[Mlir-commits] [mlir] [mlir][vector] Add fold to transfer_{read, write} vector<1xT> (PR #196598)

[Erick Ochoa Lopez]

amd-eochoalo wrote:

> Btw, is there an example that shows that this won't lower to vector.transfer_read when the permutation map is not a minor identity ?

You mean it won't lower to vector.load? The example you posted is the same example I posted (just without running       transform.apply_patterns.vector.transfer_permutation_patterns first). I ran the example and the same thing happened. On main, it stays as a vector.transfer_read. With this change, it lowers to vector.load.

https://github.com/llvm/llvm-project/pull/196598