[llvm] AMDGPU: Add tonearest and towardzero roundings for intrinsic llvm.fptrunc.round (PR #104486)

[Matt Arsenault via llvm-commits]

arsenm wrote:

> And what to do on global-isel path? I guess something similar. It will end up with similar complexity as before.

You use the same pattern, the SDNodeXForm implementation is just different and uses a custom renderer function. The overall thing will be 1/5th the code you have here for the 2x custom lowering 

https://github.com/llvm/llvm-project/pull/104486