Department of Computing GPU Cluster Guide

Update 27/9/2024

Ubuntu 24.04 upgrades are complete, please create new 24.04 compatible Python virtual environments (links in following steps) using a lab PC

A guide on how to submit GPU-enabled scripts to the departmental GPU cluster

Please note: this service is for members of the Department of Computing and its associates only. Members of other departments may want to consult the Research Computing Services (RCS) instead

Slurm is an open-source task scheduling system for managing the departmental GPU cluster. The DoC GPU cluster is a pool of NVIDIA GPUs that can be leveraged for machine learning using popular frameworks such as PyTorch and Tensorflow, or any CUDA-based code. This guide will show you how to submit your GPU-enabled scripts to work with the shared resource.

Read this guide to learn how to:

• connect to the submission host server and submit a test script
• start an interactive job (connect directly to a GPU exclusively for a time limit)
• compose a shell script that uses shared storage, a python environment, CUDA and your python scripts