1 files changed, 123 insertions, 0 deletions
diff --git a/test/CodeGen/X86/sse41-pmovxrm-intrinsics.ll b/test/CodeGen/X86/sse41-pmovxrm-intrinsics.ll
new file mode 100644
index 0000000..55faf4d
--- /dev/null
+++ b/test/CodeGen/X86/sse41-pmovxrm-intrinsics.ll
@@ -0,0 +1,123 @@
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=x86-64 -mattr=+sse4.1 | FileCheck %s --check-prefix=CHECK --check-prefix=SSE41
+; RUN: llc < %s -mtriple=x86_64-apple-darwin -mcpu=x86-64 -mattr=+avx2 | FileCheck %s --check-prefix=CHECK --check-prefix=AVX
+
+define <8 x i16> @test_llvm_x86_sse41_pmovsxbw(<16 x i8>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovsxbw
+; SSE41: pmovsxbw (%rdi), %xmm0
+; AVX:  vpmovsxbw (%rdi), %xmm0
+  %1 = load <16 x i8>* %a, align 1
+  %2 = call <8 x i16> @llvm.x86.sse41.pmovsxbw(<16 x i8> %1)
+  ret <8 x i16> %2
+}
+
+define <4 x i32> @test_llvm_x86_sse41_pmovsxbd(<16 x i8>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovsxbd
+; SSE41: pmovsxbd (%rdi), %xmm0
+; AVX:  vpmovsxbd (%rdi), %xmm0
+  %1 = load <16 x i8>* %a, align 1
+  %2 = call <4 x i32> @llvm.x86.sse41.pmovsxbd(<16 x i8> %1)
+  ret <4 x i32> %2
+}
+
+define <2 x i64> @test_llvm_x86_sse41_pmovsxbq(<16 x i8>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovsxbq
+; SSE41: pmovsxbq (%rdi), %xmm0
+; AVX:  vpmovsxbq (%rdi), %xmm0
+  %1 = load <16 x i8>* %a, align 1
+  %2 = call <2 x i64> @llvm.x86.sse41.pmovsxbq(<16 x i8> %1)
+  ret <2 x i64> %2
+}
+
+define <4 x i32> @test_llvm_x86_sse41_pmovsxwd(<8 x i16>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovsxwd
+; SSE41: pmovsxwd (%rdi), %xmm0
+; AVX:  vpmovsxwd (%rdi), %xmm0
+  %1 = load <8 x i16>* %a, align 1
+  %2 = call <4 x i32> @llvm.x86.sse41.pmovsxwd(<8 x i16> %1)
+  ret <4 x i32> %2
+}
+
+define <2 x i64> @test_llvm_x86_sse41_pmovsxwq(<8 x i16>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovsxwq
+; SSE41: pmovsxwq (%rdi), %xmm0
+; AVX:  vpmovsxwq (%rdi), %xmm0
+  %1 = load <8 x i16>* %a, align 1
+  %2 = call <2 x i64> @llvm.x86.sse41.pmovsxwq(<8 x i16> %1)
+  ret <2 x i64> %2
+}
+
+define <2 x i64> @test_llvm_x86_sse41_pmovsxdq(<4 x i32>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovsxdq
+; SSE41: pmovsxdq (%rdi), %xmm0
+; AVX:  vpmovsxdq (%rdi), %xmm0
+  %1 = load <4 x i32>* %a, align 1
+  %2 = call <2 x i64> @llvm.x86.sse41.pmovsxdq(<4 x i32> %1)
+  ret <2 x i64> %2
+}
+
+define <8 x i16> @test_llvm_x86_sse41_pmovzxbw(<16 x i8>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovzxbw
+; SSE41: pmovzxbw (%rdi), %xmm0
+; AVX:  vpmovzxbw (%rdi), %xmm0
+  %1 = load <16 x i8>* %a, align 1
+  %2 = call <8 x i16> @llvm.x86.sse41.pmovzxbw(<16 x i8> %1)
+  ret <8 x i16> %2
+}
+
+define <4 x i32> @test_llvm_x86_sse41_pmovzxbd(<16 x i8>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovzxbd
+; SSE41: pmovzxbd (%rdi), %xmm0
+; AVX:  vpmovzxbd (%rdi), %xmm0
+  %1 = load <16 x i8>* %a, align 1
+  %2 = call <4 x i32> @llvm.x86.sse41.pmovzxbd(<16 x i8> %1)
+  ret <4 x i32> %2
+}
+
+define <2 x i64> @test_llvm_x86_sse41_pmovzxbq(<16 x i8>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovzxbq
+; SSE41: pmovzxbq (%rdi), %xmm0
+; AVX:  vpmovzxbq (%rdi), %xmm0
+  %1 = load <16 x i8>* %a, align 1
+  %2 = call <2 x i64> @llvm.x86.sse41.pmovzxbq(<16 x i8> %1)
+  ret <2 x i64> %2
+}
+
+define <4 x i32> @test_llvm_x86_sse41_pmovzxwd(<8 x i16>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovzxwd
+; SSE41: pmovzxwd (%rdi), %xmm0
+; AVX:  vpmovzxwd (%rdi), %xmm0
+  %1 = load <8 x i16>* %a, align 1
+  %2 = call <4 x i32> @llvm.x86.sse41.pmovzxwd(<8 x i16> %1)
+  ret <4 x i32> %2
+}
+
+define <2 x i64> @test_llvm_x86_sse41_pmovzxwq(<8 x i16>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovzxwq
+; SSE41: pmovzxwq (%rdi), %xmm0
+; AVX:  vpmovzxwq (%rdi), %xmm0
+  %1 = load <8 x i16>* %a, align 1
+  %2 = call <2 x i64> @llvm.x86.sse41.pmovzxwq(<8 x i16> %1)
+  ret <2 x i64> %2
+}
+
+define <2 x i64> @test_llvm_x86_sse41_pmovzxdq(<4 x i32>* %a) {
+; CHECK-LABEL: test_llvm_x86_sse41_pmovzxdq
+; SSE41: pmovzxdq (%rdi), %xmm0
+; AVX:  vpmovzxdq (%rdi), %xmm0
+  %1 = load <4 x i32>* %a, align 1
+  %2 = call <2 x i64> @llvm.x86.sse41.pmovzxdq(<4 x i32> %1)
+  ret <2 x i64> %2
+}
+
+declare <2 x i64> @llvm.x86.sse41.pmovzxdq(<4 x i32>)
+declare <2 x i64> @llvm.x86.sse41.pmovzxwq(<8 x i16>)
+declare <4 x i32> @llvm.x86.sse41.pmovzxwd(<8 x i16>)
+declare <2 x i64> @llvm.x86.sse41.pmovzxbq(<16 x i8>)
+declare <4 x i32> @llvm.x86.sse41.pmovzxbd(<16 x i8>)
+declare <8 x i16> @llvm.x86.sse41.pmovzxbw(<16 x i8>)
+declare <2 x i64> @llvm.x86.sse41.pmovsxdq(<4 x i32>)
+declare <2 x i64> @llvm.x86.sse41.pmovsxwq(<8 x i16>)
+declare <4 x i32> @llvm.x86.sse41.pmovsxwd(<8 x i16>)
+declare <2 x i64> @llvm.x86.sse41.pmovsxbq(<16 x i8>)
+declare <4 x i32> @llvm.x86.sse41.pmovsxbd(<16 x i8>)
+declare <8 x i16> @llvm.x86.sse41.pmovsxbw(<16 x i8>)