Re: [Starpu-devel] Starpu stats about disk, OOC, and swapping

[Amani Alonazi <amani.alonazi@kaust.edu.sa>]

All kernels except memecpy are using the asynchronous CUDA submission, i.e., :

cudaStream_t stream = starpu_cuda_get_local_stream();
KERNEL<<<..., stream >>>();
#ifndef STARPU_CUDA_ASYNC
  cudaStreamSynchronize( stream );
#endif

I'll reinstall and update starpu to 1.3 branch.

Many thanks! 

On Thu, Jan 17, 2019 at 7:47 PM Samuel Thibault <samuel.thibault@inria.fr> wrote:

About the missing overlapping of transfers and CUDA computation, I'm
realizing that in StarPU 1.2.x that wasn't indeed as well supported as
it is now, you may want to try the starpu-1.3 branch, which I have btw
fixed to avoid making GPU workers care about disk operations.

Samuel

--

Amani

---

This message and its contents, including attachments are intended solely for the original recipient. If you are not the intended recipient or have received this message in error, please notify me immediately and delete this message from your computer system. Any unauthorized use or distribution is prohibited. Please consider the environment before printing this email.