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

[Federico Bruzzone]

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

----------------
FedericoBruzzone wrote:

Thanks 🫶

For `matmul-f32.mlir` I switched to `REQUIRES: target={{(aarch64|arm64).*}}`.
For `matmul-i8.mlir` I kept `REQUIRES: arm-emulator`. Please correct me if I'm wrong, but `FEAT_I8MM` should be optional (?). Switching this one to `target=` too would pass on my machine (it has i8mm) but could SIGILL on AArch64 hosts without it. 

Also in `Vector/CPU/`{`ArmNeon`|`ArmSVE`}`/vector-contract-i8mm.mlir` is used. E.g.,
https://github.com/llvm/llvm-project/blob/f7075d0da225b3de1cc83345a3a74d2f310d1057/mlir/test/Integration/Dialect/Vector/CPU/ArmNeon/vector-contract-i8mm.mlir#L1

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