clang/test/CodeGen/X86/sse3-builtins.c

   1 // RUN: %clang_cc1 -flax-vector-conversions=none -ffreestanding %s -triple=x86_64-apple-darwin -target-feature +sse3 -emit-llvm -o - -Wall -Werror | FileCheck %s
   2 // RUN: %clang_cc1 -flax-vector-conversions=none -ffreestanding %s -triple=i386-apple-darwin -target-feature +sse3 -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/sse3-intrinsics-fast-isel.ll
   8
   9 __m128d test_mm_addsub_pd(__m128d A, __m128d B) {
  10   // CHECK-LABEL: test_mm_addsub_pd
  11   // CHECK: call <2 x double> @llvm.x86.sse3.addsub.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}})
  12   return _mm_addsub_pd(A, B);
  13 }
  14
  15 __m128 test_mm_addsub_ps(__m128 A, __m128 B) {
  16   // CHECK-LABEL: test_mm_addsub_ps
  17   // CHECK: call <4 x float> @llvm.x86.sse3.addsub.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}})
  18   return _mm_addsub_ps(A, B);
  19 }
  20
  21 __m128d test_mm_hadd_pd(__m128d A, __m128d B) {
  22   // CHECK-LABEL: test_mm_hadd_pd
  23   // CHECK: call <2 x double> @llvm.x86.sse3.hadd.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}})
  24   return _mm_hadd_pd(A, B);
  25 }
  26
  27 __m128 test_mm_hadd_ps(__m128 A, __m128 B) {
  28   // CHECK-LABEL: test_mm_hadd_ps
  29   // CHECK: call <4 x float> @llvm.x86.sse3.hadd.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}})
  30   return _mm_hadd_ps(A, B);
  31 }
  32
  33 __m128d test_mm_hsub_pd(__m128d A, __m128d B) {
  34   // CHECK-LABEL: test_mm_hsub_pd
  35   // CHECK: call <2 x double> @llvm.x86.sse3.hsub.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}})
  36   return _mm_hsub_pd(A, B);
  37 }
  38
  39 __m128 test_mm_hsub_ps(__m128 A, __m128 B) {
  40   // CHECK-LABEL: test_mm_hsub_ps
  41   // CHECK: call <4 x float> @llvm.x86.sse3.hsub.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}})
  42   return _mm_hsub_ps(A, B);
  43 }
  44
  45 __m128i test_mm_lddqu_si128(__m128i const* P) {
  46   // CHECK-LABEL: test_mm_lddqu_si128
  47   // CHECK: call <16 x i8> @llvm.x86.sse3.ldu.dq(ptr %{{.*}})
  48   return _mm_lddqu_si128(P);
  49 }
  50
  51 __m128d test_mm_loaddup_pd(double const* P) {
  52   // CHECK-LABEL: test_mm_loaddup_pd
  53   // CHECK: load ptr
  54   // CHECK: insertelement <2 x double> undef, double %{{.*}}, i32 0
  55   // CHECK: insertelement <2 x double> %{{.*}}, double %{{.*}}, i32 1
  56   return _mm_loaddup_pd(P);
  57 }
  58
  59 __m128d test_mm_movedup_pd(__m128d A) {
  60   // CHECK-LABEL: test_mm_movedup_pd
  61   // CHECK: shufflevector <2 x double> %{{.*}}, <2 x double> %{{.*}}, <2 x i32> zeroinitializer
  62   return _mm_movedup_pd(A);
  63 }
  64
  65 __m128 test_mm_movehdup_ps(__m128 A) {
  66   // CHECK-LABEL: test_mm_movehdup_ps
  67   // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 1, i32 1, i32 3, i32 3>
  68   return _mm_movehdup_ps(A);
  69 }
  70
  71 __m128 test_mm_moveldup_ps(__m128 A) {
  72   // CHECK-LABEL: test_mm_moveldup_ps
  73   // CHECK: shufflevector <4 x float> %{{.*}}, <4 x float> %{{.*}}, <4 x i32> <i32 0, i32 0, i32 2, i32 2>
  74   return _mm_moveldup_ps(A);
  75 }