Selin Yıldırım

Thanks for your attention! I actually meant the latter. DGBMV with 2 threads is already slower than DGBMV in a single thread (in other words sequential run). Also, the input...

> Perhaps it would also make sense to try a simple driver without OpenMP to exclude any other overhead or contention ? I have adjusted my driver so that it...

> You need to set some threshold to not spin threads on small(-er than that) inputs, like e.g gemm.c for extensive example. Like to cover time scheduling thread vs actual...

> There is some small latency communicating initial tasks between threads, certainly not affecting your experiment. > > I am trying to recreate test case from your description. > >...

In my system, Multithreaded OpenBlas threads are sleeping and running in turn (although in my own openmp parallel program all threads always in running state), and their CPU utilization is...

After installing the plugin in this fork, nebuly-nvidia device plugin pod gets stuck at init state with the log that states the same thing. This is for description of my...

Thank you Kevin, I have installed the NVIDIA DRA too, but I am afraid you don't support the CRDs for deploying mps sharing yet.

As an update, I have obtained the logs of nebuly-nos gpu-agent that crashes with CrashLoopBackOff error: `selin@selin-csl:~$ kubectl logs nos-1698250532-gpu-agent-gqfdz -n nebuly-nos Defaulted container "nos-1698250532-gpu-agent" out of: nos-1698250532-gpu-agent, kube-rbac-proxy {"level":"info","ts":1698250893.187797,"logger":"setup","msg":"loaded...

I am facing the same issue. pod throwing Insufficient nvidia.com/gpu-1gb error.