In the session, we will demonstrate how to install and configure Enroot and Pyxis by NVIDIA on a Slurm cluster to run containerized applications based on Ubuntu images.
A wide variety of applications across many industries run on high-performance computing (HPC) infrastructures. Software diversity, complex workflows, and security considerations make providing and maintaining a cluster infrastructure challenging. In this context, container technologies remove the need to build complex environments and simplify the application installation and execution process.
As part of the initiatives to expand Ubuntu adoption on HPC, we aim to show how to easily run containers based on Ubuntu images regardless of the host system of the HPC cluster. Containerization enables the use of Ubuntu to run applications on such environments and creates the opportunity to demonstrate that Ubuntu can be used to deploy HPC workloads.
In collaboration, Canonical and Omnivector created a community to promote discussions around Ubuntu and HPC topics, which is open to anyone interested in contributing or learning (https://matrix.to/#/#ubuntu-hpc:matrix.org).
Session author's bio
Jaime is a software engineer and cluster administrator at Omnivector. He is a PhD. candidate in computer science and has experience in parallel and high-performance computing. He is also a member of the Ubuntu HPC community team.
|Level of Difficulty||Intermediate|