[Mlir-commits] [mlir] [vector][mlir] Canonicalize to shape_cast where possible (PR #140583)

[llvmlistbot at llvm.org]

MaheshRavishankar wrote:

@dcaballe  and @banach-space  see this post here https://discourse.llvm.org/t/rfc-update-to-general-design-section-of-operation-canonicalizations-in-mlir/79355?u=maheshravishankar . This talks about the how vector.transpose captures more information than a vector.shape_cast and how you cannot always go from shape_cast to transpose.

This is exactly the issue with treating vector.shape_cast as "canonical" representation for transposes and hoping that we can lift back to the original representation always.


https://github.com/llvm/llvm-project/pull/140583