| 1 | // RUN: %clang_cc1 -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse4.1 -emit-llvm -o - -Wall -Werror | FileCheck %s |
| 2 | // RUN: %clang_cc1 -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse4.1 -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s |
| 3 | |
| 4 | |
| 5 | #include <immintrin.h> |
| 6 | |
| 7 | // NOTE: This should match the tests in llvm/test/CodeGen/X86/sse41-intrinsics-fast-isel.ll |
| 8 | |
| 9 | __m128i test_mm_blend_epi16(__m128i V1, __m128i V2) { |
| 10 | // CHECK-LABEL: test_mm_blend_epi16 |
| 11 | // CHECK: shufflevector <8 x i16> %{{.*}}, <8 x i16> %{{.*}}, <8 x i32> <i32 0, i32 9, i32 2, i32 11, i32 4, i32 13, i32 6, i32 7> |
| 12 | return _mm_blend_epi16(V1, V2, 42); |
| 13 | } |
| 14 | |
| 15 | __m128d test_mm_blend_pd(__m128d V1, __m128d V2) { |
| 16 | // CHECK-LABEL: test_mm_blend_pd |
| 17 | // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x i32> <i32 0, i32 3> |
| 18 | return _mm_blend_pd(V1, V2, 2); |
| 19 | } |
| 20 | |
| 21 | __m128 test_mm_blend_ps(__m128 V1, __m128 V2) { |
| 22 | // CHECK-LABEL: test_mm_blend_ps |
| 23 | // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 5, i32 6, i32 3> |
| 24 | return _mm_blend_ps(V1, V2, 6); |
| 25 | } |
| 26 | |
| 27 | __m128i test_mm_blendv_epi8(__m128i V1, __m128i V2, __m128i V3) { |
| 28 | // CHECK-LABEL: test_mm_blendv_epi8 |
| 29 | // CHECK: call <16 x i8> @llvm.x86.sse41.pblendvb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}}) |
| 30 | return _mm_blendv_epi8(V1, V2, V3); |
| 31 | } |
| 32 | |
| 33 | __m128d test_mm_blendv_pd(__m128d V1, __m128d V2, __m128d V3) { |
| 34 | // CHECK-LABEL: test_mm_blendv_pd |
| 35 | // CHECK: call <2 x double> @llvm.x86.sse41.blendvpd(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}}) |
| 36 | return _mm_blendv_pd(V1, V2, V3); |
| 37 | } |
| 38 | |
| 39 | __m128 test_mm_blendv_ps(__m128 V1, __m128 V2, __m128 V3) { |
| 40 | // CHECK-LABEL: test_mm_blendv_ps |
| 41 | // CHECK: call <4 x float> @llvm.x86.sse41.blendvps(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}}) |
| 42 | return _mm_blendv_ps(V1, V2, V3); |
| 43 | } |
| 44 | |
| 45 | __m128d test_mm_ceil_pd(__m128d x) { |
| 46 | // CHECK-LABEL: test_mm_ceil_pd |
| 47 | // CHECK: call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> %{{.*}}, i32 2) |
| 48 | return _mm_ceil_pd(x); |
| 49 | } |
| 50 | |
| 51 | __m128 test_mm_ceil_ps(__m128 x) { |
| 52 | // CHECK-LABEL: test_mm_ceil_ps |
| 53 | // CHECK: call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> %{{.*}}, i32 2) |
| 54 | return _mm_ceil_ps(x); |
| 55 | } |
| 56 | |
| 57 | __m128d test_mm_ceil_sd(__m128d x, __m128d y) { |
| 58 | // CHECK-LABEL: test_mm_ceil_sd |
| 59 | // CHECK: call <2 x double> @llvm.x86.sse41.round.sd(<2 x double> %{{.*}}, <2 x double> %{{.*}}, i32 2) |
| 60 | return _mm_ceil_sd(x, y); |
| 61 | } |
| 62 | |
| 63 | __m128 test_mm_ceil_ss(__m128 x, __m128 y) { |
