diff options
Diffstat (limited to 'llvm/test/CodeGen/X86/vector-shuffle-combining.ll')
-rw-r--r-- | llvm/test/CodeGen/X86/vector-shuffle-combining.ll | 176 |
1 files changed, 116 insertions, 60 deletions
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-combining.ll b/llvm/test/CodeGen/X86/vector-shuffle-combining.ll index f670bf71a8f..58b7be1ec80 100644 --- a/llvm/test/CodeGen/X86/vector-shuffle-combining.ll +++ b/llvm/test/CodeGen/X86/vector-shuffle-combining.ll @@ -1715,37 +1715,41 @@ define <4 x i8> @combine_test1c(<4 x i8>* %a, <4 x i8>* %b) { ; SSE2-LABEL: combine_test1c: ; SSE2: # %bb.0: ; SSE2-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; SSE2-NEXT: movss {{.*#+}} xmm2 = mem[0],zero,zero,zero -; SSE2-NEXT: movaps {{.*#+}} xmm0 = [0,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255] -; SSE2-NEXT: andps %xmm0, %xmm2 -; SSE2-NEXT: andnps %xmm1, %xmm0 -; SSE2-NEXT: orps %xmm2, %xmm0 +; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero +; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] +; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] +; SSE2-NEXT: movss {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3] ; SSE2-NEXT: retq ; ; SSSE3-LABEL: combine_test1c: ; SSSE3: # %bb.0: -; SSSE3-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero +; SSSE3-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero ; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; SSSE3-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] -; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[1,2,4,6,u,u,u,u,u,u,u,u,u,u,u,u] +; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] +; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] +; SSSE3-NEXT: movss {{.*#+}} xmm0 = xmm1[0],xmm0[1,2,3] ; SSSE3-NEXT: retq ; ; SSE41-LABEL: combine_test1c: ; SSE41: # %bb.0: -; SSE41-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero -; SSE41-NEXT: movd {{.*#+}} xmm2 = mem[0],zero,zero,zero -; SSE41-NEXT: movaps {{.*#+}} xmm0 = <0,255,255,255,u,u,u,u,u,u,u,u,u,u,u,u> -; SSE41-NEXT: pblendvb %xmm0, %xmm2, %xmm1 -; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: pmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; SSE41-NEXT: pmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3,4,5,6,7] ; SSE41-NEXT: retq ; -; AVX-LABEL: combine_test1c: -; AVX: # %bb.0: -; AVX-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; AVX-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero -; AVX-NEXT: vmovdqa {{.*#+}} xmm2 = <0,255,255,255,u,u,u,u,u,u,u,u,u,u,u,u> -; AVX-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-LABEL: combine_test1c: +; AVX1: # %bb.0: +; AVX1-NEXT: vpmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX1-NEXT: vpmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3,4,5,6,7] +; AVX1-NEXT: retq +; +; AVX2-LABEL: combine_test1c: +; AVX2: # %bb.0: +; AVX2-NEXT: vpmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX2-NEXT: vpmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm0[0],xmm1[1,2,3] +; AVX2-NEXT: retq %A = load <4 x i8>, <4 x i8>* %a %B = load <4 x i8>, <4 x i8>* %b %1 = shufflevector <4 x i8> %A, <4 x i8> %B, <4 x i32> <i32 0, i32 5, i32 2, i32 7> @@ -1754,18 +1758,40 @@ define <4 x i8> @combine_test1c(<4 x i8>* %a, <4 x i8>* %b) { } define <4 x i8> @combine_test2c(<4 x i8>* %a, <4 x i8>* %b) { -; SSE-LABEL: combine_test2c: -; SSE: # %bb.0: -; SSE-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; SSE-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero -; SSE-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3] -; SSE-NEXT: retq +; SSE2-LABEL: combine_test2c: +; SSE2: # %bb.0: +; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero +; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] +; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] +; SSE2-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero +; SSE2-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] +; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] +; SSE2-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; SSE2-NEXT: retq +; +; SSSE3-LABEL: combine_test2c: +; SSSE3: # %bb.0: +; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero +; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] +; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] +; SSSE3-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero +; SSSE3-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] +; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] +; SSSE3-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; SSSE3-NEXT: retq +; +; SSE41-LABEL: combine_test2c: +; SSE41: # %bb.0: +; SSE41-NEXT: pmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; SSE41-NEXT: pmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; SSE41-NEXT: punpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] +; SSE41-NEXT: retq ; ; AVX-LABEL: combine_test2c: ; AVX: # %bb.0: -; AVX-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; AVX-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero -; AVX-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],xmm1[3] +; AVX-NEXT: vpmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX-NEXT: vpmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX-NEXT: vpunpcklqdq {{.*#+}} xmm0 = xmm0[0],xmm1[0] ; AVX-NEXT: retq %A = load <4 x i8>, <4 x i8>* %a %B = load <4 x i8>, <4 x i8>* %b @@ -1775,20 +1801,40 @@ define <4 x i8> @combine_test2c(<4 x i8>* %a, <4 x i8>* %b) { } define <4 x i8> @combine_test3c(<4 x i8>* %a, <4 x i8>* %b) { -; SSE-LABEL: combine_test3c: -; SSE: # %bb.0: -; SSE-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; SSE-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero -; SSE-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] -; SSE-NEXT: pshufd {{.