[AVX512] Add some of the FP cast intrinsics

Part of <rdar://problem/17688758>

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@214315 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Headers/avx512fintrin.h b/lib/Headers/avx512fintrin.h
index a336f6d74634ad040d0b497509ee35e490af7785..2d3a8bba45e1efe996b24eb35d4d2dd4734e4218 100644 (file)
--- a/lib/Headers/avx512fintrin.h
+++ b/lib/Headers/avx512fintrin.h
@@ -117,6 +117,33 @@ _mm512_set1_epi64(long long __d)
   return (__m512i)(__v8di){ __d, __d, __d, __d, __d, __d, __d, __d };
 }
 
+/* Cast between vector types */
+
+static __inline __m512d __attribute__((__always_inline__, __nodebug__))
+_mm512_castpd256_pd512(__m256d __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3, -1, -1, -1, -1);
+}
+
+static __inline __m512 __attribute__((__always_inline__, __nodebug__))
+_mm512_castps256_ps512(__m256 __a)
+{
+  return __builtin_shufflevector(__a, __a, 0,  1,  2,  3,  4,  5,  6,  7,
+                                          -1, -1, -1, -1, -1, -1, -1, -1);
+}
+
+static __inline __m128d __attribute__((__always_inline__, __nodebug__))
+_mm512_castpd512_pd128(__m512d __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1);
+}
+
+static __inline __m128 __attribute__((__always_inline__, __nodebug__))
+_mm512_castps512_ps128(__m512 __a)
+{
+  return __builtin_shufflevector(__a, __a, 0, 1, 2, 3);
+}
+
 /* Arithmetic */
 
 static __inline __m512d __attribute__((__always_inline__, __nodebug__))

diff --git a/test/CodeGen/avx512f-builtins.c b/test/CodeGen/avx512f-builtins.c
index ddeb3b6365fe97a4e72683dbddf1b43ecbe16222..8a2dd07230d9b523ad682f047f8e880cc4dc81b1 100644 (file)
--- a/test/CodeGen/avx512f-builtins.c
+++ b/test/CodeGen/avx512f-builtins.c
@@ -102,3 +102,10 @@ __m512d test_mm512_set1_pd(double d)
   // CHECK: insertelement <8 x double> {{.*}}, i32 7
   return _mm512_set1_pd(d);
 }
+
+__m512d test_mm512_castpd256_pd512(__m256d a)
+{
+  // CHECK-LABEL: @test_mm512_castpd256_pd512
+  // CHECK: shufflevector <4 x double> {{.*}} <i32 0, i32 1, i32 2, i32 3, i32 undef, i32 undef, i32 undef, i32 undef>
+  return _mm512_castpd256_pd512(a);
+}