Untitled

Pass CUDA stream containers through a C API?
enum QDPCudaStream { TRANSFER=0 , KERNEL=1 };

  cudaStream_t * Wcudastreams;
  cudaEvent_t * WCopied;

  void * CudaGetKernelStream() {
    return (void*)&Wcudastreams[KERNEL];
  }

  void CudaCreateStreams() {
    Wcudastreams = new cudaStream_t[2];
    for (int i=0; i<2; i++) {
      cudaStreamCreate(&Wcudastreams[i]);
    }
    WCopied = new cudaEvent_t;
    cudaEventCreate(WCopied);
  }

void * ptr_cudaStream = CudaGetKernelStream();

kernel<<< bl,th,sh , *((cudaStream_t*)ptr_cudaStream) >>>( kernel_args );