1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc -mtriple=x86_64-unknown-unknown -mattr=+sse2 < %s | FileCheck %s --check-prefix=SSE
3 ; RUN: llc -mtriple=x86_64-unknown-unknown -mattr=+avx < %s | FileCheck %s --check-prefix=AVX
5 ; Verify that we're folding the load into the math instruction.
7 define float @rcpss(float* %a) {
10 ; SSE-NEXT: rcpss (%rdi), %xmm0
15 ; AVX-NEXT: vrcpss (%rdi), %xmm0, %xmm0
17 %ld = load float, float* %a
18 %ins = insertelement <4 x float> undef, float %ld, i32 0
19 %res = tail call <4 x float> @llvm.x86.sse.rcp.ss(<4 x float> %ins)
20 %ext = extractelement <4 x float> %res, i32 0
24 define float @rsqrtss(float* %a) {
27 ; SSE-NEXT: rsqrtss (%rdi), %xmm0
32 ; AVX-NEXT: vrsqrtss (%rdi), %xmm0, %xmm0
34 %ld = load float, float* %a
35 %ins = insertelement <4 x float> undef, float %ld, i32 0
36 %res = tail call <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float> %ins)
37 %ext = extractelement <4 x float> %res, i32 0
41 define float @sqrtss(float* %a) {
44 ; SSE-NEXT: sqrtss (%rdi), %xmm0
49 ; AVX-NEXT: vsqrtss (%rdi), %xmm0, %xmm0
51 %ld = load float, float* %a
52 %ins = insertelement <4 x float> undef, float %ld, i32 0
53 %res = tail call <4 x float> @llvm.x86.sse.sqrt.ss(<4 x float> %ins)
54 %ext = extractelement <4 x float> %res, i32 0
58 define double @sqrtsd(double* %a) {
61 ; SSE-NEXT: sqrtsd (%rdi), %xmm0
66 ; AVX-NEXT: vsqrtsd (%rdi), %xmm0, %xmm0
68 %ld = load double, double* %a
69 %ins = insertelement <2 x double> undef, double %ld, i32 0
70 %res = tail call <2 x double> @llvm.x86.sse2.sqrt.sd(<2 x double> %ins)
71 %ext = extractelement <2 x double> %res, i32 0
76 declare <4 x float> @llvm.x86.sse.rcp.ss(<4 x float>) nounwind readnone
77 declare <4 x float> @llvm.x86.sse.rsqrt.ss(<4 x float>) nounwind readnone
78 declare <4 x float> @llvm.x86.sse.sqrt.ss(<4 x float>) nounwind readnone
79 declare <2 x double> @llvm.x86.sse2.sqrt.sd(<2 x double>) nounwind readnone