1 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
2 ; RUN: llc < %s -mtriple=i386-apple-darwin -mattr=+sse3 -show-mc-encoding | FileCheck %s --check-prefixes=X86,SSE,X86-SSE
3 ; RUN: llc < %s -mtriple=i386-apple-darwin -mattr=+avx -show-mc-encoding | FileCheck %s --check-prefixes=X86,AVX,X86-AVX
4 ; RUN: llc < %s -mtriple=i386-apple-darwin -mattr=+avx512f,+avx512bw,+avx512dq,+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=X86,AVX,X86-AVX
5 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse3 -show-mc-encoding | FileCheck %s --check-prefixes=X64,SSE,X64-SSE
6 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx -show-mc-encoding | FileCheck %s --check-prefixes=X64,AVX,X64-AVX
7 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+avx512f,+avx512bw,+avx512dq,+avx512vl -show-mc-encoding | FileCheck %s --check-prefixes=X64,AVX,X64-AVX
9 define <2 x double> @test_x86_sse3_addsub_pd(<2 x double> %a0, <2 x double> %a1) {
10 ; SSE-LABEL: test_x86_sse3_addsub_pd:
12 ; SSE-NEXT: addsubpd %xmm1, %xmm0 ## encoding: [0x66,0x0f,0xd0,0xc1]
13 ; SSE-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
15 ; AVX-LABEL: test_x86_sse3_addsub_pd:
17 ; AVX-NEXT: vaddsubpd %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0xd0,0xc1]
18 ; AVX-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
19 %res = call <2 x double> @llvm.x86.sse3.addsub.pd(<2 x double> %a0, <2 x double> %a1) ; <<2 x double>> [#uses=1]
22 declare <2 x double> @llvm.x86.sse3.addsub.pd(<2 x double>, <2 x double>) nounwind readnone
25 define <4 x float> @test_x86_sse3_addsub_ps(<4 x float> %a0, <4 x float> %a1) {
26 ; SSE-LABEL: test_x86_sse3_addsub_ps:
28 ; SSE-NEXT: addsubps %xmm1, %xmm0 ## encoding: [0xf2,0x0f,0xd0,0xc1]
29 ; SSE-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
31 ; AVX-LABEL: test_x86_sse3_addsub_ps:
33 ; AVX-NEXT: vaddsubps %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xfb,0xd0,0xc1]
34 ; AVX-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
35 %res = call <4 x float> @llvm.x86.sse3.addsub.ps(<4 x float> %a0, <4 x float> %a1) ; <<4 x float>> [#uses=1]
38 declare <4 x float> @llvm.x86.sse3.addsub.ps(<4 x float>, <4 x float>) nounwind readnone
41 define <2 x double> @test_x86_sse3_hadd_pd(<2 x double> %a0, <2 x double> %a1) {
42 ; SSE-LABEL: test_x86_sse3_hadd_pd:
44 ; SSE-NEXT: haddpd %xmm1, %xmm0 ## encoding: [0x66,0x0f,0x7c,0xc1]
45 ; SSE-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
47 ; AVX-LABEL: test_x86_sse3_hadd_pd:
49 ; AVX-NEXT: vhaddpd %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0x7c,0xc1]
50 ; AVX-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
51 %res = call <2 x double> @llvm.x86.sse3.hadd.pd(<2 x double> %a0, <2 x double> %a1) ; <<2 x double>> [#uses=1]
54 declare <2 x double> @llvm.x86.sse3.hadd.pd(<2 x double>, <2 x double>) nounwind readnone
57 define <4 x float> @test_x86_sse3_hadd_ps(<4 x float> %a0, <4 x float> %a1) {
58 ; SSE-LABEL: test_x86_sse3_hadd_ps:
60 ; SSE-NEXT: haddps %xmm1, %xmm0 ## encoding: [0xf2,0x0f,0x7c,0xc1]
61 ; SSE-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
63 ; AVX-LABEL: test_x86_sse3_hadd_ps:
65 ; AVX-NEXT: vhaddps %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xfb,0x7c,0xc1]
66 ; AVX-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
67 %res = call <4 x float> @llvm.x86.sse3.hadd.ps(<4 x float> %a0, <4 x float> %a1) ; <<4 x float>> [#uses=1]
70 declare <4 x float> @llvm.x86.sse3.hadd.ps(<4 x float>, <4 x float>) nounwind readnone
73 define <2 x double> @test_x86_sse3_hsub_pd(<2 x double> %a0, <2 x double> %a1) {
74 ; SSE-LABEL: test_x86_sse3_hsub_pd:
76 ; SSE-NEXT: hsubpd %xmm1, %xmm0 ## encoding: [0x66,0x0f,0x7d,0xc1]
