT: bswapq %rax -; CHECK-NOSSSE3-NEXT: movd %rax, %xmm1 +; CHECK-NOSSSE3-NEXT: pxor %xmm1, %xmm1 +; CHECK-NOSSSE3-NEXT: movdqa %xmm0, %xmm2 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm2 = 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] +; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-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] ; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1] -; CHECK-NOSSSE3-NEXT: movd %xmm0, %rax -; CHECK-NOSSSE3-NEXT: bswapq %rax -; CHECK-NOSSSE3-NEXT: movd %rax, %xmm0 -; CHECK-NOSSSE3-NEXT: punpcklqdq {{.*#+}} xmm1 = xmm1[0],xmm0[0] -; CHECK-NOSSSE3-NEXT: movdqa %xmm1, %xmm0 +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: packuswb %xmm2, %xmm0 ; CHECK-NOSSSE3-NEXT: retq ; ; CHECK-SSSE3-LABEL: test3: @@ -148,68 +118,23 @@ declare <4 x i64> @llvm.bswap.v4i64(<4 x i64>) define <16 x i16> @test4(<16 x i16> %v) { ; CHECK-NOSSSE3-LABEL: test4: ; CHECK-NOSSSE3: # BB#0: # %entry -; CHECK-NOSSSE3-NEXT: pextrw $7, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: pextrw $3, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3] -; CHECK-NOSSSE3-NEXT: pextrw $5, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm4 -; CHECK-NOSSSE3-NEXT: pextrw $1, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm4[0],xmm2[1],xmm4[1],xmm2[2],xmm4[2],xmm2[3],xmm4[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3] -; CHECK-NOSSSE3-NEXT: pextrw $6, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: pextrw $2, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm4 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm4 = xmm4[0],xmm3[0],xmm4[1],xmm3[1],xmm4[2],xmm3[2],xmm4[3],xmm3[3] -; CHECK-NOSSSE3-NEXT: pextrw $4, %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: movd %xmm0, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm0 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1],xmm0[2],xmm3[2],xmm0[3],xmm3[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm4[0],xmm0[1],xmm4[1],xmm0[2],xmm4[2],xmm0[3],xmm4[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3] -; CHECK-NOSSSE3-NEXT: pextrw $7, %xmm1, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: pextrw $3, %xmm1, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1],xmm3[2],xmm2[2],xmm3[3],xmm2[3] -; CHECK-NOSSSE3-NEXT: pextrw $5, %xmm1, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm4 -; CHECK-NOSSSE3-NEXT: pextrw $1, %xmm1, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm4[0],xmm2[1],xmm4[1],xmm2[2],xmm4[2],xmm2[3],xmm4[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1],xmm2[2],xmm3[2],xmm2[3],xmm3[3] -; CHECK-NOSSSE3-NEXT: pextrw $6, %xmm1, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: pextrw $2, %xmm1, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm4 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm4 = xmm4[0],xmm3[0],xmm4[1],xmm3[1],xmm4[2],xmm3[2],xmm4[3],xmm3[3] -; CHECK-NOSSSE3-NEXT: pextrw $4, %xmm1, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: movd %xmm1, %eax -; CHECK-NOSSSE3-NEXT: rolw $8, %ax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm1 -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm3[0],xmm1[1],xmm3[1],xmm1[2],xmm3[2],xmm1[3],xmm3[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm4[0],xmm1[1],xmm4[1],xmm1[2],xmm4[2],xmm1[3],xmm4[3] -; CHECK-NOSSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3] +; CHECK-NOSSSE3-NEXT: pxor %xmm2, %xmm2 +; CHECK-NOSSSE3-NEXT: movdqa %xmm0, %xmm3 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm3 = 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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[1,0,3,2,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,5,4,7,6] +; CHECK-NOSSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[1,0,3,2,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,5,4,7,6] +; CHECK-NOSSSE3-NEXT: packuswb %xmm3, %xmm0 +; CHECK-NOSSSE3-NEXT: movdqa %xmm1, %xmm3 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm3 = 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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[1,0,3,2,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,5,4,7,6] +; CHECK-NOSSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[1,0,3,2,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,5,4,7,6] +; CHECK-NOSSSE3-NEXT: packuswb %xmm3, %xmm1 ; CHECK-NOSSSE3-NEXT: retq ; ; CHECK-SSSE3-LABEL: test4: @@ -236,44 +161,23 @@ entry: define <8 x i32> @test5(<8 x i32> %v) { ; CHECK-NOSSSE3-LABEL: test5: ; CHECK-NOSSSE3: # BB#0: # %entry -; CHECK-NOSSSE3-NEXT: movdqa %xmm0, %xmm2 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm2[3,1,2,3] -; CHECK-NOSSSE3-NEXT: movd %xmm0, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm0 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm2[1,1,2,3] -; CHECK-NOSSSE3-NEXT: movd %xmm3, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm0[0],xmm3[1],xmm0[1] -; CHECK-NOSSSE3-NEXT: movd %xmm2, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm0 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm2[2,3,0,1] -; CHECK-NOSSSE3-NEXT: movd %xmm2, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1] -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm0 = xmm0[0],xmm3[0],xmm0[1],xmm3[1] -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm1[3,1,2,3] -; CHECK-NOSSSE3-NEXT: movd %xmm2, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm1[1,1,2,3] -; CHECK-NOSSSE3-NEXT: movd %xmm3, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm3 -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm3 = xmm3[0],xmm2[0],xmm3[1],xmm2[1] -; CHECK-NOSSSE3-NEXT: movd %xmm1, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1] -; CHECK-NOSSSE3-NEXT: movd %xmm1, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm1 -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1] -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm3[0],xmm2[1],xmm3[1] -; CHECK-NOSSSE3-NEXT: movdqa %xmm2, %xmm1 +; CHECK-NOSSSE3-NEXT: pxor %xmm2, %xmm2 +; CHECK-NOSSSE3-NEXT: movdqa %xmm0, %xmm3 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm3 = 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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: packuswb %xmm3, %xmm0 +; CHECK-NOSSSE3-NEXT: movdqa %xmm1, %xmm3 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm3 = 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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: packuswb %xmm3, %xmm1 ; CHECK-NOSSSE3-NEXT: retq ; ; CHECK-SSSE3-LABEL: test5: @@ -300,24 +204,27 @@ entry: define <4 x i64> @test6(<4 x i64> %v) { ; CHECK-NOSSSE3-LABEL: test6: ; CHECK-NOSSSE3: # BB#0: # %entry -; CHECK-NOSSSE3-NEXT: movd %xmm0, %rax -; CHECK-NOSSSE3-NEXT: bswapq %rax -; CHECK-NOSSSE3-NEXT: movd %rax, %xmm2 +; CHECK-NOSSSE3-NEXT: pxor %xmm2, %xmm2 +; CHECK-NOSSSE3-NEXT: movdqa %xmm0, %xmm3 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm3 = 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] +; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm3[2,3,0,1] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0],xmm2[0],xmm0[1],xmm2[1],xmm0[2],xmm2[2],xmm0[3],xmm2[3],xmm0[4],xmm2[4],xmm0[5],xmm2[5],xmm0[6],xmm2[6],xmm0[7],xmm2[7] ; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1] -; CHECK-NOSSSE3-NEXT: movd %xmm0, %rax -; CHECK-NOSSSE3-NEXT: bswapq %rax -; CHECK-NOSSSE3-NEXT: movd %rax, %xmm0 -; CHECK-NOSSSE3-NEXT: punpcklqdq {{.*#+}} xmm2 = xmm2[0],xmm0[0] -; CHECK-NOSSSE3-NEXT: movd %xmm1, %rax -; CHECK-NOSSSE3-NEXT: bswapq %rax -; CHECK-NOSSSE3-NEXT: movd %rax, %xmm3 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm1[2,3,0,1] -; CHECK-NOSSSE3-NEXT: movd %xmm0, %rax -; CHECK-NOSSSE3-NEXT: bswapq %rax -; CHECK-NOSSSE3-NEXT: movd %rax, %xmm0 -; CHECK-NOSSSE3-NEXT: punpcklqdq {{.*#+}} xmm3 = xmm3[0],xmm0[0] -; CHECK-NOSSSE3-NEXT: movdqa %xmm2, %xmm0 -; CHECK-NOSSSE3-NEXT: movdqa %xmm3, %xmm1 +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: packuswb %xmm3, %xmm0 +; CHECK-NOSSSE3-NEXT: movdqa %xmm1, %xmm3 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm3 = 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] +; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm3 = xmm3[2,3,0,1] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm3 = xmm3[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm3 = xmm3[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: punpcklbw {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1],xmm1[2],xmm2[2],xmm1[3],xmm2[3],xmm1[4],xmm2[4],xmm1[5],xmm2[5],xmm1[6],xmm2[6],xmm1[7],xmm2[7] +; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm1[2,3,0,1] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm1 = xmm1[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm1 = xmm1[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: packuswb %xmm3, %xmm1 ; CHECK-NOSSSE3-NEXT: retq ; ; CHECK-SSSE3-LABEL: test6: @@ -346,26 +253,16 @@ declare <4 x i16> @llvm.bswap.v4i16(<4 x i16>) define <4 x i16> @test7(<4 x i16> %v) { ; CHECK-NOSSSE3-LABEL: test7: ; CHECK-NOSSSE3: # BB#0: # %entry -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm1 = xmm0[3,1,2,3] -; CHECK-NOSSSE3-NEXT: movd %xmm1, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm1 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm2 = xmm0[1,1,2,3] -; CHECK-NOSSSE3-NEXT: movd %xmm2, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm2 -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm2 = xmm2[0],xmm1[0],xmm2[1],xmm1[1] -; CHECK-NOSSSE3-NEXT: movd %xmm0, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm1 -; CHECK-NOSSSE3-NEXT: pshufd {{.