clang/test/CodeGen/X86/sse41-builtins.c

   1 // RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse4.1 -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,X64
   2 // RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse4.1 -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,X64
   3 // RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse4.1 -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK
   4 // RUN: %clang_cc1 -x c -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse4.1 -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK
   5 // RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse4.1 -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,X64
   6 // RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse4.1 -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK,X64
   7 // RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse4.1 -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK
   8 // RUN: %clang_cc1 -x c++ -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse4.1 -fno-signed-char -emit-llvm -o - -Wall -Werror | FileCheck %s --check-prefixes=CHECK
   9
  10
  11 #include <immintrin.h>
  12
  13 // NOTE: This should match the tests in llvm/test/CodeGen/X86/sse41-intrinsics-fast-isel.ll
  14
  15 __m128i test_mm_blend_epi16(__m128i V1, __m128i V2) {
  16   // CHECK-LABEL: test_mm_blend_epi16
  17   // 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>
  18   return _mm_blend_epi16(V1, V2, 42);
  19 }
  20
  21 __m128d test_mm_blend_pd(__m128d V1, __m128d V2) {
  22   // CHECK-LABEL: test_mm_blend_pd
  23   // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x i32> <i32 0, i32 3>
  24   return _mm_blend_pd(V1, V2, 2);
  25 }
  26
  27 __m128 test_mm_blend_ps(__m128 V1, __m128 V2) {
  28   // CHECK-LABEL: test_mm_blend_ps
  29   // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 5, i32 6, i32 3>
  30   return _mm_blend_ps(V1, V2, 6);
  31 }
  32
  33 __m128i test_mm_blendv_epi8(__m128i V1, __m128i V2, __m128i V3) {
  34   // CHECK-LABEL: test_mm_blendv_epi8
  35   // CHECK: call <16 x i8> @llvm.x86.sse41.pblendvb(<16 x i8> %{{.*}}, <16 x i8> %{{.*}}, <16 x i8> %{{.*}})
  36   return _mm_blendv_epi8(V1, V2, V3);
  37 }
  38
  39 __m128d test_mm_blendv_pd(__m128d V1, __m128d V2, __m128d V3) {
  40   // CHECK-LABEL: test_mm_blendv_pd
  41   // CHECK: call {{.*}}<2 x double> @llvm.x86.sse41.blendvpd(<2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x double> %{{.*}})
  42   return _mm_blendv_pd(V1, V2, V3);
  43 }
  44
  45 __m128 test_mm_blendv_ps(__m128 V1, __m128 V2, __m128 V3) {
  46   // CHECK-LABEL: test_mm_blendv_ps
  47   // CHECK: call {{.*}}<4 x float> @llvm.x86.sse41.blendvps(<4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x float> %{{.*}})
  48   return _mm_blendv_ps(V1, V2, V3);
  49 }
  50
  51 __m128d test_mm_ceil_pd(__m128d x) {
  52   // CHECK-LABEL: test_mm_ceil_pd
