[Openmp-commits] [PATCH] D51875: [OPENMP][NVPTX] Add support for lastprivates/reductions handling in SPMD constructs with lightweight runtime.

[Jonas Hahnfeld via Phabricator via Openmp-commits]

Hahnfeld added a reviewer: Hahnfeld.
Hahnfeld added a comment.

I really, really dislike adding even more global buffers. `4096 * 32 * 56` are another 7MiB that are not usable for applications. What's wrong with using the existing ones?

Can you upload the CodeGen patch for reductions somewhere? I thought we need a global scratchpad buffer that is adressable for all teams?



================
Comment at: libomptarget/deviceRTLs/nvptx/src/option.h:37
 // memory.
-#if __CUDA_ARCH__ >= 600
+#if __CUDA_ARCH__ >= 900
+#define OMP_STATE_COUNT 32
----------------
This doesn't exist unless you have information that are not public yet. Volta is `720` at most.


Repository:
  rOMP OpenMP

https://reviews.llvm.org/D51875