[Mlir-commits] [mlir] [mlir][linalg] Pack matmul pass (PR #89782)

[Diego Caballero]

https://github.com/dcaballe commented:

Hey, thanks for the contribution! This is exciting! A couple of comments:

* Why are you framing this as a CPU-only transformation? Are there any implicit limitations that we should be aware of?
* How configurable the layout would be? Could we do plain transposes, decide which operands to transpose and how many blocking levels to use? (Sorry if I missed something while skimming through the code)

https://github.com/llvm/llvm-project/pull/89782