| 64 | // CHECK-LABEL: test_mm_ceil_ss |
| 65 | // CHECK: call <4 x float> @llvm.x86.sse41.round.ss(<4 x float> %{{.*}}, <4 x float> %{{.*}}, i32 2) |
| 66 | return _mm_ceil_ss(x, y); |
| 67 | } |
| 68 | |
| 69 | __m128i test_mm_cmpeq_epi64(__m128i A, __m128i B) { |
| 70 | // CHECK-LABEL: test_mm_cmpeq_epi64 |
| 71 | // CHECK: icmp eq <2 x i64> |
| 72 | // CHECK: sext <2 x i1> %{{.*}} to <2 x i64> |
| 73 | return _mm_cmpeq_epi64(A, B); |
| 74 | } |
| 75 | |
| 76 | __m128i test_mm_cvtepi8_epi16(__m128i a) { |
| 77 | // CHECK-LABEL: test_mm_cvtepi8_epi16 |
| 78 | // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> |
| 79 | // CHECK: sext <8 x i8> {{.*}} to <8 x i16> |
| 80 | return _mm_cvtepi8_epi16(a); |
| 81 | } |
| 82 | |
| 83 | __m128i test_mm_cvtepi8_epi32(__m128i a) { |
| 84 | // CHECK-LABEL: test_mm_cvtepi8_epi32 |
| 85 | // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| 86 | // CHECK: sext <4 x i8> {{.*}} to <4 x i32> |
| 87 | return _mm_cvtepi8_epi32(a); |
| 88 | } |
| 89 | |
| 90 | __m128i test_mm_cvtepi8_epi64(__m128i a) { |
| 91 | // CHECK-LABEL: test_mm_cvtepi8_epi64 |
| 92 | // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <2 x i32> <i32 0, i32 1> |
| 93 | // CHECK: sext <2 x i8> {{.*}} to <2 x i64> |
| 94 | return _mm_cvtepi8_epi64(a); |
| 95 | } |
| 96 | |
| 97 | __m128i test_mm_cvtepi16_epi32(__m128i a) { |
| 98 | // CHECK-LABEL: test_mm_cvtepi16_epi32 |
| 99 | // CHECK: shufflevector <8 x i16> {{.*}}, <8 x i16> {{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| 100 | // CHECK: sext <4 x i16> {{.*}} to <4 x i32> |
| 101 | return _mm_cvtepi16_epi32(a); |
| 102 | } |
| 103 | |
| 104 | __m128i test_mm_cvtepi16_epi64(__m128i a) { |
| 105 | // CHECK-LABEL: test_mm_cvtepi16_epi64 |
| 106 | // CHECK: shufflevector <8 x i16> {{.*}}, <8 x i16> {{.*}}, <2 x i32> <i32 0, i32 1> |
| 107 | // CHECK: sext <2 x i16> {{.*}} to <2 x i64> |
| 108 | return _mm_cvtepi16_epi64(a); |
| 109 | } |
| 110 | |
| 111 | __m128i test_mm_cvtepi32_epi64(__m128i a) { |
| 112 | // CHECK-LABEL: test_mm_cvtepi32_epi64 |
| 113 | // CHECK: shufflevector <4 x i32> {{.*}}, <4 x i32> {{.*}}, <2 x i32> <i32 0, i32 1> |
| 114 | // CHECK: sext <2 x i32> {{.*}} to <2 x i64> |
| 115 | return _mm_cvtepi32_epi64(a); |
| 116 | } |
| 117 | |
| 118 | __m128i test_mm_cvtepu8_epi16(__m128i a) { |
| 119 | // CHECK-LABEL: test_mm_cvtepu8_epi16 |
| 120 | // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7> |
| 121 | // CHECK: zext <8 x i8> {{.*}} to <8 x i16> |
| 122 | return _mm_cvtepu8_epi16(a); |
| 123 | } |
| 124 | |
| 125 | __m128i test_mm_cvtepu8_epi32(__m128i a) { |
| 126 | // CHECK-LABEL: test_mm_cvtepu8_epi32 |
| 127 | // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| 128 | // CHECK: zext <4 x i8> {{.*}} to <4 x i32> |
| 129 | return _mm_cvtepu8_epi32(a); |
| 130 | } |
| 131 | |
| 132 | __m128i test_mm_cvtepu8_epi64(__m128i a) { |
| 133 | // CHECK-LABEL: test_mm_cvtepu8_epi64 |
