1 ; RUN: opt < %s -scalarrepl -S > %t
2 ; RUN: grep {ret <16 x float> %A} %t
3 ; RUN: grep {ret <16 x float> zeroinitializer} %t
4 target datalayout = "E-p:64:64:64-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:64:64-v64:64:64-v128:128:128-a0:0:64"
6 define <16 x float> @foo(<16 x float> %A) nounwind {
7 %tmp = alloca <16 x float>, align 16
8 %tmp2 = alloca <16 x float>, align 16
9 store <16 x float> %A, <16 x float>* %tmp
10 %s = bitcast <16 x float>* %tmp to i8*
11 %s2 = bitcast <16 x float>* %tmp2 to i8*
12 call void @llvm.memcpy.i64(i8* %s2, i8* %s, i64 64, i32 16)
14 %R = load <16 x float>* %tmp2
18 define <16 x float> @foo2(<16 x float> %A) nounwind {
19 %tmp2 = alloca <16 x float>, align 16
21 %s2 = bitcast <16 x float>* %tmp2 to i8*
22 call void @llvm.memset.i64(i8* %s2, i8 0, i64 64, i32 16)
24 %R = load <16 x float>* %tmp2
29 declare void @llvm.memcpy.i64(i8* nocapture, i8* nocapture, i64, i32) nounwind
30 declare void @llvm.memset.i64(i8* nocapture, i8, i64, i32) nounwind