Introduction

The Advanced Research Computing (ARC) service provides access to High Performance Computing (HPC) resources, support, and advice to researchers within the University of Oxford.

At the core of the ARC service are our HPC compute clusters and high-performance data storage systems. The compute clusters provides a large number of CPU and GPU based compute nodes as well as other specialised resources such as high memory nodes for manipulating and working with large data sets and outputs. These resources support a breadth of research at the University including computational fluid dynamics, bioinformatics, and machine learning/AI to name a few application areas. The majority of ARC’s resources are connected via a high bandwidth, low latency interconnect (InfiniBand) to facilitate fast and efficient inter-processor communication across the system, as well as provide high bandwidth connectivity to storage. Data storage systems connect to these clusters, and provide both fast (scratch) storage for the fast, low latency parallel storage demanded by modern HPC applications, and bulk storage for facilitating user home directories and project data storage.