1// RUN: %clang_cc1 -ffreestanding -triple i386-apple-darwin9 -target-cpu pentium4 -target-feature +sse4.1 -g -emit-llvm %s -o - | FileCheck %s
2
3#include <xmmintrin.h>
4#include <emmintrin.h>
5#include <smmintrin.h>
6
7__m128 test_rsqrt_ss(__m128 x) {
8  // CHECK: define {{.*}} @test_rsqrt_ss
9  // CHECK: call <4 x float> @llvm.x86.sse.rsqrt.ss
10  // CHECK: extractelement <4 x float> {{.*}}, i32 0
11  // CHECK: extractelement <4 x float> {{.*}}, i32 1
12  // CHECK: extractelement <4 x float> {{.*}}, i32 2
13  // CHECK: extractelement <4 x float> {{.*}}, i32 3
14  return _mm_rsqrt_ss(x);
15}
16
17__m128 test_rcp_ss(__m128 x) {
18  // CHECK: define {{.*}} @test_rcp_ss
19  // CHECK: call <4 x float> @llvm.x86.sse.rcp.ss
20  // CHECK: extractelement <4 x float> {{.*}}, i32 0
21  // CHECK: extractelement <4 x float> {{.*}}, i32 1
22  // CHECK: extractelement <4 x float> {{.*}}, i32 2
23  // CHECK: extractelement <4 x float> {{.*}}, i32 3
24  return _mm_rcp_ss(x);
25}
26
27__m128 test_sqrt_ss(__m128 x) {
28  // CHECK: define {{.*}} @test_sqrt_ss
29  // CHECK: call <4 x float> @llvm.x86.sse.sqrt.ss
30  // CHECK: extractelement <4 x float> {{.*}}, i32 0
31  // CHECK: extractelement <4 x float> {{.*}}, i32 1
32  // CHECK: extractelement <4 x float> {{.*}}, i32 2
33  // CHECK: extractelement <4 x float> {{.*}}, i32 3
34  return _mm_sqrt_ss(x);
35}
36
37__m128 test_loadl_pi(__m128 x, void* y) {
38  // CHECK: define {{.*}} @test_loadl_pi
39  // CHECK: load <2 x float>* {{.*}}, align 1{{$}}
40  // CHECK: shufflevector {{.*}} <4 x i32> <i32 0, i32 1
41  // CHECK: shufflevector {{.*}} <4 x i32> <i32 4, i32 5, i32 2, i32 3>
42  return _mm_loadl_pi(x,y);
43}
44
45__m128 test_loadh_pi(__m128 x, void* y) {
46  // CHECK: define {{.*}} @test_loadh_pi
47  // CHECK: load <2 x float>* {{.*}}, align 1{{$}}
48  // CHECK: shufflevector {{.*}} <4 x i32> <i32 0, i32 1
49  // CHECK: shufflevector {{.*}} <4 x i32> <i32 0, i32 1, i32 4, i32 5>
50  return _mm_loadh_pi(x,y);
51}
52
53__m128 test_load_ss(void* y) {
54  // CHECK: define {{.*}} @test_load_ss
55  // CHECK: load float* {{.*}}, align 1{{$}}
56  return _mm_load_ss(y);
57}
58
59__m128 test_load1_ps(void* y) {
60  // CHECK: define {{.*}} @test_load1_ps
61  // CHECK: load float* {{.*}}, align 1{{$}}
62  return _mm_load1_ps(y);
63}
64
65void test_store_ss(__m128 x, void* y) {
66  // CHECK: define void @test_store_ss
67  // CHECK: store {{.*}} float* {{.*}}, align 1,
68  _mm_store_ss(y, x);
69}
70
71__m128d test_load1_pd(__m128 x, void* y) {
72  // CHECK: define {{.*}} @test_load1_pd
73  // CHECK: load double* {{.*}}, align 1{{$}}
74  return _mm_load1_pd(y);
75}
76
77__m128d test_loadr_pd(__m128 x, void* y) {
78  // CHECK: define {{.*}} @test_loadr_pd
79  // CHECK: load <2 x double>* {{.*}}, align 16{{$}}
80  return _mm_loadr_pd(y);
81}
82
83__m128d test_load_sd(void* y) {
84  // CHECK: define {{.*}} @test_load_sd
85  // CHECK: load double* {{.*}}, align 1{{$}}
86  return _mm_load_sd(y);
87}
88
89__m128d test_loadh_pd(__m128d x, void* y) {