*#+}} xmm0 = xmm0[2,3,0,1] -; CHECK-NOSSSE3-NEXT: movd %xmm0, %eax -; CHECK-NOSSSE3-NEXT: bswapl %eax -; CHECK-NOSSSE3-NEXT: movd %eax, %xmm0 -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1] -; CHECK-NOSSSE3-NEXT: punpckldq {{.*#+}} xmm1 = xmm1[0],xmm2[0],xmm1[1],xmm2[1] -; CHECK-NOSSSE3-NEXT: psrld $16, %xmm1 -; CHECK-NOSSSE3-NEXT: movdqa %xmm1, %xmm0 +; CHECK-NOSSSE3-NEXT: pxor %xmm1, %xmm1 +; CHECK-NOSSSE3-NEXT: movdqa %xmm0, %xmm2 +; CHECK-NOSSSE3-NEXT: punpckhbw {{.*#+}} xmm2 = 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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm2 = xmm2[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm2 = xmm2[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-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] +; CHECK-NOSSSE3-NEXT: pshuflw {{.*#+}} xmm0 = xmm0[3,2,1,0,4,5,6,7] +; CHECK-NOSSSE3-NEXT: pshufhw {{.*#+}} xmm0 = xmm0[0,1,2,3,7,6,5,4] +; CHECK-NOSSSE3-NEXT: packuswb %xmm2, %xmm0 +; CHECK-NOSSSE3-NEXT: psrld $16, %xmm0 ; CHECK-NOSSSE3-NEXT: retq ; ; CHECK-SSSE3-LABEL: test7: diff --git a/llvm/test/CodeGen/X86/mmx-arg-passing-x86-64.ll b/llvm/test/CodeGen/X86/mmx-arg-passing-x86-64.ll index b4a51e4..c536a39 100644 --- a/llvm/test/CodeGen/X86/mmx-arg-passing-x86-64.ll +++ b/llvm/test/CodeGen/X86/mmx-arg-passing-x86-64.ll @@ -10,11 +10,6 @@ define void @t3() nounwind { ; X86-64: ## BB#0: ; X86-64-NEXT: movq _g_v8qi@{{.*}}(%rip), %rax ; X86-64-NEXT: movq {{.*#+}} xmm0 = mem[0],zero -; X86-64-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] -; X86-64-NEXT: pand {{.*}}(%rip), %xmm0 -; X86-64-NEXT: packuswb %xmm0, %xmm0 -; X86-64-NEXT: movd %xmm0, %rax -; X86-64-NEXT: movd %rax, %xmm0 ; X86-64-NEXT: movb $1, %al ; X86-64-NEXT: jmp _pass_v8qi ## TAILCALL %tmp3 = load <8 x i8>* @g_v8qi, align 8 @@ -33,9 +28,6 @@ define void @t4(x86_mmx %v1, x86_mmx %v2) nounwind { ; X86-64-NEXT: movq {{.*#+}} xmm0 = mem[0],zero ; X86-64-NEXT: movq {{.*#+}} xmm1 = mem[0],zero ; X86-64-NEXT: paddb %xmm0, %xmm1 -; X86-64-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] -; X86-64-NEXT: pand {{.*}}(%rip), %xmm1 -; X86-64-NEXT: packuswb %xmm1, %xmm1 ; X86-64-NEXT: movd %xmm1, %rax ; X86-64-NEXT: movd %rax, %xmm0 ; X86-64-NEXT: movb $1, %al diff --git a/llvm/test/CodeGen/X86/sse3.ll b/llvm/test/CodeGen/X86/sse3.ll index 4223b79..6c0b701 100644 --- a/llvm/test/CodeGen/X86/sse3.ll +++ b/llvm/test/CodeGen/X86/sse3.ll @@ -255,13 +255,8 @@ entry: define <16 x i8> @t16(<16 x i8> %T0) nounwind readnone { ; X64-LABEL: t16: ; X64: ## BB#0: ## %entry -; X64-NEXT: movdqa {{.*#+}} xmm2 = [0,0,0,0,1,1,1,1,0,0,0,0,0,0,0,0] -; X64-NEXT: punpcklwd {{.*#+}} xmm2 = xmm2[0],xmm0[0],xmm2[1],xmm0[1],xmm2[2],xmm0[2],xmm2[3],xmm0[3] -; X64-NEXT: pslld $16, %xmm0 -; X64-NEXT: movdqa {{.*#+}} xmm1 = [255,255,255,0,255,255,255,255,255,255,255,255,255,255,255,255] -; X64-NEXT: pand %xmm1, %xmm2 -; X64-NEXT: pandn %xmm0, %xmm1 -; X64-NEXT: por %xmm2, %xmm1 +; X64-NEXT: movdqa {{.