[Mlir-commits] [mlir] Introduce `arith.scaling_extf` and `arith.scaling_truncf` (PR #141965)

[Krzysztof Drewniak]

krzysz00 wrote:

Note: I'd rather we not land this just yet because I'm still waiting to find out if potential hardware-specific lowerings of `arith.scaling_truncf` will perform the exponent subtraction that this code does.

I have a suspicion that the answer is "no" - that that adjustment is part of the scale _computation_ process, not the scale _application_ process, and so the semantics of scaling_truncf shouldn't include it.

https://github.com/llvm/llvm-project/pull/141965