  53   // CHECK: call {{.*}}<2 x double> @llvm.x86.sse41.round.pd(<2 x double> %{{.*}}, i32 2)
  54   return _mm_ceil_pd(x);
  55 }
  56
  57 __m128 test_mm_ceil_ps(__m128 x) {
  58   // CHECK-LABEL: test_mm_ceil_ps
  59   // CHECK: call {{.*}}<4 x float> @llvm.x86.sse41.round.ps(<4 x float> %{{.*}}, i32 2)
  60   return _mm_ceil_ps(x);
  61 }
  62
  63 __m128d test_mm_ceil_sd(__m128d x, __m128d y) {
  64   // CHECK-LABEL: test_mm_ceil_sd
  65   // CHECK: call {{.*}}<2 x double> @llvm.x86.sse41.round.sd(<2 x double> %{{.*}}, <2 x double> %{{.*}}, i32 2)
  66   return _mm_ceil_sd(x, y);
  67 }
  68
  69 __m128 test_mm_ceil_ss(__m128 x, __m128 y) {
  70   // CHECK-LABEL: test_mm_ceil_ss
  71   // CHECK: call {{.*}}<4 x float> @llvm.x86.sse41.round.ss(<4 x float> %{{.*}}, <4 x float> %{{.*}}, i32 2)
  72   return _mm_ceil_ss(x, y);
  73 }
  74
  75 __m128i test_mm_cmpeq_epi64(__m128i A, __m128i B) {
  76   // CHECK-LABEL: test_mm_cmpeq_epi64
  77   // CHECK: icmp eq <2 x i64>
  78   // CHECK: sext <2 x i1> %{{.*}} to <2 x i64>
  79   return _mm_cmpeq_epi64(A, B);
  80 }
  81
  82 __m128i test_mm_cvtepi8_epi16(__m128i a) {
  83   // CHECK-LABEL: test_mm_cvtepi8_epi16
  84   // 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>
  85   // CHECK: sext <8 x i8> {{.*}} to <8 x i16>
  86   return _mm_cvtepi8_epi16(a);
  87 }
  88
  89 __m128i test_mm_cvtepi8_epi32(__m128i a) {
  90   // CHECK-LABEL: test_mm_cvtepi8_epi32
  91   // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
  92   // CHECK: sext <4 x i8> {{.*}} to <4 x i32>
  93   return _mm_cvtepi8_epi32(a);
  94 }
  95
  96 __m128i test_mm_cvtepi8_epi64(__m128i a) {
  97   // CHECK-LABEL: test_mm_cvtepi8_epi64
  98   // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <2 x i32> <i32 0, i32 1>
  99   // CHECK: sext <2 x i8> {{.*}} to <2 x i64>
 100   return _mm_cvtepi8_epi64(a);
 101 }
 102
 103 __m128i test_mm_cvtepi16_epi32(__m128i a) {
 104   // CHECK-LABEL: test_mm_cvtepi16_epi32
 105   // CHECK: shufflevector <8 x i16> {{.*}}, <8 x i16> {{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 106   // CHECK: sext <4 x i16> {{.*}} to <4 x i32>
 107   return _mm_cvtepi16_epi32(a);
 108 }
 109
 110 __m128i test_mm_cvtepi16_epi64(__m128i a) {
 111   // CHECK-LABEL: test_mm_cvtepi16_epi64
 112   // CHECK: shufflevector <8 x i16> {{.*}}, <8 x i16> {{.*}}, <2 x i32> <i32 0, i32 1>
 113   // CHECK: sext <2 x i16> {{.*}} to <2 x i64>
 114   return _mm_cvtepi16_epi64(a);
 115 }
 116
 117 __m128i test_mm_cvtepi32_epi64(__m128i a) {
 118   // CHECK-LABEL: test_mm_cvtepi32_epi64
 119   // CHECK: shufflevector <4 x i32> {{.*}}, <4 x i32> {{.*}}, <2 x i32> <i32 0, i32 1>
 120   // CHECK: sext <2 x i32> {{.*}} to <2 x i64>
 121   return _mm_cvtepi32_epi64(a);
 122 }
 123
 124 __m128i test_mm_cvtepu8_epi16(__m128i a) {
 125   // CHECK-LABEL: test_mm_cvtepu8_epi16
 126   // 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>
