[PATCH] D136311: [CUDA,NVPTX] Implement __bf16 support for NVPTX.

[Jakub Chlanda via Phabricator via llvm-commits]

jchlanda accepted this revision.
jchlanda added a comment.
This revision is now accepted and ready to land.

Looks good.



================
Comment at: llvm/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp:838
   }
 }
+static int getLdStRegType(EVT VT) {
----------------
New line here.


================
Comment at: llvm/lib/Target/NVPTX/NVPTXISelDAGToDAG.cpp:844
+    case MVT::bf16:
+    case MVT::v2f16:
+    case MVT::v2bf16:
----------------
Nice, this fixes the reg type for `v2f16` from `Float` to `Untyped`, looks like no test picked up on that before.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D136311/new/

https://reviews.llvm.org/D136311