*#+}} xmm1 = [0,0,0,0,1,1,1,1,0,0,0,0,0,0,0,0] +; X64-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] ; X64-NEXT: movdqa %xmm1, %xmm0 ; X64-NEXT: retq entry: diff --git a/llvm/test/CodeGen/X86/sse41.ll b/llvm/test/CodeGen/X86/sse41.ll index 189b4ab..a5b07e7 100644 --- a/llvm/test/CodeGen/X86/sse41.ll +++ b/llvm/test/CodeGen/X86/sse41.ll @@ -551,18 +551,12 @@ define <4 x float> @shuf_X00A(<4 x float> %x, <4 x float> %a) { define <4 x float> @shuf_X00X(<4 x float> %x, <4 x float> %a) { ; X32-LABEL: shuf_X00X: ; X32: ## BB#0: -; X32-NEXT: xorps %xmm1, %xmm1 -; X32-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3] -; X32-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[0] -; X32-NEXT: movaps %xmm1, %xmm0 +; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],zero,zero,xmm0[0] ; X32-NEXT: retl ; ; X64-LABEL: shuf_X00X: ; X64: ## BB#0: -; X64-NEXT: xorps %xmm1, %xmm1 -; X64-NEXT: blendps {{.*#+}} xmm1 = xmm0[0],xmm1[1,2,3] -; X64-NEXT: insertps {{.*#+}} xmm1 = xmm1[0,1,2],xmm0[0] -; X64-NEXT: movaps %xmm1, %xmm0 +; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],zero,zero,xmm0[0] ; X64-NEXT: retq %vecext = extractelement <4 x float> %x, i32 0 %vecinit = insertelement <4 x float> undef, float %vecext, i32 0 @@ -575,20 +569,14 @@ define <4 x float> @shuf_X00X(<4 x float> %x, <4 x float> %a) { define <4 x float> @shuf_X0YC(<4 x float> %x, <4 x float> %a) { ; X32-LABEL: shuf_X0YC: ; X32: ## BB#0: -; X32-NEXT: xorps %xmm2, %xmm2 -; X32-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1],xmm0[2,3] -; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1],xmm0[1],zero -; X32-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[2] -; X32-NEXT: movaps %xmm2, %xmm0 +; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero +; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm1[2] ; X32-NEXT: retl ; ; X64-LABEL: shuf_X0YC: ; X64: ## BB#0: -; X64-NEXT: xorps %xmm2, %xmm2 -; X64-NEXT: blendps {{.*#+}} xmm2 = xmm0[0],xmm2[1],xmm0[2,3] -; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1],xmm0[1],zero -; X64-NEXT: insertps {{.*#+}} xmm2 = xmm2[0,1,2],xmm1[2] -; X64-NEXT: movaps %xmm2, %xmm0 +; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0],zero,xmm0[1],zero +; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,2],xmm1[2] ; X64-NEXT: retq %vecext = extractelement <4 x float> %x, i32 0 %vecinit = insertelement <4 x float> undef, float %vecext, i32 0 @@ -642,12 +630,16 @@ define <4 x i32> @i32_shuf_XY00(<4 x i32> %x, <4 x i32> %a) { define <4 x i32> @i32_shuf_XYY0(<4 x i32> %x, <4 x i32> %a) { ; X32-LABEL: i32_shuf_XYY0: ; X32: ## BB#0: -; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,1],zero +; X32-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,1,3] +; X32-NEXT: pxor %xmm0, %xmm0 +; X32-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1,2,3,4,5],xmm0[6,7] ; X32-NEXT: retl ; ; X64-LABEL: i32_shuf_XYY0: ; X64: ## BB#0: -; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,1],zero +; X64-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,1,3] +; X64-NEXT: pxor %xmm0, %xmm0 +; X64-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1,2,3,4,5],xmm0[6,7] ; X64-NEXT: retq %vecext = extractelement <4 x i32> %x, i32 0 %vecinit = insertelement <4 x i32> undef, i32 %vecext, i32 0 @@ -661,12 +653,16 @@ define <4 x i32> @i32_shuf_XYY0(<4 x i32> %x, <4 x i32> %a) { define <4 x i32> @i32_shuf_XYW0(<4 x i32> %x, <4 x i32> %a) { ; X32-LABEL: i32_shuf_XYW0: ; X32: ## BB#0: -; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,3],zero +; X32-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,3,3] +; X32-NEXT: pxor %xmm0, %xmm0 +; X32-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1,2,3,4,5],xmm0[6,7] ; X32-NEXT: retl ; ; X64-LABEL: i32_shuf_XYW0: ; X64: ## BB#0: -; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[0,1,3],zero +; X64-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,3,3] +; X64-NEXT: pxor %xmm0, %xmm0 +; X64-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1,2,3,4,5],xmm0[6,7] ; X64-NEXT: retq %vecext = extractelement <4 x i32> %x, i32 0 %vecinit = insertelement <4 x i32> undef, i32 %vecext, i32 0 @@ -681,12 +677,16 @@ define <4 x i32> @i32_shuf_XYW0(<4 x i32> %x, <4 x i32> %a) { define <4 x i32> @i32_shuf_W00W(<4 x i32> %x, <4 x i32> %a) { ; X32-LABEL: i32_shuf_W00W: ; X32: ## BB#0: -; X32-NEXT: insertps {{.*#+}} xmm0 = xmm0[3],zero,zero,xmm0[3] +; X32-NEXT: pshufd {{.*#+}} xmm1 = xmm0[3,1,2,3] +; X32-NEXT: pxor %xmm0, %xmm0 +; X32-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5],xmm1[6,7] ; X32-NEXT: retl ; ; X64-LABEL: i32_shuf_W00W: ; X64: ## BB#0: -; X64-NEXT: insertps {{.*#+}} xmm0 = xmm0[3],zero,zero,xmm0[3] +; X64-NEXT: pshufd {{.*#+}} xmm1 = xmm0[3,1,2,3] +; X64-NEXT: pxor %xmm0, %xmm0 +; X64-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5],xmm1[6,7] ; X64-NEXT: retq %vecext = extractelement <4 x i32> %x, i32 3 %vecinit = insertelement <4 x i32> undef, i32 %vecext, i32 0 @@ -723,18 +723,16 @@ define <4 x i32> @i32_shuf_X00A(<4 x i32> %x, <4 x i32> %a) { define <4 x i32> @i32_shuf_X00X(<4 x i32> %x, <4 x i32> %a) { ; X32-LABEL: i32_shuf_X00X: ; X32: ## BB#0: -; X32-NEXT: pxor %xmm1, %xmm1 -; X32-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1],xmm1[2,3,4,5,6,7] -; X32-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,2,0] -; X32-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1,2,3,4,5],xmm0[6,7] +; X32-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,2,0] +; X32-NEXT: pxor %xmm0, %xmm0 +; X32-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5],xmm1[6,7] ; X32-NEXT: retl ; ; X64-LABEL: i32_shuf_X00X: ; X64: ## BB#0: -; X64-NEXT: pxor %xmm1, %xmm1 -; X64-NEXT: pblendw {{.*#+}} xmm1 = xmm0[0,1],xmm1[2,3,4,5,6,7] -; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm0[0,1,2,0] -; X64-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1,2,3,4,5],xmm0[6,7] +; X64-NEXT: pshufd {{.*#+}} xmm1 = xmm0[0,1,2,0] +; X64-NEXT: pxor %xmm0, %xmm0 +; X64-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5],xmm1[6,7] ; X64-NEXT: retq %vecext = extractelement <4 x i32> %x, i32 0 %vecinit = insertelement <4 x i32> undef, i32 %vecext, i32 0 @@ -748,7 +746,6 @@ define <4 x i32> @i32_shuf_X0YC(<4 x i32> %x, <4 x i32> %a) { ; X32-LABEL: i32_shuf_X0YC: ; X32: ## BB#0: ; X32-NEXT: pmovzxdq {{.*#+}} xmm2 = xmm0[0],zero,xmm0[1],zero -; X32-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,5,6,7] ; X32-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,1,2,2] ; X32-NEXT: pblendw {{.*#+}} xmm0 = xmm2[0,1,2,3,4,5],xmm0[6,7] ; X32-NEXT: retl @@ -756,7 +753,6 @@ define <4 x i32> @i32_shuf_X0YC(<4 x i32> %x, <4 x i32> %a) { ; X64-LABEL: i32_shuf_X0YC: ; X64: ## BB#0: ; X64-NEXT: pmovzxdq {{.*#+}} xmm2 = xmm0[0],zero,xmm0[1],zero -; X64-NEXT: pblendw {{.