Seff for GPU vRAM

Hi @Yann.Sagon , I was wondering is there a way to check how much of the GPU vRAM an already finished run has used? (something similar to seff)
Or is it only possible while the job is running with nvidia-smi?

Thank you very much in advance

Hi, right now it is only possible to get this value during the run.

We are installing Open XDMoD, maybe this information will be available as time-series, stay tuned.

1 Like

Hello,

You can also check the metrics using Grafana : hpc:hpc_clusters [eResearch Doc]

Rémy

1 Like