[llvm] [NVPTX] Add support for f16 fabs (PR #116107)

[Justin Fargnoli via llvm-commits]

================
@@ -864,10 +864,15 @@ NVPTXTargetLowering::NVPTXTargetLowering(const NVPTXTargetMachine &TM,
     AddPromotedToType(Op, MVT::bf16, MVT::f32);
   }
   for (const auto &Op : {ISD::FABS}) {
-    setOperationAction(Op, MVT::f16, Promote);
     setOperationAction(Op, MVT::f32, Legal);
     setOperationAction(Op, MVT::f64, Legal);
----------------
justinfargnoli wrote:

> abs.f64 requires sm_13 or higher.

9.7.3.9. [Floating Point Instructions: abs](https://docs.nvidia.com/cuda/parallel-thread-execution/index.html#floating-point-instructions-abs)

https://github.com/llvm/llvm-project/pull/116107