1 ; Tests for SSE2 and below, without SSE3+.
2 ; RUN: llc < %s -mtriple=i386-apple-darwin10 -mcpu=pentium4 | FileCheck %s
4 define void @t1(<2 x double>* %r, <2 x double>* %A, double %B) nounwind {
5 %tmp3 = load <2 x double>* %A, align 16
6 %tmp7 = insertelement <2 x double> undef, double %B, i32 0
7 %tmp9 = shufflevector <2 x double> %tmp3, <2 x double> %tmp7, <2 x i32> < i32 2, i32 1 >
8 store <2 x double> %tmp9, <2 x double>* %r, align 16
12 ; CHECK: movl 8(%esp), %eax
13 ; CHECK-NEXT: movapd (%eax), %xmm0
14 ; CHECK-NEXT: movlpd 12(%esp), %xmm0
15 ; CHECK-NEXT: movl 4(%esp), %eax
16 ; CHECK-NEXT: movapd %xmm0, (%eax)
20 define void @t2(<2 x double>* %r, <2 x double>* %A, double %B) nounwind {
21 %tmp3 = load <2 x double>* %A, align 16
22 %tmp7 = insertelement <2 x double> undef, double %B, i32 0
23 %tmp9 = shufflevector <2 x double> %tmp3, <2 x double> %tmp7, <2 x i32> < i32 0, i32 2 >
24 store <2 x double> %tmp9, <2 x double>* %r, align 16
28 ; CHECK: movl 8(%esp), %eax
29 ; CHECK-NEXT: movapd (%eax), %xmm0
30 ; CHECK-NEXT: movhpd 12(%esp), %xmm0
31 ; CHECK-NEXT: movl 4(%esp), %eax
32 ; CHECK-NEXT: movapd %xmm0, (%eax)