diff options
| author | Craig Topper <craig.topper@gmail.com> | 2017-01-31 05:18:29 +0000 |
|---|---|---|
| committer | Craig Topper <craig.topper@gmail.com> | 2017-01-31 05:18:29 +0000 |
| commit | 06e038c6de0051b4028cf3c5ccf6656019e7e52f (patch) | |
| tree | 9b3d44a4a0e5217546bfd978cce429967a536f13 /llvm/test/CodeGen/X86/avx-vbroadcast.ll | |
| parent | 88b0a473122c8f2a3df68dd01d9d67480e57fdb8 (diff) | |
| download | bcm5719-llvm-06e038c6de0051b4028cf3c5ccf6656019e7e52f.tar.gz bcm5719-llvm-06e038c6de0051b4028cf3c5ccf6656019e7e52f.zip | |
[X86] Update the broadcast fallback patterns to use shuffle instructions from the appropriate execution domain.
llvm-svn: 293603
Diffstat (limited to 'llvm/test/CodeGen/X86/avx-vbroadcast.ll')
| -rw-r--r-- | llvm/test/CodeGen/X86/avx-vbroadcast.ll | 34 |
1 files changed, 17 insertions, 17 deletions
diff --git a/llvm/test/CodeGen/X86/avx-vbroadcast.ll b/llvm/test/CodeGen/X86/avx-vbroadcast.ll index 312f4237cd2..77b9c0dc8f1 100644 --- a/llvm/test/CodeGen/X86/avx-vbroadcast.ll +++ b/llvm/test/CodeGen/X86/avx-vbroadcast.ll @@ -49,7 +49,7 @@ define <4 x i64> @A2(i64* %ptr, i64* %ptr2) nounwind uwtable readnone ssp { ; X64-NEXT: movq (%rdi), %rax ; X64-NEXT: vmovq %rax, %xmm0 ; X64-NEXT: movq %rax, (%rsi) -; X64-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1] +; X64-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0] ; X64-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 ; X64-NEXT: retq entry: @@ -114,7 +114,7 @@ define <8 x i32> @B3(i32* %ptr, i32* %ptr2) nounwind uwtable readnone ssp { ; X32-NEXT: movl (%ecx), %ecx ; X32-NEXT: vmovd %ecx, %xmm0 ; X32-NEXT: movl %ecx, (%eax) -; X32-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0] +; X32-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0] ; X32-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 ; X32-NEXT: retl ; @@ -123,7 +123,7 @@ define <8 x i32> @B3(i32* %ptr, i32* %ptr2) nounwind uwtable readnone ssp { ; X64-NEXT: movl (%rdi), %eax ; X64-NEXT: vmovd %eax, %xmm0 ; X64-NEXT: movl %eax, (%rsi) -; X64-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0] +; X64-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0] ; X64-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 ; X64-NEXT: retq entry: @@ -167,7 +167,7 @@ define <4 x double> @C2(double* %ptr, double* %ptr2) nounwind uwtable readnone s ; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx ; X32-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero ; X32-NEXT: vmovsd %xmm0, (%eax) -; X32-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1] +; X32-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0] ; X32-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 ; X32-NEXT: retl ; @@ -175,7 +175,7 @@ define <4 x double> @C2(double* %ptr, double* %ptr2) nounwind uwtable readnone s ; X64: ## BB#0: ## %entry ; X64-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero ; X64-NEXT: vmovsd %xmm0, (%rsi) -; X64-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,1,0,1] +; X64-NEXT: vmovddup {{.*#+}} xmm0 = xmm0[0,0] ; X64-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 ; X64-NEXT: retq entry: @@ -237,17 +237,17 @@ define <8 x float> @D3(float* %ptr, float* %ptr2) nounwind uwtable readnone ssp ; X32: ## BB#0: ## %entry ; X32-NEXT: movl {{[0-9]+}}(%esp), %eax ; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X32-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; X32-NEXT: vmovd %xmm0, (%eax) -; X32-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0] +; X32-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero +; X32-NEXT: vmovss %xmm0, (%eax) +; X32-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0] ; X32-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 ; X32-NEXT: retl ; ; X64-LABEL: D3: ; X64: ## BB#0: ## %entry -; X64-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; X64-NEXT: vmovd %xmm0, (%rsi) -; X64-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0] +; X64-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero +; X64-NEXT: vmovss %xmm0, (%rsi) +; X64-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0] ; X64-NEXT: vinsertf128 $1, %xmm0, %ymm0, %ymm0 ; X64-NEXT: retq entry: @@ -291,16 +291,16 @@ define <4 x float> @e2(float* %ptr, float* %ptr2) nounwind uwtable readnone ssp ; X32: ## BB#0: ## %entry ; X32-NEXT: movl {{[0-9]+}}(%esp), %eax ; X32-NEXT: movl {{[0-9]+}}(%esp), %ecx -; X32-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; X32-NEXT: vmovd %xmm0, (%eax) -; X32-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0] +; X32-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero +; X32-NEXT: vmovss %xmm0, (%eax) +; X32-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0] ; X32-NEXT: retl ; ; X64-LABEL: e2: ; X64: ## BB#0: ## %entry -; X64-NEXT: vmovd {{.*#+}} xmm0 = mem[0],zero,zero,zero -; X64-NEXT: vmovd %xmm0, (%rsi) -; X64-NEXT: vpshufd {{.*#+}} xmm0 = xmm0[0,0,0,0] +; X64-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero +; X64-NEXT: vmovss %xmm0, (%rsi) +; X64-NEXT: vpermilps {{.*#+}} xmm0 = xmm0[0,0,0,0] ; X64-NEXT: retq entry: %q = load float, float* %ptr, align 4 |