77 ; SSE-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
79 ; AVX-LABEL: test_x86_sse3_hsub_pd:
81 ; AVX-NEXT: vhsubpd %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xf9,0x7d,0xc1]
82 ; AVX-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
83 %res = call <2 x double> @llvm.x86.sse3.hsub.pd(<2 x double> %a0, <2 x double> %a1) ; <<2 x double>> [#uses=1]
86 declare <2 x double> @llvm.x86.sse3.hsub.pd(<2 x double>, <2 x double>) nounwind readnone
89 define <4 x float> @test_x86_sse3_hsub_ps(<4 x float> %a0, <4 x float> %a1) {
90 ; SSE-LABEL: test_x86_sse3_hsub_ps:
92 ; SSE-NEXT: hsubps %xmm1, %xmm0 ## encoding: [0xf2,0x0f,0x7d,0xc1]
93 ; SSE-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
95 ; AVX-LABEL: test_x86_sse3_hsub_ps:
97 ; AVX-NEXT: vhsubps %xmm1, %xmm0, %xmm0 ## encoding: [0xc5,0xfb,0x7d,0xc1]
98 ; AVX-NEXT: ret{{[l|q]}} ## encoding: [0xc3]
99 %res = call <4 x float> @llvm.x86.sse3.hsub.ps(<4 x float> %a0, <4 x float> %a1) ; <<4 x float>> [#uses=1]
102 declare <4 x float> @llvm.x86.sse3.hsub.ps(<4 x float>, <4 x float>) nounwind readnone
105 define <16 x i8> @test_x86_sse3_ldu_dq(ptr %a0) {
106 ; X86-SSE-LABEL: test_x86_sse3_ldu_dq:
108 ; X86-SSE-NEXT: movl {{[0-9]+}}(%esp), %eax ## encoding: [0x8b,0x44,0x24,0x04]
109 ; X86-SSE-NEXT: lddqu (%eax), %xmm0 ## encoding: [0xf2,0x0f,0xf0,0x00]
110 ; X86-SSE-NEXT: retl ## encoding: [0xc3]
112 ; X86-AVX-LABEL: test_x86_sse3_ldu_dq:
114 ; X86-AVX-NEXT: movl {{[0-9]+}}(%esp), %eax ## encoding: [0x8b,0x44,0x24,0x04]
115 ; X86-AVX-NEXT: vlddqu (%eax), %xmm0 ## encoding: [0xc5,0xfb,0xf0,0x00]
116 ; X86-AVX-NEXT: retl ## encoding: [0xc3]
118 ; X64-SSE-LABEL: test_x86_sse3_ldu_dq:
120 ; X64-SSE-NEXT: lddqu (%rdi), %xmm0 ## encoding: [0xf2,0x0f,0xf0,0x07]
121 ; X64-SSE-NEXT: retq ## encoding: [0xc3]
123 ; X64-AVX-LABEL: test_x86_sse3_ldu_dq:
125 ; X64-AVX-NEXT: vlddqu (%rdi), %xmm0 ## encoding: [0xc5,0xfb,0xf0,0x07]
126 ; X64-AVX-NEXT: retq ## encoding: [0xc3]
127 %res = call <16 x i8> @llvm.x86.sse3.ldu.dq(ptr %a0) ; <<16 x i8>> [#uses=1]
130 declare <16 x i8> @llvm.x86.sse3.ldu.dq(ptr) nounwind readonly
132 ; Make sure instructions with no AVX equivalents, but are associated with SSEX feature flags still work
134 define void @monitor(ptr %P, i32 %E, i32 %H) nounwind {
135 ; X86-LABEL: monitor:
137 ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ## encoding: [0x8b,0x44,0x24,0x04]
138 ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ## encoding: [0x8b,0x4c,0x24,0x08]
139 ; X86-NEXT: movl {{[0-9]+}}(%esp), %edx ## encoding: [0x8b,0x54,0x24,0x0c]
140 ; X86-NEXT: monitor ## encoding: [0x0f,0x01,0xc8]
141 ; X86-NEXT: retl ## encoding: [0xc3]
143 ; X64-LABEL: monitor:
145 ; X64-NEXT: movl %esi, %ecx ## encoding: [0x89,0xf1]
146 ; X64-NEXT: movq %rdi, %rax ## encoding: [0x48,0x89,0xf8]
147 ; X64-NEXT: monitor ## encoding: [0x0f,0x01,0xc8]
148 ; X64-NEXT: retq ## encoding: [0xc3]
149 tail call void @llvm.x86.sse3.monitor(ptr %P, i32 %E, i32 %H)
152 declare void @llvm.x86.sse3.monitor(ptr, i32, i32) nounwind
154 define void @mwait(i32 %E, i32 %H) nounwind {
157 ; X86-NEXT: movl {{[0-9]+}}(%esp), %ecx ## encoding: [0x8b,0x4c,0x24,0x04]
158 ; X86-NEXT: movl {{[0-9]+}}(%esp), %eax ## encoding: [0x8b,0x44,0x24,0x08]
159 ; X86-NEXT: mwait ## encoding: [0x0f,0x01,0xc9]
160 ; X86-NEXT: retl ## encoding: [0xc3]
164 ; X64-NEXT: movl %esi, %eax ## encoding: [0x89,0xf0]
165 ; X64-NEXT: movl %edi, %ecx ## encoding: [0x89,0xf9]
166 ; X64-NEXT: mwait ## encoding: [0x0f,0x01,0xc9]
167 ; X64-NEXT: retq ## encoding: [0xc3]
168 tail call void @llvm.x86.sse3.mwait(i32 %E, i32 %H)
171 declare void @llvm.x86.sse3.mwait(i32, i32) nounwind