*#+}} xmm0 = xmm1[1,1,2,3] -; SSE-NEXT: retq +; SSE2-LABEL: combine_test3c: +; SSE2: # %bb.0: +; SSE2-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero +; SSE2-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] +; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] +; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero +; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] +; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] +; SSE2-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1] +; SSE2-NEXT: retq +; +; SSSE3-LABEL: combine_test3c: +; SSSE3: # %bb.0: +; SSSE3-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero +; SSSE3-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] +; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] +; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero +; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] +; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] +; SSSE3-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1] +; SSSE3-NEXT: retq +; +; SSE41-LABEL: combine_test3c: +; SSE41: # %bb.0: +; SSE41-NEXT: pmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; SSE41-NEXT: pmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; SSE41-NEXT: punpckhqdq {{.*#+}} xmm0 = xmm0[1],xmm1[1] +; SSE41-NEXT: retq ; ; AVX-LABEL: combine_test3c: ; AVX: # %bb.0: -; AVX-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; AVX-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero -; AVX-NEXT: vpunpcklwd {{.*#+}} xmm0 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] -; AVX-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[1,1,2,3] +; AVX-NEXT: vpmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX-NEXT: vpmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX-NEXT: vpunpckhqdq {{.*#+}} xmm0 = xmm1[1],xmm0[1] ; AVX-NEXT: retq %A = load <4 x i8>, <4 x i8>* %a %B = load <4 x i8>, <4 x i8>* %b @@ -1800,38 +1846,48 @@ define <4 x i8> @combine_test3c(<4 x i8>* %a, <4 x i8>* %b) { define <4 x i8> @combine_test4c(<4 x i8>* %a, <4 x i8>* %b) { ; SSE2-LABEL: combine_test4c: ; SSE2: # %bb.0: +; SSE2-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero +; SSE2-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] +; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] ; SSE2-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero -; SSE2-NEXT: movss {{.*#+}} xmm2 = mem[0],zero,zero,zero -; SSE2-NEXT: movaps {{.*#+}} xmm0 = [255,0,255,255,255,255,255,255,255,255,255,255,255,255,255,255] -; SSE2-NEXT: andps %xmm0, %xmm2 -; SSE2-NEXT: andnps %xmm1, %xmm0 -; SSE2-NEXT: orps %xmm2, %xmm0 +; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,0],xmm1[0,0] +; SSE2-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] +; SSE2-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] +; SSE2-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,0],xmm1[2,3] ; SSE2-NEXT: retq ; ; SSSE3-LABEL: combine_test4c: ; SSSE3: # %bb.0: -; SSSE3-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero ; SSSE3-NEXT: movd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; SSSE3-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] -; SSSE3-NEXT: pshufb {{.*#+}} xmm0 = xmm0[0,3,4,6,u,u,u,u,u,u,u,u,u,u,u,u] +; SSSE3-NEXT: punpcklbw {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3,4,4,5,5,6,6,7,7] +; SSSE3-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] +; SSSE3-NEXT: movss {{.*#+}} xmm1 = mem[0],zero,zero,zero +; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[1,0],xmm1[0,0] +; SSSE3-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] +; SSSE3-NEXT: punpcklwd {{.*#+}} xmm1 = xmm1[0],xmm0[0],xmm1[1],xmm0[1],xmm1[2],xmm0[2],xmm1[3],xmm0[3] +; SSSE3-NEXT: shufps {{.*#+}} xmm0 = xmm0[2,0],xmm1[2,3] ; SSSE3-NEXT: retq ; ; SSE41-LABEL: combine_test4c: ; SSE41: # %bb.0: -; SSE41-NEXT: movd {{.*#+}} xmm1 = mem[0],zero,zero,zero -; SSE41-NEXT: movd {{.*#+}} xmm2 = mem[0],zero,zero,zero -; SSE41-NEXT: movaps {{.*#+}} xmm0 = <255,0,255,255,u,u,u,u,u,u,u,u,u,u,u,u> -; SSE41-NEXT: pblendvb %xmm0, %xmm2, %xmm1 -; SSE41-NEXT: movdqa %xmm1, %xmm0 +; SSE41-NEXT: pmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; SSE41-NEXT: pmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; SSE41-NEXT: pblendw {{.*#+}} xmm0 = xmm0[0,1],xmm1[2,3],xmm0[4,5,6,7] ; SSE41-NEXT: retq ; -; AVX-LABEL: combine_test4c: -; AVX: # %bb.0: -; AVX-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; AVX-NEXT: vmovd {{.*#+}} xmm1 = mem[0],zero,zero,zero -; AVX-NEXT: vmovdqa {{.*#+}} xmm2 = <255,0,255,255,u,u,u,u,u,u,u,u,u,u,u,u> -; AVX-NEXT: vpblendvb %xmm2, %xmm1, %xmm0, %xmm0 -; AVX-NEXT: retq +; AVX1-LABEL: combine_test4c: +; AVX1: # %bb.0: +; AVX1-NEXT: vpmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX1-NEXT: vpmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX1-NEXT: vpblendw {{.*#+}} xmm0 = xmm1[0,1],xmm0[2,3],xmm1[4,5,6,7] +; AVX1-NEXT: retq +; +; AVX2-LABEL: combine_test4c: +; AVX2: # %bb.0: +; AVX2-NEXT: vpmovzxbd {{.*#+}} xmm0 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX2-NEXT: vpmovzxbd {{.*#+}} xmm1 = mem[0],zero,zero,zero,mem[1],zero,zero,zero,mem[2],zero,zero,zero,mem[3],zero,zero,zero +; AVX2-NEXT: vpblendd {{.*#+}} xmm0 = xmm1[0],xmm0[1],xmm1[2,3] +; AVX2-NEXT: retq %A = load <4 x i8>, <4 x i8>* %a %B = load <4 x i8>, <4 x i8>* %b %1 = shufflevector <4 x i8> %A, <4 x i8> %B, <4 x i32> <i32 4, i32 1, i32 6, i32 3> |