1; RUN: llc -O3 -mtriple=x86_64-unknown -mcpu=core-avx2 -mattr=avx2 < %s | FileCheck %s
2
3define <8 x i16> @commute_fold_vpblendw_128(<8 x i16> %a, <8 x i16>* %b) #0 {
4  %1 = load <8 x i16>, <8 x i16>* %b
5  %2 = call <8 x i16> @llvm.x86.sse41.pblendw(<8 x i16> %1, <8 x i16> %a, i8 17)
6  ret <8 x i16> %2
7
8  ;LABEL:      commute_fold_vpblendw_128
9  ;CHECK:      vpblendw {{.*#+}} xmm0 = xmm0[0],mem[1,2,3],xmm0[4],mem[5,6,7]
10  ;CHECK-NEXT: retq
11}
12declare <8 x i16> @llvm.x86.sse41.pblendw(<8 x i16>, <8 x i16>, i8) nounwind readnone
13
14define <16 x i16> @commute_fold_vpblendw_256(<16 x i16> %a, <16 x i16>* %b) #0 {
15  %1 = load <16 x i16>, <16 x i16>* %b
16  %2 = call <16 x i16> @llvm.x86.avx2.pblendw(<16 x i16> %1, <16 x i16> %a, i8 17)
17  ret <16 x i16> %2
18
19  ;LABEL:      commute_fold_vpblendw_256
20  ;CHECK:      vpblendw {{.*#+}} ymm0 = ymm0[0],mem[1,2,3],ymm0[4],mem[5,6,7],ymm0[8],mem[9,10,11],ymm0[12],mem[13,14,15]
21  ;CHECK-NEXT: retq
22}
23declare <16 x i16> @llvm.x86.avx2.pblendw(<16 x i16>, <16 x i16>, i8) nounwind readnone
24
25define <4 x i32> @commute_fold_vpblendd_128(<4 x i32> %a, <4 x i32>* %b) #0 {
26  %1 = load <4 x i32>, <4 x i32>* %b
27  %2 = call <4 x i32> @llvm.x86.avx2.pblendd.128(<4 x i32> %1, <4 x i32> %a, i8 1)
28  ret <4 x i32> %2
29
30  ;LABEL:      commute_fold_vpblendd_128
31  ;CHECK:      vpblendd {{.*#+}} xmm0 = xmm0[0],mem[1,2,3]
32  ;CHECK-NEXT: retq
33}
34declare <4 x i32> @llvm.x86.avx2.pblendd.128(<4 x i32>, <4 x i32>, i8) nounwind readnone
35
36define <8 x i32> @commute_fold_vpblendd_256(<8 x i32> %a, <8 x i32>* %b) #0 {
37  %1 = load <8 x i32>, <8 x i32>* %b
38  %2 = call <8 x i32> @llvm.x86.avx2.pblendd.256(<8 x i32> %1, <8 x i32> %a, i8 129)
39  ret <8 x i32> %2
40
41  ;LABEL:      commute_fold_vpblendd_256
42  ;CHECK:      vpblendd {{.*#+}} ymm0 = ymm0[0],mem[1,2,3,4,5,6],ymm0[7]
43  ;CHECK-NEXT: retq
44}
45declare <8 x i32> @llvm.x86.avx2.pblendd.256(<8 x i32>, <8 x i32>, i8) nounwind readnone
46
47define <4 x float> @commute_fold_vblendps_128(<4 x float> %a, <4 x float>* %b) #0 {
48  %1 = load <4 x float>, <4 x float>* %b
49  %2 = call <4 x float> @llvm.x86.sse41.blendps(<4 x float> %1, <4 x float> %a, i8 5)
50  ret <4 x float> %2
51
52  ;LABEL:      commute_fold_vblendps_128
53  ;CHECK:      vblendps {{.*#+}} xmm0 = xmm0[0],mem[1],xmm0[2],mem[3]
54  ;CHECK-NEXT: retq
55}
56declare <4 x float> @llvm.x86.sse41.blendps(<4 x float>, <4 x float>, i8) nounwind readnone
57
58define <8 x float> @commute_fold_vblendps_256(<8 x float> %a, <8 x float>* %b) #0 {
59  %1 = load <8 x float>, <8 x float>* %b
60  %2 = call <8 x float> @llvm.x86.avx.blend.ps.256(<8 x float> %1, <8 x float> %a, i8 7)
61  ret <8 x float> %2
62
63  ;LABEL:      commute_fold_vblendps_256
64  ;CHECK:      vblendps {{.*#+}} ymm0 = ymm0[0,1,2],mem[3,4,5,6,7]
65  ;CHECK-NEXT: retq
66}
67declare <8 x float> @llvm.x86.avx.blend.ps.256(<8 x float>, <8 x float>, i8) nounwind readnone
68
69define <2 x double> @commute_fold_vblendpd_128(<2 x double> %a, <2 x double>* %b) #0 {
70  %1 = load <2 x double>, <2 x double>* %b
71  %2 = call <2 x double> @llvm.x86.sse41.blendpd(<2 x double> %1, <2 x double> %a, i8 1)
72  ret <2 x double> %2
73
74  ;LABEL:      commute_fold_vblendpd_128
75  ;CHECK:      vblendpd {{.*#+}} xmm0 = xmm0[0],mem[1]
76  ;CHECK-NEXT: retq
77}
78declare <2 x double> @llvm.x86.sse41.blendpd(<2 x double>, <2 x double>, i8) nounwind readnone
79
80define <4 x double> @commute_fold_vblendpd_256(<4 x double> %a, <4 x double>* %b) #0 {
81  %1 = load <4 x double>, <4 x double>* %b
82  %2 = call <4 x double> @llvm.x86.avx.blend.pd.256(<4 x double> %1, <4 x double> %a, i8 7)
83  ret <4 x double> %2
84
85  ;LABEL:      commute_fold_vblendpd_256
86  ;CHECK:      vblendpd {{.*#+}} ymm0 = ymm0[0,1,2],mem[3]
87  ;CHECK-NEXT: retq
88}
89declare <4 x double> @llvm.x86.avx.blend.pd.256(<4 x double>, <4 x double>, i8) nounwind readnone
90