[Mlir-commits] [mlir] [mlir][ArmNeon] Add linalg.matmul e2e tests (PR #212809)

[Andrzej Warzyński]

================

----------------
banach-space wrote:

Do we require the sizes to be dynamic? We will for SVE, but we can cross that bridge when we get there.

What you proposed in #213698 makes sense - we should only target I8MM _after_ packing (i.e. when the data is known no to require masking).

https://github.com/llvm/llvm-project/pull/212809