[llvm] [AArch64] Add @llvm.experimental.vector.match (PR #101974)

[Ricardo Jesus via llvm-commits]

https://github.com/rj-jesus commented:

Thanks very much for the feedback @david-arm!
I've addressed most of your comments, except for the TTI vs TLI hook and cost model. As for the cost model, I thought the decision to use the intrinsic would effectively be binary - unless the hardware supports MATCH (or similar), I doubt we'd want to use the generic lowering. I'm happy to add this though if you think it would be useful. What do you think?

https://github.com/llvm/llvm-project/pull/101974