Setting up Dockerized Slurm Cluster on Raspberry Pis

1 · · May 29, 2024, 6:30 p.m.
Background Here is the scenario - you need to find a way to share the compute power that ‘you’ (or your organization) possess, in an effective manner. In my case, we need to share the GPU servers we have with an entire department of researchers. However, the catch is this - you are not allowed to touch the bare metal, you can do whatever you need to, but only within containers. At present, a group of researchers are allocated one node to work on, which proves to be very inefficient. Top research...