1; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py 2; RUN: llc < %s -O3 -disable-peephole -mtriple=i686-apple-macosx10.9.0 -mcpu=corei7-avx -mattr=+avx | FileCheck %s --check-prefix=X86 3; RUN: llc < %s -O3 -disable-peephole -mtriple=x86_64-apple-macosx10.9.0 -mcpu=corei7-avx -mattr=+avx | FileCheck %s --check-prefix=X64 4 5; Function Attrs: nounwind ssp uwtable 6define void @test1(float* %A, float* %C) #0 { 7; X86-LABEL: test1: 8; X86: ## %bb.0: 9; X86-NEXT: movl {{[0-9]+}}(%esp), %eax 10; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx 11; X86-NEXT: vmovaps (%ecx), %ymm0 12; X86-NEXT: vandps LCPI0_0, %ymm0, %ymm0 13; X86-NEXT: vmovss %xmm0, (%eax) 14; X86-NEXT: vzeroupper 15; X86-NEXT: retl 16; 17; X64-LABEL: test1: 18; X64: ## %bb.0: 19; X64-NEXT: vmovaps (%rdi), %ymm0 20; X64-NEXT: vandps {{.*}}(%rip), %ymm0, %ymm0 21; X64-NEXT: vmovss %xmm0, (%rsi) 22; X64-NEXT: vzeroupper 23; X64-NEXT: retq 24 %tmp1 = bitcast float* %A to <8 x float>* 25 %tmp2 = load <8 x float>, <8 x float>* %tmp1, align 32 26 %tmp3 = bitcast <8 x float> %tmp2 to <8 x i32> 27 %tmp4 = and <8 x i32> %tmp3, <i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647> 28 %tmp5 = bitcast <8 x i32> %tmp4 to <8 x float> 29 %tmp6 = extractelement <8 x float> %tmp5, i32 0 30 store float %tmp6, float* %C 31 ret void 32} 33 34; Function Attrs: nounwind ssp uwtable 35define void @test2(float* %A, float* %C) #0 { 36; X86-LABEL: test2: 37; X86: ## %bb.0: 38; X86-NEXT: movl {{[0-9]+}}(%esp), %eax 39; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx 40; X86-NEXT: vmovaps (%ecx), %ymm0 41; X86-NEXT: vorps LCPI1_0, %ymm0, %ymm0 42; X86-NEXT: vmovss %xmm0, (%eax) 43; X86-NEXT: vzeroupper 44; X86-NEXT: retl 45; 46; X64-LABEL: test2: 47; X64: ## %bb.0: 48; X64-NEXT: vmovaps (%rdi), %ymm0 49; X64-NEXT: vorps {{.*}}(%rip), %ymm0, %ymm0 50; X64-NEXT: vmovss %xmm0, (%rsi) 51; X64-NEXT: vzeroupper 52; X64-NEXT: retq 53 %tmp1 = bitcast float* %A to <8 x float>* 54 %tmp2 = load <8 x float>, <8 x float>* %tmp1, align 32 55 %tmp3 = bitcast <8 x float> %tmp2 to <8 x i32> 56 %tmp4 = or <8 x i32> %tmp3, <i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647> 57 %tmp5 = bitcast <8 x i32> %tmp4 to <8 x float> 58 %tmp6 = extractelement <8 x float> %tmp5, i32 0 59 store float %tmp6, float* %C 60 ret void 61} 62 63; Function Attrs: nounwind ssp uwtable 64define void @test3(float* %A, float* %C) #0 { 65; X86-LABEL: test3: 66; X86: ## %bb.0: 67; X86-NEXT: movl {{[0-9]+}}(%esp), %eax 68; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx 69; X86-NEXT: vmovaps (%ecx), %ymm0 70; X86-NEXT: vxorps LCPI2_0, %ymm0, %ymm0 71; X86-NEXT: vmovss %xmm0, (%eax) 72; X86-NEXT: vzeroupper 73; X86-NEXT: retl 74; 75; X64-LABEL: test3: 76; X64: ## %bb.0: 77; X64-NEXT: vmovaps (%rdi), %ymm0 78; X64-NEXT: vxorps {{.*}}(%rip), %ymm0, %ymm0 79; X64-NEXT: vmovss %xmm0, (%rsi) 80; X64-NEXT: vzeroupper 81; X64-NEXT: retq 82 %tmp1 = bitcast float* %A to <8 x float>* 83 %tmp2 = load <8 x float>, <8 x float>* %tmp1, align 32 84 %tmp3 = bitcast <8 x float> %tmp2 to <8 x i32> 85 %tmp4 = xor <8 x i32> %tmp3, <i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647> 86 %tmp5 = bitcast <8 x i32> %tmp4 to <8 x float> 87 %tmp6 = extractelement <8 x float> %tmp5, i32 0 88 store float %tmp6, float* %C 89 ret void 90} 91 92define void @test4(float* %A, float* %C) #0 { 93; X86-LABEL: test4: 94; X86: ## %bb.0: 95; X86-NEXT: movl {{[0-9]+}}(%esp), %eax 96; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx 97; X86-NEXT: vmovaps (%ecx), %ymm0 98; X86-NEXT: vandnps LCPI3_0, %ymm0, %ymm0 99; X86-NEXT: vmovss %xmm0, (%eax) 100; X86-NEXT: vzeroupper 101; X86-NEXT: retl 102; 103; X64-LABEL: test4: 104; X64: ## %bb.0: 105; X64-NEXT: vmovaps (%rdi), %ymm0 106; X64-NEXT: vandnps {{.*}}(%rip), %ymm0, %ymm0 107; X64-NEXT: vmovss %xmm0, (%rsi) 108; X64-NEXT: vzeroupper 109; X64-NEXT: retq 110 %tmp1 = bitcast float* %A to <8 x float>* 111 %tmp2 = load <8 x float>, <8 x float>* %tmp1, align 32 112 %tmp3 = bitcast <8 x float> %tmp2 to <8 x i32> 113 %tmp4 = xor <8 x i32> %tmp3, <i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1> 114 %tmp5 = and <8 x i32> %tmp4, <i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647, i32 2147483647> 115 %tmp6 = bitcast <8 x i32> %tmp5 to <8 x float> 116 %tmp7 = extractelement <8 x float> %tmp6, i32 0 117 store float %tmp7, float * %C 118 ret void 119} 120