1 ; RUN: llc < %s -march=x86-64 -mtriple=x86_64-apple-darwin -mcpu=knl | FileCheck %s
3 define <16 x i32> @test1(i32* %x) {
6 ; CHECK-NEXT: vmovd (%rdi), %xmm0
7 ; CHECK-NEXT: vmovdqa32 {{.*}}(%rip), %zmm1
8 ; CHECK-NEXT: vpxord %zmm2, %zmm2, %zmm2
9 ; CHECK-NEXT: vpermt2d %zmm0, %zmm1, %zmm2
10 ; CHECK-NEXT: vmovdqa32 {{.*}}(%rip), %zmm0
11 ; CHECK-NEXT: vpermd %zmm2, %zmm0, %zmm0
13 %y = load i32* %x, align 4
14 %res = insertelement <16 x i32>zeroinitializer, i32 %y, i32 4
18 define <16 x i32> @test2(<16 x i32> %x) {
21 ; CHECK-NEXT: vpaddd {{.*}}(%rip){1to16}, %zmm0, %zmm0
23 %res = add <16 x i32><i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1, i32 -1>, %x
27 define <16 x float> @test3(<4 x float> %a) {
30 ; CHECK-NEXT: vmovhlps %xmm0, %xmm0, %xmm1
31 ; CHECK-NEXT: vxorps %xmm2, %xmm2, %xmm2
32 ; CHECK-NEXT: vmovss %xmm0, %xmm2, %xmm0
33 ; CHECK-NEXT: vmovss %xmm1, %xmm2, %xmm1
34 ; CHECK-NEXT: vshufps {{.*#+}} xmm0 = xmm1[1,0],xmm0[0,1]
35 ; CHECK-NEXT: vinsertf128 $1, %xmm0, %ymm2, %ymm0
36 ; CHECK-NEXT: vxorps %ymm1, %ymm1, %ymm1
37 ; CHECK-NEXT: vinsertf64x4 $1, %ymm1, %zmm0, %zmm0
39 %b = extractelement <4 x float> %a, i32 2
40 %c = insertelement <16 x float> <float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float undef, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00, float 0.000000e+00>, float %b, i32 5
41 %b1 = extractelement <4 x float> %a, i32 0
42 %c1 = insertelement <16 x float> %c, float %b1, i32 6