[Mlir-commits] [mlir] [mlir][gpu] Support `minimumnumf` and `maximumnumf` in GPU dialect (PR #226903)

[Artem Gindinson]

AGindinson wrote:

> Could you pls clarify “lower through appropriate arith ops” in the commit message? The connection is indirect: this patch maps the new GPU reduction kinds to the existing vector combining kinds.

@chelini I've reworded it to reference the vector combining kinds. Focusing this on Arith as one of the conversion targets was an unfortunate choice of words, sorry for the confusion.

https://github.com/llvm/llvm-project/pull/226903