 127   // CHECK: zext <8 x i8> {{.*}} to <8 x i16>
 128   return _mm_cvtepu8_epi16(a);
 129 }
 130
 131 __m128i test_mm_cvtepu8_epi32(__m128i a) {
 132   // CHECK-LABEL: test_mm_cvtepu8_epi32
 133   // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 134   // CHECK: zext <4 x i8> {{.*}} to <4 x i32>
 135   return _mm_cvtepu8_epi32(a);
 136 }
 137
 138 __m128i test_mm_cvtepu8_epi64(__m128i a) {
 139   // CHECK-LABEL: test_mm_cvtepu8_epi64
 140   // CHECK: shufflevector <16 x i8> {{.*}}, <16 x i8> {{.*}}, <2 x i32> <i32 0, i32 1>
 141   // CHECK: zext <2 x i8> {{.*}} to <2 x i64>
 142   return _mm_cvtepu8_epi64(a);
 143 }
 144
 145 __m128i test_mm_cvtepu16_epi32(__m128i a) {
 146   // CHECK-LABEL: test_mm_cvtepu16_epi32
 147   // CHECK: shufflevector <8 x i16> {{.*}}, <8 x i16> {{.*}}, <4 x i32> <i32 0, i32 1, i32 2, i32 3>
 148   // CHECK: zext <4 x i16> {{.*}} to <4 x i32>
 149   return _mm_cvtepu16_epi32(a);
 150 }
 151
 152 __m128i test_mm_cvtepu16_epi64(__m128i a) {
 153   // CHECK-LABEL: test_mm_cvtepu16_epi64
 154   // CHECK: shufflevector <8 x i16> {{.*}}, <8 x i16> {{.*}}, <2 x i32> <i32 0, i32 1>
 155   // CHECK: zext <2 x i16> {{.*}} to <2 x i64>
 156   return _mm_cvtepu16_epi64(a);
 157 }
 158
 159 __m128i test_mm_cvtepu32_epi64(__m128i a) {
 160   // CHECK-LABEL: test_mm_cvtepu32_epi64
 161   // CHECK: shufflevector <4 x i32> {{.*}}, <4 x i32> {{.*}}, <2 x i32> <i32 0, i32 1>
 162   // CHECK: zext <2 x i32> {{.*}} to <2 x i64>
 163   return _mm_cvtepu32_epi64(a);
 164 }
 165
 166 __m128d test_mm_dp_pd(__m128d x, __m128d y) {
 167   // CHECK-LABEL: test_mm_dp_pd
 168   // CHECK: call {{.*}}<2 x double> @llvm.x86.sse41.dppd(<2 x double> {{.*}}, <2 x double> {{.*}}, i8 7)
 169   return _mm_dp_pd(x, y, 7);
 170 }
 171
 172 __m128 test_mm_dp_ps(__m128 x, __m128 y) {
 173   // CHECK-LABEL: test_mm_dp_ps
 174   // CHECK: call {{.*}}<4 x float> @llvm.x86.sse41.dpps(<4 x float> {{.*}}, <4 x float> {{.*}}, i8 7)
 175   return _mm_dp_ps(x, y, 7);
 176 }
 177
 178 int test_mm_extract_epi8(__m128i x) {
 179   // CHECK-LABEL: test_mm_extract_epi8
 180   // CHECK: extractelement <16 x i8> %{{.*}}, {{i32|i64}} 1
 181   // CHECK: zext i8 %{{.*}} to i32
 182   return _mm_extract_epi8(x, 1);
 183 }
 184
 185 int test_mm_extract_epi32(__m128i x) {
 186   // CHECK-LABEL: test_mm_extract_epi32
 187   // CHECK: extractelement <4 x i32> %{{.*}}, {{i32|i64}} 1
 188   return _mm_extract_epi32(x, 1);
 189 }
 190
 191 long long test_mm_extract_epi64(__m128i x) {
 192   // CHECK-LABEL: test_mm_extract_epi64
 193   // CHECK: extractelement <2 x i64> %{{.*}}, {{i32|i64}} 1
 194   return _mm_extract_epi64(x, 1);
 195 }
 196
 197 int test_mm_extract_ps(__m128 x) {
 198   // CHECK-LABEL: test_mm_extract_ps
 199   // CHECK: extractelement <4 x float> %{{.*}}, {{i32|i64}} 1
 200   return _mm_extract_ps(x, 1);
 201 }
 202
 203 __m128d test_mm_floor_pd(__m128d x) {
