search

det-lab / jupyterhub-deploy-kubernetes-jetstream

CDMS JupyterHub deployment on XSEDE Jetstream
0 stars 1 forks source link

Interface for GPU allocation? #89

Open pibion opened 1 year ago

pibion commented 1 year ago

@zonca I'm not sure if we'd need funding for you to work on this, but wanted to ask -

We have an allocation for GPU hardware on Jetstream2 but right now don't have an interface set up to access it through e.g. a jupyterhub instance. Is this something you could set up? Or something I might be able to set up by deploying the current settings on different machines?

zonca commented 1 year ago

I have written a tutorial about it: https://docs.jetstream-cloud.org/general/k8sgpu/ However, I haven't found a way yet to have a single Kubernetes deployment with some GPU and some CPU nodes. So we will need to investigate that possibility, or have 2 separate JupyterHub deployments, one with CPU nodes and one with GPU nodes.

This is a significant amount of work, so I will need funding.

pibion commented 1 year ago

I think having two separate deployments, one with CPU nodes and one with GPU nodes, would be just fine.

I'll look into funding options. I'm not sure if ACCESS provides anything anymore - as far as I know the ECSS program is no more.