Convert vperm2f128 and vperm2i128 intrinsics back to using llvm intrinsics. Unfortunately, these instructions have behavior that can't be modeled with shuffle vector.

llvm-svn: 154906

3 files changed, 7 insertions, 4 deletions

diff --git a/clang/test/CodeGen/avx-shuffle-builtins.c b/clang/test/CodeGen/avx-shuffle-builtins.c
index 538ae50c938..d071f825aa9 100644
--- a/clang/test/CodeGen/avx-shuffle-builtins.c
+++ b/clang/test/CodeGen/avx-shuffle-builtins.c
@@ -48,18 +48,18 @@ __m256 test_mm256_permute_ps(__m256 a) {
 
 __m256d test_mm256_permute2f128_pd(__m256d a, __m256d b) {
   // Check if the mask is correct
-  // CHECK: shufflevector{{.*}}<i32 2, i32 3, i32 6, i32 7>
+  // CHECK: @llvm.x86.avx.vperm2f128.pd.256
   return _mm256_permute2f128_pd(a, b, 0x31);
 }
 
 __m256 test_mm256_permute2f128_ps(__m256 a, __m256 b) {
   // Check if the mask is correct
-  // CHECK: shufflevector{{.*}}<i32 12, i32 13, i32 14, i32 15, i32 4, i32 5, i32 6, i32 7>
+  // CHECK: @llvm.x86.avx.vperm2f128.ps.256
   return _mm256_permute2f128_ps(a, b, 0x13);
 }
 
 __m256i test_mm256_permute2f128_si256(__m256i a, __m256i b) {
   // Check if the mask is correct
-  // CHECK: shufflevector{{.*}}<i32 0, i32 1, i32 2, i32 3, i32 8, i32 9, i32 10, i32 11>
+  // CHECK: @llvm.x86.avx.vperm2f128.si.256
   return _mm256_permute2f128_si256(a, b, 0x20);
 }
diff --git a/clang/test/CodeGen/avx2-builtins.c b/clang/test/CodeGen/avx2-builtins.c
index e56e55d205a..7d166b56263 100644
--- a/clang/test/CodeGen/avx2-builtins.c
+++ b/clang/test/CodeGen/avx2-builtins.c
@@ -677,7 +677,7 @@ __m256i test_mm256_permute4x64_epi64(__m256i a) {
 }
 
 __m256i test_mm256_permute2x128_si256(__m256i a, __m256i b) {
-  // CHECK: shufflevector{{.*}}<i32 2, i32 3, i32 6, i32 7>
+  // CHECK: @llvm.x86.avx2.vperm2i128
   return _mm256_permute2x128_si256(a, b, 0x31);
 }
 
diff --git a/clang/test/CodeGen/builtins-x86.c b/clang/test/CodeGen/builtins-x86.c
index 30138d6374f..acb5554db4d 100644
--- a/clang/test/CodeGen/builtins-x86.c
+++ b/clang/test/CodeGen/builtins-x86.c
@@ -414,6 +414,9 @@ void f0() {
   tmp_V4i = __builtin_ia32_cvttpd2dq256(tmp_V4d);
   tmp_V4i = __builtin_ia32_cvtpd2dq256(tmp_V4d);
   tmp_V8i = __builtin_ia32_cvttps2dq256(tmp_V8f);
+  tmp_V4d = __builtin_ia32_vperm2f128_pd256(tmp_V4d, tmp_V4d, 0x7);
+  tmp_V8f = __builtin_ia32_vperm2f128_ps256(tmp_V8f, tmp_V8f, 0x7);
+  tmp_V8i = __builtin_ia32_vperm2f128_si256(tmp_V8i, tmp_V8i, 0x7);
   tmp_V4d = __builtin_ia32_vinsertf128_pd256(tmp_V4d, tmp_V2d, 0x7);
   tmp_V8f = __builtin_ia32_vinsertf128_ps256(tmp_V8f, tmp_V4f, 0x7);
   tmp_V8i = __builtin_ia32_vinsertf128_si256(tmp_V8i, tmp_V4i, 0x7);