[Mlir-commits] [mlir] [mlir][x86] Lower packed type vector.contract to AMX dot-product (online-packing) (PR #188192)

[Arun Thangamani]

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

----------------
arun-thmn wrote:

Kind of true. I re-factored the code slightly. 
After the re-write for reduction loop with dept 1, We assign the `outerLoop` with `innerLoop` This helps the final store back to the acc uses the same code for the both reduction loop with depth `1` or `2`.

https://github.com/llvm/llvm-project/pull/188192