[Openmp-commits] [openmp] [Libomptarget] Fix RPC-based malloc on NVPTX (PR #72440)

Johannes Doerfert via Openmp-commits openmp-commits at lists.llvm.org
Mon Nov 20 12:40:30 PST 2023


================
@@ -486,6 +494,16 @@ struct CUDADeviceTy : public GenericDeviceTy {
       Res = cuMemAllocManaged(&DevicePtr, Size, CU_MEM_ATTACH_GLOBAL);
       MemAlloc = (void *)DevicePtr;
       break;
+    case TARGET_ALLOC_DEVICE_NON_BLOCKING: {
+      CUstream Stream;
+      if (Res = cuStreamCreate(&Stream, CU_STREAM_NON_BLOCKING))
----------------
jdoerfert wrote:

So, there is no AsyncInfoTy object here, but I am also confused why we can't simply call getStream and later return it again. Stream creation and destruction is not free, hence our resource pools. 

https://github.com/llvm/llvm-project/pull/72440


More information about the Openmp-commits mailing list