[PATCH] D118470: [X86][avx512fp16] Promote fp16 to fp32 for frem.

[Phoebe Wang via Phabricator via llvm-commits]

pengfei added inline comments.


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:808-809
   // Some FP actions are always expanded for vector types.
   for (auto VT : { MVT::v4f32, MVT::v8f32, MVT::v16f32,
                    MVT::v2f64, MVT::v4f64, MVT::v8f64 }) {
     setOperationAction(ISD::FSIN,      VT, Expand);
----------------
Can we add `MVT::v8f16, MVT::v16f16, MVT::v32f16` here instead? I think it's fine not adding tests to others at the present.


================
Comment at: llvm/lib/Target/X86/X86ISelLowering.cpp:2013
+      setOperationAction(ISD::FREM,                   MVT::v32f16, Expand);
+      setOperationAction(ISD::STRICT_FREM,            MVT::v32f16, Expand);
+
----------------
STRICT nodes are set Expand by default. We don't need to set again.
The same below.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D118470/new/

https://reviews.llvm.org/D118470