[clang] [clang-tools-extra] [compiler-rt] [llvm] [X86] AMD Zen 6 Initial enablement (PR #179150)

[Simon Pilgrim via cfe-commits]

================
@@ -7341,4 +7341,31 @@ def int_x86_movrsdi : ClangBuiltin<"__builtin_ia32_movrsdi">,
                   [IntrReadMem]>;
 def int_x86_prefetchrs : ClangBuiltin<"__builtin_ia32_prefetchrs">,
         Intrinsic<[], [llvm_ptr_ty], []>;
+
+//===----------------------------------------------------------------------===//
+// BMM intrinsics
+def int_x86_avx512_vbitrevb_128 :
+  DefaultAttrsIntrinsic<[llvm_v16i8_ty], [llvm_v16i8_ty],
+                        [IntrNoMem]>;
+def int_x86_avx512_vbitrevb_256 :
+  DefaultAttrsIntrinsic<[llvm_v32i8_ty], [llvm_v32i8_ty],
+                        [IntrNoMem]>;
+def int_x86_avx512_vbitrevb_512 :
+  DefaultAttrsIntrinsic<[llvm_v64i8_ty], [llvm_v64i8_ty],
----------------
RKSimon wrote:

we shouldn't need the bitrev intrinsics - we can use bitreverse generics

https://github.com/llvm/llvm-project/pull/179150