[Openmp-commits] [PATCH] D62393: [OPENMP][NVPTX]Mark parallel level counter as volatile.

[Jon Chesterfield via Phabricator via Openmp-commits]

JonChesterfield added a comment.

Thanks for the explicit asm. One of the hazards of compiling cuda for amdgcn is that volatile doesn't imply atomic, so this is a clear warning that I'll have a bug on merge.



================
Comment at: libomptarget/deviceRTLs/nvptx/src/supporti.h:121
+INLINE void changeParallelLevel(int WarpId, int Val, bool IsIncrement) {
+  uint8_t &parLevel = parallelLevel[WarpId];
+#if defined(CUDA_VERSION) && CUDA_VERSION <= 8000
----------------
Could this function call getParallelLevel, increment/decrement the result, then call setParallelLevel?


Repository:
  rOMP OpenMP

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D62393/new/

https://reviews.llvm.org/D62393