1; RUN: llc < %s -mattr=+sse2 -mtriple=i686-apple-darwin -mcpu=core2 | FileCheck %s -check-prefix=SSE2 2; RUN: llc < %s -mattr=+sse,-sse2 -mtriple=i686-apple-darwin -mcpu=core2 | FileCheck %s -check-prefix=SSE1 3; RUN: llc < %s -mattr=-sse -mtriple=i686-apple-darwin -mcpu=core2 | FileCheck %s -check-prefix=NOSSE 4; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=core2 | FileCheck %s -check-prefix=X86-64 5 6@.str = internal constant [25 x i8] c"image\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00\00" 7@.str2 = internal constant [30 x i8] c"xxxxxxxxxxxxxxxxxxxxxxxxxxxxx\00", align 4 8 9define void @t1(i32 %argc, i8** %argv) nounwind { 10entry: 11; SSE2: t1: 12; SSE2: movaps _.str, %xmm0 13; SSE2: movaps %xmm0 14; SSE2: movb $0 15; SSE2: movl $0 16; SSE2: movl $0 17 18; SSE1: t1: 19; SSE1: movaps _.str, %xmm0 20; SSE1: movaps %xmm0 21; SSE1: movb $0 22; SSE1: movl $0 23; SSE1: movl $0 24 25; NOSSE: t1: 26; NOSSE: movb $0 27; NOSSE: movl $0 28; NOSSE: movl $0 29; NOSSE: movl $0 30; NOSSE: movl $0 31; NOSSE: movl $101 32; NOSSE: movl $1734438249 33 34; X86-64: t1: 35; X86-64: movaps _.str(%rip), %xmm0 36; X86-64: movaps %xmm0 37; X86-64: movb $0 38; X86-64: movq $0 39 %tmp1 = alloca [25 x i8] 40 %tmp2 = bitcast [25 x i8]* %tmp1 to i8* 41 call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* getelementptr inbounds ([25 x i8]* @.str, i32 0, i32 0), i32 25, i32 1, i1 false) 42 unreachable 43} 44 45;rdar://7774704 46%struct.s0 = type { [2 x double] } 47 48define void @t2(%struct.s0* nocapture %a, %struct.s0* nocapture %b) nounwind ssp { 49entry: 50; SSE2: t2: 51; SSE2: movaps (%eax), %xmm0 52; SSE2: movaps %xmm0, (%eax) 53 54; SSE1: t2: 55; SSE1: movaps (%eax), %xmm0 56; SSE1: movaps %xmm0, (%eax) 57 58; NOSSE: t2: 59; NOSSE: movl 60; NOSSE: movl 61; NOSSE: movl 62; NOSSE: movl 63; NOSSE: movl 64; NOSSE: movl 65; NOSSE: movl 66; NOSSE: movl 67; NOSSE: movl 68; NOSSE: movl 69 70; X86-64: t2: 71; X86-64: movaps (%rsi), %xmm0 72; X86-64: movaps %xmm0, (%rdi) 73 %tmp2 = bitcast %struct.s0* %a to i8* ; <i8*> [#uses=1] 74 %tmp3 = bitcast %struct.s0* %b to i8* ; <i8*> [#uses=1] 75 tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* %tmp3, i32 16, i32 16, i1 false) 76 ret void 77} 78 79define void @t3(%struct.s0* nocapture %a, %struct.s0* nocapture %b) nounwind ssp { 80entry: 81; SSE2: t3: 82; SSE2: movsd (%eax), %xmm0 83; SSE2: movsd 8(%eax), %xmm1 84; SSE2: movsd %xmm1, 8(%eax) 85; SSE2: movsd %xmm0, (%eax) 86 87; SSE1: t3: 88; SSE1: movl 89; SSE1: movl 90; SSE1: movl 91; SSE1: movl 92; SSE1: movl 93; SSE1: movl 94; SSE1: movl 95; SSE1: movl 96; SSE1: movl 97; SSE1: movl 98 99; NOSSE: t3: 100; NOSSE: movl 101; NOSSE: movl 102; NOSSE: movl 103; NOSSE: movl 104; NOSSE: movl 105; NOSSE: movl 106; NOSSE: movl 107; NOSSE: movl 108; NOSSE: movl 109; NOSSE: movl 110 111; X86-64: t3: 112; X86-64: movq (%rsi), %rax 113; X86-64: movq 8(%rsi), %rcx 114; X86-64: movq %rcx, 8(%rdi) 115; X86-64: movq %rax, (%rdi) 116 %tmp2 = bitcast %struct.s0* %a to i8* ; <i8*> [#uses=1] 117 %tmp3 = bitcast %struct.s0* %b to i8* ; <i8*> [#uses=1] 118 tail call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* %tmp3, i32 16, i32 8, i1 false) 119 ret void 120} 121 122define void @t4() nounwind { 123entry: 124; SSE2: t4: 125; SSE2: movw $120 126; SSE2: movl $2021161080 127; SSE2: movl $2021161080 128; SSE2: movl $2021161080 129; SSE2: movl $2021161080 130; SSE2: movl $2021161080 131; SSE2: movl $2021161080 132; SSE2: movl $2021161080 133 134; SSE1: t4: 135; SSE1: movw $120 136; SSE1: movl $2021161080 137; SSE1: movl $2021161080 138; SSE1: movl $2021161080 139; SSE1: movl $2021161080 140; SSE1: movl $2021161080 141; SSE1: movl $2021161080 142; SSE1: movl $2021161080 143 144; NOSSE: t4: 145; NOSSE: movw $120 146; NOSSE: movl $2021161080 147; NOSSE: movl $2021161080 148; NOSSE: movl $2021161080 149; NOSSE: movl $2021161080 150; NOSSE: movl $2021161080 151; NOSSE: movl $2021161080 152; NOSSE: movl $2021161080 153 154; X86-64: t4: 155; X86-64: movabsq $8680820740569200760, %rax 156; X86-64: movq %rax 157; X86-64: movq %rax 158; X86-64: movq %rax 159; X86-64: movw $120 160; X86-64: movl $2021161080 161 %tmp1 = alloca [30 x i8] 162 %tmp2 = bitcast [30 x i8]* %tmp1 to i8* 163 call void @llvm.memcpy.p0i8.p0i8.i32(i8* %tmp2, i8* getelementptr inbounds ([30 x i8]* @.str2, i32 0, i32 0), i32 30, i32 1, i1 false) 164 unreachable 165} 166 167declare void @llvm.memcpy.p0i8.p0i8.i32(i8* nocapture, i8* nocapture, i32, i32, i1) nounwind 168