*#+}} xmm2 = xmm2[0,1,2,3,4,5,6,7] ; X64-NEXT: pshufd {{.*#+}} xmm0 = xmm1[0,1,2,2] ; X64-NEXT: pblendw {{.*#+}} xmm0 = xmm2[0,1,2,3,4,5],xmm0[6,7] ; X64-NEXT: retq @@ -1012,15 +1008,17 @@ define void @insertps_pr20411(i32* noalias nocapture %RET) #1 { ; X32: ## BB#0: ; X32-NEXT: movl {{[0-9]+}}(%esp), %eax ; X32-NEXT: pshufd {{.*#+}} xmm0 = mem[2,3,0,1] -; X32-NEXT: pblendw {{.*#+}} xmm0 = mem[0,1],xmm0[2,3],mem[4,5,6,7] -; X32-NEXT: movdqu %xmm0, (%eax) +; X32-NEXT: pshufd {{.*#+}} xmm1 = mem[3,1,2,3] +; X32-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3],xmm1[4,5,6,7] +; X32-NEXT: movdqu %xmm1, (%eax) ; X32-NEXT: retl ; ; X64-LABEL: insertps_pr20411: ; X64: ## BB#0: ; X64-NEXT: pshufd {{.*#+}} xmm0 = mem[2,3,0,1] -; X64-NEXT: pblendw {{.*#+}} xmm0 = mem[0,1],xmm0[2,3],mem[4,5,6,7] -; X64-NEXT: movdqu %xmm0, (%rdi) +; X64-NEXT: pshufd {{.*#+}} xmm1 = mem[3,1,2,3] +; X64-NEXT: pblendw {{.*#+}} xmm1 = xmm1[0,1],xmm0[2,3],xmm1[4,5,6,7] +; X64-NEXT: movdqu %xmm1, (%rdi) ; X64-NEXT: retq %gather_load = shufflevector <8 x i32> , <8 x i32> undef, <8 x i32> %shuffle109 = shufflevector <4 x i32> , <4 x i32> undef, <4 x i32> ; 4 5 6 7 diff --git a/llvm/test/CodeGen/X86/vector-blend.ll b/llvm/test/CodeGen/X86/vector-blend.ll index 51995da..e15daaa 100644 --- a/llvm/test/CodeGen/X86/vector-blend.ll +++ b/llvm/test/CodeGen/X86/vector-blend.ll @@ -255,31 +255,30 @@ entry: define <16 x i8> @vsel_i8(<16 x i8> %v1, <16 x i8> %v2) { ; SSE2-LABEL: vsel_i8: ; SSE2: # BB#0: # %entry -; SSE2-NEXT: movaps {{.*#+}} xmm2 = [0,255,255,255,0,255,255,255,0,255,255,255,0,255,255,255] -; SSE2-NEXT: andps %xmm2, %xmm1 -; SSE2-NEXT: andnps %xmm0, %xmm2 -; SSE2-NEXT: orps %xmm1, %xmm2 -; SSE2-NEXT: movaps %xmm2, %xmm0 +; SSE2-NEXT: movaps {{.*#+}} xmm2 = [255,0,0,0,255,0,0,0,255,255,255,255,255,255,255,255] +; SSE2-NEXT: andps %xmm2, %xmm0 +; SSE2-NEXT: andnps %xmm1, %xmm2 +; SSE2-NEXT: orps %xmm2, %xmm0 ; SSE2-NEXT: retq ; ; SSSE3-LABEL: vsel_i8: ; SSSE3: # BB#0: # %entry -; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[8],zero,zero,zero,xmm0[12],zero,zero,zero -; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = zero,xmm1[1,2,3],zero,xmm1[5,6,7],zero,xmm1[9,10,11],zero,xmm1[13,14,15] +; SSSE3-NEXT: pshufb {{.*#+}} xmm1 = zero,xmm1[1,2,3],zero,xmm1[5,6,7],zero,zero,zero,zero,zero,zero,zero,zero +; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0],zero,zero,zero,xmm0[4],zero,zero,zero,xmm0[8,9,10,11,12,13,14,15] ; SSSE3-NEXT: por %xmm1, %xmm0 ; SSSE3-NEXT: retq ; ; SSE41-LABEL: vsel_i8: ; SSE41: # BB#0: # %entry ; SSE41-NEXT: movdqa %xmm0, %xmm2 -; SSE41-NEXT: movaps {{.*#+}} xmm0 = [255,0,0,0,255,0,0,0,255,0,0,0,255,0,0,0] +; SSE41-NEXT: movaps {{.*#+}} xmm0 = [255,0,0,0,255,0,0,0,255,255,255,255,255,255,255,255] ; SSE41-NEXT: pblendvb %xmm2, %xmm1 ; SSE41-NEXT: movdqa %xmm1, %xmm0 ; SSE41-NEXT: retq ; ; AVX-LABEL: vsel_i8: ; AVX: # BB#0: # %entry -; AVX-NEXT: vmovdqa {{.*#+}} xmm2 = [255,0,0,0,255,0,0,0,255,0,0,0,255,0,0,0] +; AVX-NEXT: vmovdqa {{.*#+}} xmm2 = [255,0,0,0,255,0,0,0,255,255,255,255,255,255,255,255] ; AVX-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm0 ; AVX-NEXT: retq entry: @@ -624,34 +623,32 @@ entry: define <32 x i8> @constant_pblendvb_avx2(<32 x i8> %xyzw, <32 x i8> %abcd) { ; SSE2-LABEL: constant_pblendvb_avx2: ; SSE2: # BB#0: # %entry -; SSE2-NEXT: movaps {{.