1; RUN: llc < %s -mtriple=x86_64-unknown-linux -mcpu=corei7-avx | FileCheck %s
2
3; These tests check that an insert_subvector which replaces one of the halves
4; of a concat_vectors is optimized into a single vinsertf128.
5
6
7declare <8 x float> @llvm.x86.avx.vinsertf128.ps.256(<8 x float>, <4 x float>, i8)
8
9define <8 x float> @lower_half(<4 x float> %v1, <4 x float> %v2, <4 x float> %v3) {
10  %1 = shufflevector <4 x float> %v1, <4 x float> %v2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
11  %2 = tail call <8 x float> @llvm.x86.avx.vinsertf128.ps.256(<8 x float> %1, <4 x float> %v3, i8 0)
12  ret <8 x float> %2
13
14; CHECK-LABEL: lower_half
15; CHECK-NOT: vinsertf128
16; CHECK: vinsertf128 $1, %xmm1, %ymm2, %ymm0
17; CHECK-NEXT: ret
18}
19
20define <8 x float> @upper_half(<4 x float> %v1, <4 x float> %v2, <4 x float> %v3) {
21  %1 = shufflevector <4 x float> %v1, <4 x float> %v2, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>
22  %2 = tail call <8 x float> @llvm.x86.avx.vinsertf128.ps.256(<8 x float> %1, <4 x float> %v3, i8 1)
23  ret <8 x float> %2
24
25; CHECK-LABEL: upper_half
26; CHECK-NOT: vinsertf128
27; CHECK: vinsertf128 $1, %xmm2, %ymm0, %ymm0
28; CHECK-NEXT: ret
29}
30