diff options
| -rw-r--r-- | llvm/test/CodeGen/X86/vector-shuffle-combining-avx2.ll | 22 |
1 files changed, 22 insertions, 0 deletions
diff --git a/llvm/test/CodeGen/X86/vector-shuffle-combining-avx2.ll b/llvm/test/CodeGen/X86/vector-shuffle-combining-avx2.ll index bedf6823277..f2f45fc6700 100644 --- a/llvm/test/CodeGen/X86/vector-shuffle-combining-avx2.ll +++ b/llvm/test/CodeGen/X86/vector-shuffle-combining-avx2.ll @@ -342,6 +342,28 @@ define <8 x float> @combine_permps_as_permpd(<8 x float> %a) { ret <8 x float> %1 } +define <8 x float> @combine_permps_as_vpermilps(<8 x float> %a, i32 %a1) { +; X86-LABEL: combine_permps_as_vpermilps: +; X86: # %bb.0: +; X86-NEXT: vmovss {{.*#+}} xmm1 = mem[0],zero,zero,zero +; X86-NEXT: vblendps {{.*#+}} ymm1 = ymm1[0],mem[1,2,3,4,5,6,7] +; X86-NEXT: vpermps %ymm0, %ymm1, %ymm0 +; X86-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[1,1,2,3,4,5,6,7] +; X86-NEXT: retl +; +; X64-LABEL: combine_permps_as_vpermilps: +; X64: # %bb.0: +; X64-NEXT: vmovd %edi, %xmm1 +; X64-NEXT: vpblendd {{.*#+}} ymm1 = ymm1[0],mem[1,2,3,4,5,6,7] +; X64-NEXT: vpermd %ymm0, %ymm1, %ymm0 +; X64-NEXT: vpermilps {{.*#+}} ymm0 = ymm0[1,1,2,3,4,5,6,7] +; X64-NEXT: retq + %1 = insertelement <8 x i32> <i32 3, i32 2, i32 1, i32 0, i32 7, i32 6, i32 5, i32 4>, i32 %a1, i32 0 + %2 = call <8 x float> @llvm.x86.avx2.permps(<8 x float> %a, <8 x i32> %1) + %3 = shufflevector <8 x float> %2, <8 x float> undef, <8 x i32> <i32 1, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> + ret <8 x float> %3 +} + define <4 x i64> @combine_pshufb_as_zext(<32 x i8> %a0) { ; CHECK-LABEL: combine_pshufb_as_zext: ; CHECK: # %bb.0: |
