[Mlir-commits] [mlir] [mlir][Vector] Fold vector.step compared to constant (PR #161615)

[Kunwar Grover]

Groverkss wrote:

> Nice contribution!
> 
> Couldn't we make `vector.step` a `ConstantLike` op? That would make the arith.constant canonicalizations to also trigger for vector.step?

Interesting, do you mean make it something like : `arith.constant iota<end>` ? Or is there a ConstantLike interface (maybe we should add that if not?) I'll have a deeper look into the codebase too, but could use more info on what you had in mind.

https://github.com/llvm/llvm-project/pull/161615