How do we monitor GPU ressources?

HPC Support
1

Hi,

I am struggling to visualise the consumption of my node.

I run an AI model on a gpu node in baobab and the run can be really memory intensive. I would like to monitor it to optimize the application but I don’t understand the layout of the Grafana.

Do you have some time to explain it to me

Thanks,
Thibaut

Monitoring ressource
2

Did you checked this dashboard?

image
image887×443 33.6 KB

ps: there is a bug right now: only the Yggdrasil GPUs are listed, we’ll fix that soon.

3

Okay,
The PS should explain why I was not able to find the baobab gpu I was interested in. Thanks.

4

Hello,

The bug is now fixed, you should be able to see baobab gpus in this dashboard.

Remy