[Mlir-commits] [mlir] [mlir][arith][transforms] Adds Truncf f32 to f4e2m1 (PR #144157)

[Jakub Kuderski]

https://github.com/kuhar commented:

I don't understand why the scalar and vector cases use different implementation. I think we should use the same implementation for both. This would be easier with a sequence of selects so that you don't need to look up in 2d vectors.

https://github.com/llvm/llvm-project/pull/144157