diff options
| -rw-r--r-- | llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll | 64 |
1 files changed, 64 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll b/llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll index 45c2abc03cd..939a27c22c4 100644 --- a/llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll +++ b/llvm/test/CodeGen/X86/vector-shuffle-combining-avx.ll @@ -417,3 +417,67 @@ entry: store <8 x float> %1, <8 x float>* undef, align 16 unreachable } + +define <4 x i64> @concat_self_v4i64(<2 x i64> %x) { +; AVX1-LABEL: concat_self_v4i64: +; AVX1: # %bb.0: +; AVX1-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0 +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 +; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm0[2,3,0,1] +; AVX1-NEXT: vpermilpd {{.*#+}} ymm1 = ymm1[0,0,3,2] +; AVX1-NEXT: vblendpd {{.*#+}} ymm0 = ymm0[0],ymm1[1,2],ymm0[3] +; AVX1-NEXT: ret{{[l|q]}} +; +; AVX2-LABEL: concat_self_v4i64: +; AVX2: # %bb.0: +; AVX2-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0 +; AVX2-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 +; AVX2-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,2,1,3] +; AVX2-NEXT: ret{{[l|q]}} +; +; AVX512-LABEL: concat_self_v4i64: +; AVX512: # %bb.0: +; AVX512-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0 +; AVX512-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 +; AVX512-NEXT: vpermpd {{.*#+}} ymm0 = ymm0[0,2,1,3] +; AVX512-NEXT: ret{{[l|q]}} + %cat = shufflevector <2 x i64> %x, <2 x i64> undef, <4 x i32> <i32 0, i32 1, i32 0, i32 1> + %s = shufflevector <4 x i64> %cat, <4 x i64> undef, <4 x i32> <i32 0, i32 2, i32 1, i32 3> + ret <4 x i64> %s +} + +define <8 x i32> @concat_self_v8i32(<4 x i32> %x) { +; AVX1-LABEL: concat_self_v8i32: +; AVX1: # %bb.0: +; AVX1-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0 +; AVX1-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm1 +; AVX1-NEXT: vperm2f128 {{.*#+}} ymm1 = ymm1[2,3,0,1] +; AVX1-NEXT: vpermilps {{.*#+}} ymm1 = ymm1[3,2,1,0,4,6,5,7] +; AVX1-NEXT: vextractf128 $1, %ymm1, %xmm2 +; AVX1-NEXT: vpaddd %xmm0, %xmm2, %xmm2 +; AVX1-NEXT: vpaddd %xmm0, %xmm1, %xmm0 +; AVX1-NEXT: vinsertf128 $1, %xmm2, %ymm0, %ymm0 +; AVX1-NEXT: ret{{[l|q]}} +; +; AVX2-LABEL: concat_self_v8i32: +; AVX2: # %bb.0: +; AVX2-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0 +; AVX2-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0 +; AVX2-NEXT: vmovdqa {{.*#+}} ymm1 = [7,6,5,4,0,2,1,3] +; AVX2-NEXT: vpermd %ymm0, %ymm1, %ymm1 +; AVX2-NEXT: vpaddd %ymm0, %ymm1, %ymm0 +; AVX2-NEXT: ret{{[l|q]}} +; +; AVX512-LABEL: concat_self_v8i32: +; AVX512: # %bb.0: +; AVX512-NEXT: # kill: def $xmm0 killed $xmm0 def $ymm0 +; AVX512-NEXT: vinserti128 $1, %xmm0, %ymm0, %ymm0 +; AVX512-NEXT: vmovdqa {{.*#+}} ymm1 = [7,6,5,4,0,2,1,3] +; AVX512-NEXT: vpermd %ymm0, %ymm1, %ymm1 +; AVX512-NEXT: vpaddd %ymm0, %ymm1, %ymm0 +; AVX512-NEXT: ret{{[l|q]}} + %cat = shufflevector <4 x i32> %x, <4 x i32> undef, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 0, i32 1, i32 2, i32 3> + %s = shufflevector <8 x i32> %cat, <8 x i32> undef, <8 x i32> <i32 7, i32 6, i32 5, i32 4, i32 0, i32 2, i32 1, i32 3> + %a = add <8 x i32> %s, %cat + ret <8 x i32> %a +} |
