[Mlir-commits] [mlir] [mlir][ArithToAMDGPU] Only use the scaled conversions for f8E8M0FNU scales (PR #221386)

[Hung-Kuan Tseng]

Tim096 wrote:

Ping @krzysz00 @tgymnich. I merged main to resolve a conflict with #219195. The question in my first comment is still open.

https://github.com/llvm/llvm-project/pull/221386