 204   // CHECK-LABEL: test_mm_floor_pd
 205   // CHECK: call {{.*}}<2 x double> @llvm.x86.sse41.round.pd(<2 x double> %{{.*}}, i32 1)
 206   return _mm_floor_pd(x);
 207 }
 208
 209 __m128 test_mm_floor_ps(__m128 x) {
 210   // CHECK-LABEL: test_mm_floor_ps
 211   // CHECK: call {{.*}}<4 x float> @llvm.x86.sse41.round.ps(<4 x float> %{{.*}}, i32 1)
 212   return _mm_floor_ps(x);
 213 }
 214
 215 __m128d test_mm_floor_sd(__m128d x, __m128d y) {
 216   // CHECK-LABEL: test_mm_floor_sd
 217   // CHECK: call {{.*}}<2 x double> @llvm.x86.sse41.round.sd(<2 x double> %{{.*}}, <2 x double> %{{.*}}, i32 1)
 218   return _mm_floor_sd(x, y);
 219 }
 220
 221 __m128 test_mm_floor_ss(__m128 x, __m128 y) {
 222   // CHECK-LABEL: test_mm_floor_ss
 223   // CHECK: call {{.*}}<4 x float> @llvm.x86.sse41.round.ss(<4 x float> %{{.*}}, <4 x float> %{{.*}}, i32 1)
 224   return _mm_floor_ss(x, y);
 225 }
 226
 227 __m128i test_mm_insert_epi8(__m128i x, char b) {
 228   // CHECK-LABEL: test_mm_insert_epi8
 229   // CHECK: insertelement <16 x i8> %{{.*}}, i8 %{{.*}}, {{i32|i64}} 1
 230   return _mm_insert_epi8(x, b, 1);
 231 }
 232
 233 __m128i test_mm_insert_epi32(__m128i x, int b) {
 234   // CHECK-LABEL: test_mm_insert_epi32
 235   // CHECK: insertelement <4 x i32> %{{.*}}, i32 %{{.*}}, {{i32|i64}} 1
 236   return _mm_insert_epi32(x, b, 1);
 237 }
 238
 239 #ifdef __x86_64__
 240 __m128i test_mm_insert_epi64(__m128i x, long long b) {
 241   // X64-LABEL: test_mm_insert_epi64
 242   // X64: insertelement <2 x i64> %{{.*}}, i64 %{{.*}}, {{i32|i64}} 1
 243   return _mm_insert_epi64(x, b, 1);
 244 }
 245 #endif
 246
 247 __m128 test_mm_insert_ps(__m128 x, __m128 y) {
 248   // CHECK-LABEL: test_mm_insert_ps
 249   // CHECK: call {{.*}}<4 x float> @llvm.x86.sse41.insertps(<4 x float> %{{.*}}, <4 x float> %{{.*}}, i8 4)
 250   return _mm_insert_ps(x, y, 4);
 251 }
 252
 253 __m128i test_mm_max_epi8(__m128i x, __m128i y) {
 254   // CHECK-LABEL: test_mm_max_epi8
 255   // CHECK: call <16 x i8> @llvm.smax.v16i8(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
 256   return _mm_max_epi8(x, y);
 257 }
 258
 259 __m128i test_mm_max_epi32(__m128i x, __m128i y) {
 260   // CHECK-LABEL: test_mm_max_epi32
 261   // CHECK: call <4 x i32> @llvm.smax.v4i32(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
 262   return _mm_max_epi32(x, y);
 263 }
 264
 265 __m128i test_mm_max_epu16(__m128i x, __m128i y) {
 266   // CHECK-LABEL: test_mm_max_epu16
 267   // CHECK: call <8 x i16> @llvm.umax.v8i16(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
 268   return _mm_max_epu16(x, y);
 269 }
 270
 271 __m128i test_mm_max_epu32(__m128i x, __m128i y) {
 272   // CHECK-LABEL: test_mm_max_epu32
 273   // CHECK: call <4 x i32> @llvm.umax.v4i32(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
 274   return _mm_max_epu32(x, y);
 275 }
 276
 277 __m128i test_mm_min_epi8(__m128i x, __m128i y) {
 278   // CHECK-LABEL: test_mm_min_epi8
