[clang] [llvm] [AArch64][clang][llvm] Add ACLE Armv9.7 matrix multiply-accumulate intrinsics (PR #193017)

[Jonathan Thackray via llvm-commits]

================
@@ -2834,6 +2834,11 @@ def int_aarch64_sve_fp8_fmmla
   : DefaultAttrsIntrinsic<[llvm_anyvector_ty],
                           [LLVMMatchType<0>, llvm_nxv16i8_ty, llvm_nxv16i8_ty],
                           [IntrReadMem, IntrInaccessibleMemOnly]>;
+def int_aarch64_sve_bfmmla_bf16
----------------
jthackray wrote:

> Honestly I don't think we need int_aarch64_sve_bfmmla and it would be great if that one could be removed as well

Is it worth removing `int_aarch64_neon_bfmmla` for similar reasons? Looks like it can share `int_aarch64_neon_fmmla`.

https://github.com/llvm/llvm-project/pull/193017