| 134 | // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <2 x i32> <i32 0, i32 1> |
| 135 | // CHECK: zext <2 x i8> {{.*}} to <2 x i64> |
| 136 | return _mm_cvtepu8_epi64(a); |
| 137 | } |
| 138 | |
| 139 | __m128i test_mm_cvtepu16_epi32(__m128i a) { |
| 140 | // CHECK-LABEL: test_mm_cvtepu16_epi32 |
| 141 | // CHECK: shufflevector <8 x i16> {{.*}}, <8 x i16> {{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3> |
| 142 | // CHECK: zext <4 x i16> {{.*}} to <4 x i32> |
| 143 | return _mm_cvtepu16_epi32(a); |
| 144 | } |
| 145 | |
| 146 | __m128i test_mm_cvtepu16_epi64(__m128i a) { |
| 147 | // CHECK-LABEL: test_mm_cvtepu16_epi64 |
| 148 | // CHECK: shufflevector <8 x i16> {{.*}}, <8 x i16> {{.*}}, <2 x i32> <i32 0, i32 1> |
| 149 | // CHECK: zext <2 x i16> {{.*}} to <2 x i64> |
| 150 | return _mm_cvtepu16_epi64(a); |
| 151 | } |
| 152 | |
| 153 | __m128i test_mm_cvtepu32_epi64(__m128i a) { |
| 154 | // CHECK-LABEL: test_mm_cvtepu32_epi64 |
| 155 | // CHECK: shufflevector <4 x i32> {{.*}}, <4 x i32> {{.*}}, <2 x i32> <i32 0, i32 1> |
| 156 | // CHECK: zext <2 x i32> {{.*}} to <2 x i64> |
| 157 | return _mm_cvtepu32_epi64(a); |
| 158 | } |
| 159 | |
| 160 | __m128d test_mm_dp_pd(__m128d x, __m128d y) { |
| 161 | // CHECK-LABEL: test_mm_dp_pd |
| 162 | // CHECK: call <2 x double> @llvm.x86.sse41.dppd(<2 x double> {{.*}}, <2 x double> {{.*}}, i8 7) |
| 163 | return _mm_dp_pd(x, y, 7); |
| 164 | } |
| 165 | |
| 166 | __m128 test_mm_dp_ps(__m128 x, __m128 y) { |
| 167 | // CHECK-LABEL: test_mm_dp_ps |
| 168 | // CHECK: call <4 x float> @llvm.x86.sse41.dpps(<4 x float> {{.*}}, <4 x float> {{.*}}, i8 7) |
| 169 | return _mm_dp_ps(x, y, 7); |
| 170 | } |
| 171 | |
| 172 | int test_mm_extract_epi8(__m128i x) { |
| 173 | // CHECK-LABEL: test_mm_extract_epi8 |
| 174 | // CHECK: extractelement <16 x i8> %{{.*}}, {{i32|i64}} 1 |
| 175 | // CHECK: zext i8 %{{.*}} to i32 |
| 176 | return _mm_extract_epi8(x, 1); |
| 177 | } |
| 178 | |
| 179 | int test_mm_extract_epi32(__m128i x) { |
| 180 | // CHECK-LABEL: test_mm_extract_epi32 |
| 181 | // CHECK: extractelement <4 x i32> %{{.*}}, {{i32|i64}} 1 |
| 182 | return _mm_extract_epi32(x, 1); |
| 183 | } |
| 184 | |
| 185 | long long test_mm_extract_epi64(__m128i x) { |
| 186 | // CHECK-LABEL: test_mm_extract_epi64 |
| 187 | // CHECK: extractelement <2 x i64> %{{.*}}, {{i32|i64}} 1 |
| 188 | return _mm_extract_epi64(x, 1); |
| 189 | } |
| 190 | |
| 191 | int test_mm_extract_ps(__m128 x) { |
| 192 | // CHECK-LABEL: test_mm_extract_ps |
| 193 | // CHECK: extractelement <4 x float> %{{.*}}, {{i32|i64}} 1 |
| 194 | return _mm_extract_ps(x, 1); |
| 195 | } |
| 196 | |
| 197 | __m128d test_mm_floor_pd(__m128d x) { |
| 198 | // CHECK-LABEL: test_mm_floor_pd |
| 199 | // CHECK: call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> %{{.*}}, i32 1) |
| 200 | return _mm_floor_pd(x); |
| 201 | } |
| 202 | |
| 203 | __m128 test_mm_floor_ps(__m128 x) { |
