[PATCH] D118398: [NVPTX] Remove fmin/fmax.NaN.f64 again

Thu Jan 27 13:36:56 PST 2022

csigg added a comment.

In D118398#3277316 <https://reviews.llvm.org/D118398#3277316>, @tra wrote:

>> Is it fine to leave the change in llvm/lib/Target/NVPTX/NVPTXInstrInfo.td as is?
>
> I'd add a comment around `FMINNAN` that f64 variants do not actually exist. As long as we never lower to it it's not worth the trouble of refactoring tablegen to deal with a minor exception.

Makes sense. Thanks.

================
Comment at: llvm/lib/Target/NVPTX/NVPTXISelLowering.cpp:575-577
     setFP16OperationAction(Op, MVT::f16, GetMinMaxAction(Expand), Expand);
     setOperationAction(Op, MVT::f32, GetMinMaxAction(Expand));
     setFP16OperationAction(Op, MVT::v2f16, GetMinMaxAction(Expand), Expand);
----------------
tra wrote:
> Shouldn't all these actions be `Legal`, given that we have instructions for them?
They are `Legal` for `sm_80+`, see the `GetMinMaxAction()` implementation.
For `f16` and `v2f16`, we additionally require that fp16 is not disabled (see `setFP16OperationAction()`)

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D118398/new/

https://reviews.llvm.org/D118398