[Openmp-commits] [PATCH] D95752: [OpenMP][DeviceRTL] Extract shuffle idiom and port it to declare variant

[Johannes Doerfert via Phabricator via Openmp-commits]

jdoerfert planned changes to this revision.
jdoerfert added a comment.

I'll update this tomorrow, certain parts are not great.



================
Comment at: openmp/libomptarget/deviceRTLs/common/include/shuffle.h:88
+                                     int32_t SrcLane) {
+// In Cuda 9.0, the *_sync() version takes an extra argument 'mask'.
+#if CUDA_VERSION >= 9000
----------------
JonChesterfield wrote:
> Seems bad, both because it's a macro instead of variant, and because I thought we'd already got rid of that macro
we will, ptx selection. I'll actually update this patch.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D95752/new/

https://reviews.llvm.org/D95752