| 204 | // CHECK-LABEL: test_mm_floor_ps |
| 205 | // CHECK: call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> %{{.*}}, i32 1) |
| 206 | return _mm_floor_ps(x); |
| 207 | } |
| 208 | |
| 209 | __m128d test_mm_floor_sd(__m128d x, __m128d y) { |
| 210 | // CHECK-LABEL: test_mm_floor_sd |
| 211 | // CHECK: call <2 x double> @llvm.x86.sse41.round.sd(<2 x double> %{{.*}}, <2 x double> %{{.*}}, i32 1) |
| 212 | return _mm_floor_sd(x, y); |
| 213 | } |
| 214 | |
| 215 | __m128 test_mm_floor_ss(__m128 x, __m128 y) { |
| 216 | // CHECK-LABEL: test_mm_floor_ss |
| 217 | // CHECK: call <4 x float> @llvm.x86.sse41.round.ss(<4 x float> %{{.*}}, <4 x float> %{{.*}}, i32 1) |
| 218 | return _mm_floor_ss(x, y); |
| 219 | } |
| 220 | |
| 221 | __m128i test_mm_insert_epi8(__m128i x, char b) { |
| 222 | // CHECK-LABEL: test_mm_insert_epi8 |
| 223 | // CHECK: insertelement <16 x i8> %{{.*}}, i8 %{{.*}}, {{i32|i64}} 1 |
| 224 | return _mm_insert_epi8(x, b, 1); |
| 225 | } |
| 226 | |
| 227 | __m128i test_mm_insert_epi32(__m128i x, int b) { |
| 228 | // CHECK-LABEL: test_mm_insert_epi32 |
| 229 | // CHECK: insertelement <4 x i32> %{{.*}}, i32 %{{.*}}, {{i32|i64}} 1 |
| 230 | return _mm_insert_epi32(x, b, 1); |
| 231 | } |
| 232 | |
| 233 | __m128i test_mm_insert_epi64(__m128i x, long long b) { |
| 234 | // CHECK-LABEL: test_mm_insert_epi64 |
| 235 | // CHECK: insertelement <2 x i64> %{{.*}}, i64 %{{.*}}, {{i32|i64}} 1 |
| 236 | return _mm_insert_epi64(x, b, 1); |
| 237 | } |
| 238 | |
| 239 | __m128 test_mm_insert_ps(__m128 x, __m128 y) { |
| 240 | // CHECK-LABEL: test_mm_insert_ps |
| 241 | // CHECK: call <4 x float> @llvm.x86.sse41.insertps(<4 x float> %{{.*}}, <4 x float> %{{.*}}, i8 4) |
| 242 | return _mm_insert_ps(x, y, 4); |
| 243 | } |
| 244 | |
| 245 | __m128i test_mm_max_epi8(__m128i x, __m128i y) { |
| 246 | // CHECK-LABEL: test_mm_max_epi8 |
| 247 | // CHECK: [[CMP:%.*]] = icmp sgt <16 x i8> [[X:%.*]], [[Y:%.*]] |
| 248 | // CHECK-NEXT: select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]] |
| 249 | return _mm_max_epi8(x, y); |
| 250 | } |
| 251 | |
| 252 | __m128i test_mm_max_epi32(__m128i x, __m128i y) { |
| 253 | // CHECK-LABEL: test_mm_max_epi32 |
| 254 | // CHECK: [[CMP:%.*]] = icmp sgt <4 x i32> [[X:%.*]], [[Y:%.*]] |
| 255 | // CHECK-NEXT: select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]] |
| 256 | return _mm_max_epi32(x, y); |
| 257 | } |
| 258 | |
| 259 | __m128i test_mm_max_epu16(__m128i x, __m128i y) { |
| 260 | // CHECK-LABEL: test_mm_max_epu16 |
| 261 | // CHECK: [[CMP:%.*]] = icmp ugt <8 x i16> [[X:%.*]], [[Y:%.*]] |
| 262 | // CHECK-NEXT: select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]] |
| 263 | return _mm_max_epu16(x, y); |
| 264 | } |
| 265 | |
| 266 | __m128i test_mm_max_epu32(__m128i x, __m128i y) { |
| 267 | // CHECK-LABEL: test_mm_max_epu32 |
| 268 | // CHECK: [[CMP:%.*]] = icmp ugt <4 x i32> [[X:%.*]], [[Y:%.*]] |
