[clang] [llvm] [X86][AVX10.2] Support AVX10.2-MINMAX new instructions. (PR #101598)

[Phoebe Wang via llvm-commits]

================
@@ -0,0 +1,219 @@
+/*===--------------- avx10_2_512minmaxintrin.h - AVX10_2_512MINMAX intrinsics
+ *-----------------===
+ *
+ * Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+ * See https://llvm.org/LICENSE.txt for license information.
+ * SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+ *
+ *===-----------------------------------------------------------------------===
+ */
+#ifndef __IMMINTRIN_H
+#error                                                                         \
+    "Never use <avx10_2_512minmaxintrin.h> directly; include <immintrin.h> instead."
+#endif // __IMMINTRIN_H
+
+#ifndef __AVX10_2_512MINMAXINTRIN_H
+#define __AVX10_2_512MINMAXINTRIN_H
+
+#define _mm512_minmaxne_pbh(A, B, C)                                           \
+  ((__m512bh)__builtin_ia32_vminmaxnepbf16512(                                 \
+      (__v32bf)(__m512bh)(A), (__v32bf)(__m512bh)(A), (int)(C)))
+
+#define _mm512_mask_minmaxne_pbh(W, U, A, B, C)                                \
+  ((__m512bh)__builtin_ia32_selectpbf_512(                                     \
+      (__mmask32)(U),                                                          \
+      (__v32bf)__builtin_ia32_vminmaxnepbf16512(                               \
----------------
phoebewang wrote:

We prefer to use `_mm512_minmaxne_pbh` instead of builtin. The same below.

https://github.com/llvm/llvm-project/pull/101598