clang/test/CodeGen/X86/sse4a-builtins.c

   1 // RUN: %clang_cc1 -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse4a -emit-llvm -o - -Wall -Werror | FileCheck %s
   2 // RUN: %clang_cc1 -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse4a -emit-llvm -o - -Wall -Werror | FileCheck %s
   3
   4
   5 #include <x86intrin.h>
   6
   7 // NOTE: This should match the tests in llvm/test/CodeGen/X86/sse4a-intrinsics-fast-isel.ll
   8
   9 __m128i test_mm_extracti_si64(__m128i x) {
  10   // CHECK-LABEL: test_mm_extracti_si64
  11   // CHECK: call <2 x i64> @llvm.x86.sse4a.extrqi(<2 x i64> %{{[^,]+}}, i8 3, i8 2)
  12   return _mm_extracti_si64(x, 3, 2);
  13 }
  14
  15 __m128i test_mm_extract_si64(__m128i x, __m128i y) {
  16   // CHECK-LABEL: test_mm_extract_si64
  17   // CHECK: call <2 x i64> @llvm.x86.sse4a.extrq(<2 x i64> %{{[^,]+}}, <16 x i8> %{{[^,]+}})
  18   return _mm_extract_si64(x, y);
  19 }
  20
  21 __m128i test_mm_inserti_si64(__m128i x, __m128i y) {
  22   // CHECK-LABEL: test_mm_inserti_si64
  23   // CHECK: call <2 x i64> @llvm.x86.sse4a.insertqi(<2 x i64> %{{[^,]+}}, <2 x i64> %{{[^,]+}}, i8 5, i8 6)
  24   return _mm_inserti_si64(x, y, 5, 6);
  25 }
  26
  27 __m128i test_mm_insert_si64(__m128i x, __m128i y) {
  28   // CHECK-LABEL: test_mm_insert_si64
  29   // CHECK: call <2 x i64> @llvm.x86.sse4a.insertq(<2 x i64> %{{[^,]+}}, <2 x i64> %{{[^,]+}})
  30   return _mm_insert_si64(x, y);
  31 }
  32
  33 void test_mm_stream_sd(double *p, __m128d a) {
  34   // CHECK-LABEL: test_mm_stream_sd
  35   // CHECK: extractelement <2 x double> %{{.*}}, i64 0
  36   // CHECK: store double %{{.*}}, ptr %{{.*}}, align 1, !nontemporal
  37    _mm_stream_sd(p, a);
  38 }
  39
  40 void test_mm_stream_sd_void(void *p, __m128d a) {
  41   // CHECK-LABEL: test_mm_stream_sd_void
  42   // CHECK: extractelement <2 x double> %{{.*}}, i64 0
  43   // CHECK: store double %{{.*}}, ptr %{{.*}}, align 1, !nontemporal
  44    _mm_stream_sd(p, a);
  45 }
  46
  47 void test_mm_stream_ss(float *p, __m128 a) {
  48   // CHECK-LABEL: test_mm_stream_ss
  49   // CHECK: extractelement <4 x float> %{{.*}}, i64 0
  50   // CHECK: store float %{{.*}}, ptr %{{.*}}, align 1, !nontemporal
  51   _mm_stream_ss(p, a);
  52 }
  53
  54 void test_mm_stream_s_void(void *p, __m128 a) {
  55   // CHECK-LABEL: test_mm_stream_s_void
  56   // CHECK: extractelement <4 x float> %{{.*}}, i64 0
  57   // CHECK: store float %{{.*}}, ptr %{{.*}}, align 1, !nontemporal
  58   _mm_stream_ss(p, a);
  59 }