[Mlir-commits] [mlir] [mlir][linalg] Add mixed precision folding pattern in vectorize_children_and_apply_patterns TD Op (PR #148684)

[Adam Siemieniuk]

================
@@ -2348,6 +2348,8 @@ def VectorizeChildrenAndApplyPatternsOp :
     operation that is contained inside the vectorization target.
 
     This transformation supports the following attributes:
+    - `fold_mixed_precision_into_contract`: a `UnitAttr` to enable the folding of
----------------
adam-smnk wrote:

That intuitively makes me think about `vector.*_cast` ops which is unrelated.
`numerical_casts` also could indicate `vector.bitcast` etc.

Hmm, maybe just directly `fold_numerical_extensions_into_contract`? (a bit long but oh well)

https://github.com/llvm/llvm-project/pull/148684