| 269 | // CHECK-NEXT: select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]] |
| 270 | return _mm_max_epu32(x, y); |
| 271 | } |
| 272 | |
| 273 | __m128i test_mm_min_epi8(__m128i x, __m128i y) { |
| 274 | // CHECK-LABEL: test_mm_min_epi8 |
| 275 | // CHECK: [[CMP:%.*]] = icmp slt <16 x i8> [[X:%.*]], [[Y:%.*]] |
| 276 | // CHECK-NEXT: select <16 x i1> [[CMP]], <16 x i8> [[X]], <16 x i8> [[Y]] |
| 277 | return _mm_min_epi8(x, y); |
| 278 | } |
| 279 | |
| 280 | __m128i test_mm_min_epi32(__m128i x, __m128i y) { |
| 281 | // CHECK-LABEL: test_mm_min_epi32 |
| 282 | // CHECK: [[CMP:%.*]] = icmp slt <4 x i32> [[X:%.*]], [[Y:%.*]] |
| 283 | // CHECK-NEXT: select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]] |
| 284 | return _mm_min_epi32(x, y); |
| 285 | } |
| 286 | |
| 287 | __m128i test_mm_min_epu16(__m128i x, __m128i y) { |
| 288 | // CHECK-LABEL: test_mm_min_epu16 |
| 289 | // CHECK: [[CMP:%.*]] = icmp ult <8 x i16> [[X:%.*]], [[Y:%.*]] |
| 290 | // CHECK-NEXT: select <8 x i1> [[CMP]], <8 x i16> [[X]], <8 x i16> [[Y]] |
| 291 | return _mm_min_epu16(x, y); |
| 292 | } |
| 293 | |
| 294 | __m128i test_mm_min_epu32(__m128i x, __m128i y) { |
| 295 | // CHECK-LABEL: test_mm_min_epu32 |
| 296 | // CHECK: [[CMP:%.*]] = icmp ult <4 x i32> [[X:%.*]], [[Y:%.*]] |
| 297 | // CHECK-NEXT: select <4 x i1> [[CMP]], <4 x i32> [[X]], <4 x i32> [[Y]] |
| 298 | return _mm_min_epu32(x, y); |
| 299 | } |
| 300 | |
| 301 | __m128i test_mm_minpos_epu16(__m128i x) { |
| 302 | // CHECK-LABEL: test_mm_minpos_epu16 |
| 303 | // CHECK: call <8 x i16> @llvm.x86.sse41.phminposuw(<8 x i16> %{{.*}}) |
| 304 | return _mm_minpos_epu16(x); |
| 305 | } |
| 306 | |
| 307 | __m128i test_mm_mpsadbw_epu8(__m128i x, __m128i y) { |
| 308 | // CHECK-LABEL: test_mm_mpsadbw_epu8 |
| 309 | // CHECK: call <8 x i16> @llvm.x86.sse41.mpsadbw(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, i8 1) |
| 310 | return _mm_mpsadbw_epu8(x, y, 1); |
| 311 | } |
| 312 | |
| 313 | __m128i test_mm_mul_epi32(__m128i x, __m128i y) { |
| 314 | // CHECK-LABEL: test_mm_mul_epi32 |
| 315 | // CHECK: shl <2 x i64> %{{.*}}, <i64 32, i64 32> |
| 316 | // CHECK: ashr <2 x i64> %{{.*}}, <i64 32, i64 32> |
| 317 | // CHECK: shl <2 x i64> %{{.*}}, <i64 32, i64 32> |
| 318 | // CHECK: ashr <2 x i64> %{{.*}}, <i64 32, i64 32> |
| 319 | // CHECK: mul <2 x i64> %{{.*}}, %{{.*}} |
| 320 | return _mm_mul_epi32(x, y); |
| 321 | } |
| 322 | |
| 323 | __m128i test_mm_mullo_epi32(__m128i x, __m128i y) { |
| 324 | // CHECK-LABEL: test_mm_mullo_epi32 |
| 325 | // CHECK: mul <4 x i32> |
| 326 | return _mm_mullo_epi32(x, y); |
| 327 | } |
| 328 | |
| 329 | __m128i test_mm_packus_epi32(__m128i x, __m128i y) { |
| 330 | // CHECK-LABEL: test_mm_packus_epi32 |
| 331 | // CHECK: call <8 x i16> @llvm.x86.sse41.packusdw(<4 x i32> %{{.*}}, <4 x i32> %{{.*}}) |
| 332 | return _mm_packus_epi32(x, y); |
| 333 | } |
| 334 | |