*#+}} xmm4 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255] +; SSE2-NEXT: movaps {{.*#+}} xmm4 = [0,0,255,0,255,255,255,0,255,255,255,255,255,255,255,255] ; SSE2-NEXT: movaps %xmm4, %xmm5 -; SSE2-NEXT: andnps %xmm0, %xmm5 -; SSE2-NEXT: andps %xmm4, %xmm2 -; SSE2-NEXT: orps %xmm2, %xmm5 -; SSE2-NEXT: andps %xmm4, %xmm3 -; SSE2-NEXT: andnps %xmm1, %xmm4 -; SSE2-NEXT: orps %xmm3, %xmm4 -; SSE2-NEXT: movaps %xmm5, %xmm0 -; SSE2-NEXT: movaps %xmm4, %xmm1 +; SSE2-NEXT: andnps %xmm2, %xmm5 +; SSE2-NEXT: andps %xmm4, %xmm0 +; SSE2-NEXT: orps %xmm5, %xmm0 +; SSE2-NEXT: andps %xmm4, %xmm1 +; SSE2-NEXT: andnps %xmm3, %xmm4 +; SSE2-NEXT: orps %xmm4, %xmm1 ; SSE2-NEXT: retq ; ; SSSE3-LABEL: constant_pblendvb_avx2: ; SSSE3: # BB#0: # %entry -; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [128,128,2,128,4,5,6,128,128,128,10,128,12,13,14,128] -; SSSE3-NEXT: pshufb %xmm4, %xmm0 -; SSSE3-NEXT: movdqa {{.*#+}} xmm5 = [0,1,128,3,128,128,128,7,8,9,128,11,128,128,128,15] -; SSSE3-NEXT: pshufb %xmm5, %xmm2 +; SSSE3-NEXT: movdqa {{.*#+}} xmm4 = [0,1,128,3,128,128,128,7,128,128,128,128,128,128,128,128] +; SSSE3-NEXT: pshufb %xmm4, %xmm2 +; SSSE3-NEXT: movdqa {{.*#+}} xmm5 = [128,128,2,128,4,5,6,128,8,9,10,11,12,13,14,15] +; SSSE3-NEXT: pshufb %xmm5, %xmm0 ; SSSE3-NEXT: por %xmm2, %xmm0 -; SSSE3-NEXT: pshufb %xmm4, %xmm1 -; SSSE3-NEXT: pshufb %xmm5, %xmm3 +; SSSE3-NEXT: pshufb %xmm4, %xmm3 +; SSSE3-NEXT: pshufb %xmm5, %xmm1 ; SSSE3-NEXT: por %xmm3, %xmm1 ; SSSE3-NEXT: retq ; ; SSE41-LABEL: constant_pblendvb_avx2: ; SSE41: # BB#0: # %entry ; SSE41-NEXT: movdqa %xmm0, %xmm4 -; SSE41-NEXT: movaps {{.*#+}} xmm0 = [0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0] +; SSE41-NEXT: movaps {{.*#+}} xmm0 = [0,0,255,0,255,255,255,0,255,255,255,255,255,255,255,255] ; SSE41-NEXT: pblendvb %xmm4, %xmm2 ; SSE41-NEXT: pblendvb %xmm1, %xmm3 ; SSE41-NEXT: movdqa %xmm2, %xmm0 @@ -660,17 +657,15 @@ define <32 x i8> @constant_pblendvb_avx2(<32 x i8> %xyzw, <32 x i8> %abcd) { ; ; AVX1-LABEL: constant_pblendvb_avx2: ; AVX1: # BB#0: # %entry -; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 -; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm3 -; AVX1-NEXT: vmovdqa {{.*#+}} xmm4 = [255,255,0,255,0,0,0,255,255,255,0,255,0,0,0,255] -; AVX1-NEXT: vpblendvb %xmm4, %xmm2, %xmm3, %xmm2 -; AVX1-NEXT: vpblendvb %xmm4, %xmm1, %xmm0, %xmm0 -; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0 +; AVX1-NEXT: vmovdqa {{.*#+}} xmm2 = [0,0,255,0,255,255,255,0,255,255,255,255,255,255,255,255] +; AVX1-NEXT: vpblendvb %xmm2, %xmm0, %xmm1, %xmm1 +; AVX1-NEXT: vextractf128 $1, %ymm0, %xmm0 +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm1, %ymm0 ; AVX1-NEXT: retq ; ; AVX2-LABEL: constant_pblendvb_avx2: ; AVX2: # BB#0: # %entry -; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = [0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0,0,0,255,0,255,255,255,0] +; AVX2-NEXT: vmovdqa {{.*#+}} ymm2 = [0,0,255,0,255,255,255,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] ; AVX2-NEXT: vpblendvb %ymm2, %ymm0, %ymm1, %ymm0 ; AVX2-NEXT: retq entry: -- 2.7.4