90  // CHECK: define {{.*}} @test_loadh_pd
91  // CHECK: load double* {{.*}}, align 1{{$}}
92  return _mm_loadh_pd(x, y);
93}
94
95__m128d test_loadl_pd(__m128d x, void* y) {
96  // CHECK: define {{.*}} @test_loadl_pd
97  // CHECK: load double* {{.*}}, align 1{{$}}
98  return _mm_loadl_pd(x, y);
99}
100
101void test_store_sd(__m128d x, void* y) {
102  // CHECK: define void @test_store_sd
103  // CHECK: store {{.*}} double* {{.*}}, align 1{{$}}
104  _mm_store_sd(y, x);
105}
106
107void test_store1_pd(__m128d x, void* y) {
108  // CHECK: define void @test_store1_pd
109  // CHECK: store {{.*}} double* {{.*}}, align 1{{$}}
110  // CHECK: store {{.*}} double* {{.*}}, align 1{{$}}
111  _mm_store1_pd(y, x);
112}
113
114void test_storer_pd(__m128d x, void* y) {
115  // CHECK: define void @test_storer_pd
116  // CHECK: store {{.*}} <2 x double>* {{.*}}, align 16{{$}}
117  _mm_storer_pd(y, x);
118}
119
120void test_storeh_pd(__m128d x, void* y) {
121  // CHECK: define void @test_storeh_pd
122  // CHECK: store {{.*}} double* {{.*}}, align 1{{$}}
123  _mm_storeh_pd(y, x);
124}
125
126void test_storel_pd(__m128d x, void* y) {
127  // CHECK: define void @test_storel_pd
128  // CHECK: store {{.*}} double* {{.*}}, align 1{{$}}
129  _mm_storel_pd(y, x);
130}
131
132__m128i test_loadl_epi64(void* y) {
133  // CHECK: define {{.*}} @test_loadl_epi64
134  // CHECK: load i64* {{.*}}, align 1{{$}}
135  return _mm_loadl_epi64(y);
136}
137
138__m128i test_mm_minpos_epu16(__m128i x) {
139  // CHECK: define {{.*}} @test_mm_minpos_epu16
140  // CHECK: @llvm.x86.sse41.phminposuw
141  return _mm_minpos_epu16(x);
142}
143
144__m128i test_mm_mpsadbw_epu8(__m128i x, __m128i y) {
145  // CHECK: define {{.*}} @test_mm_mpsadbw_epu8
146  // CHECK: @llvm.x86.sse41.mpsadbw
147  return _mm_mpsadbw_epu8(x, y, 1);
148}
149
150__m128 test_mm_dp_ps(__m128 x, __m128 y) {
151  // CHECK: define {{.*}} @test_mm_dp_ps
152  // CHECK: @llvm.x86.sse41.dpps
153  return _mm_dp_ps(x, y, 2);
154}
155
156__m128d test_mm_dp_pd(__m128d x, __m128d y) {
157  // CHECK: define {{.*}} @test_mm_dp_pd
158  // CHECK: @llvm.x86.sse41.dppd
159  return _mm_dp_pd(x, y, 2);
160}
161
162__m128 test_mm_round_ps(__m128 x) {
163  // CHECK: define {{.*}} @test_mm_round_ps
164  // CHECK: @llvm.x86.sse41.round.ps
165  return _mm_round_ps(x, 2);
166}
167
168__m128 test_mm_round_ss(__m128 x, __m128 y) {
169  // CHECK: define {{.*}} @test_mm_round_ss
170  // CHECK: @llvm.x86.sse41.round.ss
171  return _mm_round_ss(x, y, 2);
172}
173
174__m128d test_mm_round_pd(__m128d x) {
175  // CHECK: define {{.*}} @test_mm_round_pd
176  // CHECK: @llvm.x86.sse41.round.pd
177  return _mm_round_pd(x, 2);
178}
179
180__m128d test_mm_round_sd(__m128d x, __m128d y) {
181  // CHECK: define {{.*}} @test_mm_round_sd
182  // CHECK: @llvm.x86.sse41.round.sd
183  return _mm_round_sd(x, y, 2);
184}
185
186void test_storel_epi64(__m128i x, void* y) {
187  // CHECK: define void @test_storel_epi64
188  // CHECK: store {{.*}} i64* {{.*}}, align 1{{$}}
189  _mm_storel_epi64(y, x);
190}
191