[Mlir-commits] [mlir] [mlir][GPUToNVVM] enable fallback to generic LLVM lowering for math dialect in convert-gpu-to-nvvm pass (PR #165728)

[Yang Bai]

yangtetris wrote:

> In general the change seems ok. However, it's important to note that the convert-gpu-to-nvvm is not really meant to be used in production, instead it's expected that people use convert to llvm.

Thanks! Could you please elaborate a bit more why it's not recommended to use it in production? In fact I'm seeing this pass in some projects' production code.

https://github.com/llvm/llvm-project/pull/165728