[Openmp-commits] [openmp] [OpenMP][libomptarget][RFC] extend libomptarget with mechanism to execute fill memory on the target (PR #73801)

Johannes Doerfert via Openmp-commits openmp-commits at lists.llvm.org
Wed Nov 29 08:55:38 PST 2023


================
@@ -1387,6 +1391,14 @@ Error CUDADeviceTy::dataExchangeImpl(const void *SrcPtr,
   return Plugin::check(Res, "Error in cuMemcpyDtoDAsync: %s");
 }
 
+/// Fill memory on the target device (aka memset)
+Error CUDADeviceTy::fillMemoryImpl(void *Ptr, int32_t Val, uint64_t NumValues,
+                                   AsyncInfoWrapperTy &AsyncInfoWrapperTy) {
+  CUdeviceptr DevPtr = reinterpret_cast<CUdeviceptr>(Ptr);
+  CUresult Res = cuMemsetD32(DevPtr, Val, static_cast<size_t>(NumValues));
----------------
jdoerfert wrote:

why not the async version?

https://github.com/llvm/llvm-project/pull/73801


More information about the Openmp-commits mailing list