| 335 | __m128d test_mm_round_pd(__m128d x) { |
| 336 | // CHECK-LABEL: test_mm_round_pd |
| 337 | // CHECK: call <2 x double> @llvm.x86.sse41.round.pd(<2 x double> %{{.*}}, i32 4) |
| 338 | return _mm_round_pd(x, 4); |
| 339 | } |
| 340 | |
| 341 | __m128 test_mm_round_ps(__m128 x) { |
| 342 | // CHECK-LABEL: test_mm_round_ps |
| 343 | // CHECK: call <4 x float> @llvm.x86.sse41.round.ps(<4 x float> %{{.*}}, i32 4) |
| 344 | return _mm_round_ps(x, 4); |
| 345 | } |
| 346 | |
| 347 | __m128d test_mm_round_sd(__m128d x, __m128d y) { |
| 348 | // CHECK-LABEL: test_mm_round_sd |
| 349 | // CHECK: call <2 x double> @llvm.x86.sse41.round.sd(<2 x double> %{{.*}}, <2 x double> %{{.*}}, i32 4) |
| 350 | return _mm_round_sd(x, y, 4); |
| 351 | } |
| 352 | |
| 353 | __m128 test_mm_round_ss(__m128 x, __m128 y) { |
| 354 | // CHECK-LABEL: test_mm_round_ss |
| 355 | // CHECK: call <4 x float> @llvm.x86.sse41.round.ss(<4 x float> %{{.*}}, <4 x float> %{{.*}}, i32 4) |
| 356 | return _mm_round_ss(x, y, 4); |
| 357 | } |
| 358 | |
| 359 | __m128i test_mm_stream_load_si128(__m128i const *a) { |
| 360 | // CHECK-LABEL: test_mm_stream_load_si128 |
| 361 | // CHECK: load <2 x i64>, <2 x i64>* %{{.*}}, align 16, !nontemporal |
| 362 | return _mm_stream_load_si128(a); |
| 363 | } |
| 364 | |
| 365 | int test_mm_test_all_ones(__m128i x) { |
| 366 | // CHECK-LABEL: test_mm_test_all_ones |
| 367 | // CHECK: call i32 @llvm.x86.sse41.ptestc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) |
| 368 | return _mm_test_all_ones(x); |
| 369 | } |
| 370 | |
| 371 | int test_mm_test_all_zeros(__m128i x, __m128i y) { |
| 372 | // CHECK-LABEL: test_mm_test_all_zeros |
| 373 | // CHECK: call i32 @llvm.x86.sse41.ptestz(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) |
| 374 | return _mm_test_all_zeros(x, y); |
| 375 | } |
| 376 | |
| 377 | int test_mm_test_mix_ones_zeros(__m128i x, __m128i y) { |
| 378 | // CHECK-LABEL: test_mm_test_mix_ones_zeros |
| 379 | // CHECK: call i32 @llvm.x86.sse41.ptestnzc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) |
| 380 | return _mm_test_mix_ones_zeros(x, y); |
| 381 | } |
| 382 | |
| 383 | int test_mm_testc_si128(__m128i x, __m128i y) { |
| 384 | // CHECK-LABEL: test_mm_testc_si128 |
| 385 | // CHECK: call i32 @llvm.x86.sse41.ptestc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) |
| 386 | return _mm_testc_si128(x, y); |
| 387 | } |
| 388 | |
| 389 | int test_mm_testnzc_si128(__m128i x, __m128i y) { |
| 390 | // CHECK-LABEL: test_mm_testnzc_si128 |
| 391 | // CHECK: call i32 @llvm.x86.sse41.ptestnzc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) |
| 392 | return _mm_testnzc_si128(x, y); |
| 393 | } |
| 394 | |
| 395 | int test_mm_testz_si128(__m128i x, __m128i y) { |
| 396 | // CHECK-LABEL: test_mm_testz_si128 |
| 397 | // CHECK: call i32 @llvm.x86.sse41.ptestz(<2 x i64> %{{.*}}, <2 x i64> %{{.*}}) |
| 398 | return _mm_testz_si128(x, y); |
| 399 | } |
| 400 | |