 279   // CHECK: call <16 x i8> @llvm.smin.v16i8(<16 x i8> %{{.*}}, <16 x i8> %{{.*}})
 280   return _mm_min_epi8(x, y);
 281 }
 282
 283 __m128i test_mm_min_epi32(__m128i x, __m128i y) {
 284   // CHECK-LABEL: test_mm_min_epi32
 285   // CHECK: call <4 x i32> @llvm.smin.v4i32(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
 286   return _mm_min_epi32(x, y);
 287 }
 288
 289 __m128i test_mm_min_epu16(__m128i x, __m128i y) {
 290   // CHECK-LABEL: test_mm_min_epu16
 291   // CHECK: call <8 x i16> @llvm.umin.v8i16(<8 x i16> %{{.*}}, <8 x i16> %{{.*}})
 292   return _mm_min_epu16(x, y);
 293 }
 294
 295 __m128i test_mm_min_epu32(__m128i x, __m128i y) {
 296   // CHECK-LABEL: test_mm_min_epu32
 297   // CHECK: call <4 x i32> @llvm.umin.v4i32(<4 x i32> %{{.*}}, <4 x i32> %{{.*}})
 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> %{{.*}}, splat (i64 32)
 316   // CHECK: ashr <2 x i64> %{{.*}}, splat (i64 32)
 317   // CHECK: shl <2 x i64> %{{.*}}, splat (i64 32)
 318   // CHECK: ashr <2 x i64> %{{.*}}, splat (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>, ptr %{{.*}}, align 16, !nontemporal
 362   return _mm_stream_load_si128(a);
 363 }
 364
 365 __m128i test_mm_stream_load_si128_void(const void *a) {
 366   // CHECK-LABEL: test_mm_stream_load_si128_void
 367   // CHECK: load <2 x i64>, ptr %{{.*}}, align 16, !nontemporal
 368   return _mm_stream_load_si128(a);
 369 }
 370
 371 int test_mm_test_all_ones(__m128i x) {
 372   // CHECK-LABEL: test_mm_test_all_ones
 373   // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
 374   return _mm_test_all_ones(x);
 375 }
 376
 377 int test_mm_test_all_zeros(__m128i x, __m128i y) {
 378   // CHECK-LABEL: test_mm_test_all_zeros
 379   // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestz(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
 380   return _mm_test_all_zeros(x, y);
 381 }
 382
 383 int test_mm_test_mix_ones_zeros(__m128i x, __m128i y) {
 384   // CHECK-LABEL: test_mm_test_mix_ones_zeros
 385   // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestnzc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
 386   return _mm_test_mix_ones_zeros(x, y);
 387 }
 388
 389 int test_mm_testc_si128(__m128i x, __m128i y) {
 390   // CHECK-LABEL: test_mm_testc_si128
 391   // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
 392   return _mm_testc_si128(x, y);
 393 }
 394
 395 int test_mm_testnzc_si128(__m128i x, __m128i y) {
 396   // CHECK-LABEL: test_mm_testnzc_si128
 397   // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestnzc(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
 398   return _mm_testnzc_si128(x, y);
 399 }
 400
 401 int test_mm_testz_si128(__m128i x, __m128i y) {
 402   // CHECK-LABEL: test_mm_testz_si128
 403   // CHECK: call {{.*}}i32 @llvm.x86.sse41.ptestz(<2 x i64> %{{.*}}, <2 x i64> %{{.*}})
 404   return _mm_testz_si128(x, y);
 405 }
 406