#include <x86intrin.h>
+// NOTE: This should match the tests in llvm/test/CodeGen/X86/sse3-intrinsics-fast-isel.ll
+
__m128d test_mm_addsub_pd(__m128d A, __m128d B) {
// CHECK-LABEL: test_mm_addsub_pd
- // CHECK: call <2 x double> @llvm.x86.sse3.addsub.pd
+ // CHECK: call <2 x double> @llvm.x86.sse3.addsub.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}})
return _mm_addsub_pd(A, B);
}
__m128 test_mm_addsub_ps(__m128 A, __m128 B) {
// CHECK-LABEL: test_mm_addsub_ps
- // CHECK: call <4 x float> @llvm.x86.sse3.addsub.ps
+ // CHECK: call <4 x float> @llvm.x86.sse3.addsub.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}})
return _mm_addsub_ps(A, B);
}
__m128d test_mm_hadd_pd(__m128d A, __m128d B) {
// CHECK-LABEL: test_mm_hadd_pd
- // CHECK: call <2 x double> @llvm.x86.sse3.hadd.pd
+ // CHECK: call <2 x double> @llvm.x86.sse3.hadd.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}})
return _mm_hadd_pd(A, B);
}
__m128 test_mm_hadd_ps(__m128 A, __m128 B) {
// CHECK-LABEL: test_mm_hadd_ps
- // CHECK: call <4 x float> @llvm.x86.sse3.hadd.ps
+ // CHECK: call <4 x float> @llvm.x86.sse3.hadd.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}})
return _mm_hadd_ps(A, B);
}
__m128d test_mm_hsub_pd(__m128d A, __m128d B) {
// CHECK-LABEL: test_mm_hsub_pd
- // CHECK: call <2 x double> @llvm.x86.sse3.hsub.pd
+ // CHECK: call <2 x double> @llvm.x86.sse3.hsub.pd(<2 x double> %{{.*}}, <2 x double> %{{.*}})
return _mm_hsub_pd(A, B);
}
__m128 test_mm_hsub_ps(__m128 A, __m128 B) {
// CHECK-LABEL: test_mm_hsub_ps
- // CHECK: call <4 x float> @llvm.x86.sse3.hsub.ps
+ // CHECK: call <4 x float> @llvm.x86.sse3.hsub.ps(<4 x float> %{{.*}}, <4 x float> %{{.*}})
return _mm_hsub_ps(A, B);
}
__m128i test_mm_lddqu_si128(__m128i const* P) {
// CHECK-LABEL: test_mm_lddqu_si128
- // CHECK: call <16 x i8> @llvm.x86.sse3.ldu.dq
+ // CHECK: call <16 x i8> @llvm.x86.sse3.ldu.dq(i8* %{{.*}})
return _mm_lddqu_si128(P);
}