What is 'Hex'?

Hex is a collection of GPU-equipped hosts onto which single- multi- or GPU-processor jobs can be executed. The current, exact specification of all nodes can be found on the hardware page and we are actively looking to expand our capabilities over time to align better with larger HPC systems.

Hex is not HEC - it is not yet another high-performance system but is rather a modular development system, which attempts to reach feature parity with larger systems to allow students, researchers and staff to design software to run on HEC, BEDE, AWS, Google Compute, and other systems, but in a much more development-friendly environment.

Note that Hex is not directly affiliated with any of these systems (although Hex and HEC are both located at Lancaster University) and simply attempts to reach feature parity wherever it can with these; where budget and practicality allow.

What technologies does Hex use?

The software stack for Hex is a combination of Docker, Swarm and Kubernetes, Docker Compose, Svelte, Javascript and Python, plus the required GPU drivers and operating systems for the hosts. The hardware stack includes a large number of small form factor desktop PCs running Ubuntu, each equipped with a 20-core Intel-based CPU, an A3000 GPU with 12GB of VRAM, and 32GB of RAM or equivalent technology. In the future we would like to expand this to include ARM64-based hardware hosts to enable very large numbers of densely co-located instances to be deployed

User software is executed in one or more docker containers distributed across the cluster and we support a number of base images targetting the spaCy pipeline, with and without GPU support. However, intel-based docker container images can be executed on Hex, although Hex's security requirements may mean that some have limited functionality

How do I reference Hex in published works?

If your work uses Hex, please use the following reference in your bibliography (in LaTeX/BibTex format):

@misc{UcrelHex,

title = {{UCREL - Hex}; A shared, hybrid multiprocessor system},

author = {Vidler, John AND Rayson, Paul},

abstract = {Hex is a collection of GPU equipped hosts onto which single- multi- or GPU-processor jobs can be executed hosted at Lancaster University, UK as part of the School of Computing and Communications and the UCREL group.},

howpublished = {\url{https://github.com/UCREL/hex}}, note = {Accessed: 2024} }