[Mlir-commits] [llvm] [mlir] [mlir][EmitC] Add MathToEmitC pass for math function lowering to EmitC (PR #113799)

[Corentin Ferry]

cferry-AMD wrote:

Thanks for this PR @recursion-man! I second the approach to go straight to EmitC instead of through a libm detour with `func.call` to allow overloading, and besides this also gives us the opportunity to call functions outside of libm (or go with inline implementations) if, say, the type isn't supported by libm.

https://github.com/llvm/llvm-project/pull/113799