Major boost to research computing
16 Apr 2026
Research Lifecycle Programme (RLP) successfully upgrades University’s flagship high performance computing platform, CSF3.
The Research Lifecycle Programme (RLP), in collaboration with the Research IT (RIT) team, have successfully delivered a major upgrade to the University’s flagship high‑performance computing platform CSF3.
The CSF3 upgrade directly supports Manchester 2035 and the University’s Research Leap by strengthening our core research enablers. These improvements strengthen performance, security, and reliability across our High-Performance Computing (HPC) environment, ensuring our research community continues to benefit from world‑class on‑premise compute infrastructure.
Approaching its 15th anniversary, CSF3 remains one of the University’s most valuable and trusted research assets. It’s a platform that has supported thousands of projects and underpinned high‑impact discoveries. This upgrade continues that legacy, ensuring the service remains robust, sustainable and ready for the future needs of our research community.
Delivering a major upgrade with minimal disruption
Throughout 2025, CSF3 received a substantial refresh, including:
- Modern AMD processors,
- An updated operating system and system libraries,
- Installation of NVIDIA L40S GPUs.
Despite the scale and complexity of this work, CSF3 remained online and accessible to researchers throughout, with only four days of planned downtime
Researchers are already experiencing the benefits. As one user, Tom Flint, Department of Materials, shared: “The new AMD nodes allow me to have a greater throughput of simulations; running more simulations simultaneously to explore the parameter space and investigate the scaling behaviours of various iterations of the developed numerical implementation.”
Why this upgrade matters
- Additional new AMD processors delivers a substantial performance boost, offering higher throughput and improved efficiency. This upgrade enables researchers to run larger, more complex simulations more quickly and effectively.
- Upgrading the operating system and system libraries enhances overall performance, security, and stability.
- With access to the latest compiler optimisations, modern CPU features, and up‑to‑date patches, researchers can work more efficiently and use a broader range of current scientific software.
- The introduction of NVIDIA L40S GPUs further expands CSF3’s capabilities, providing powerful acceleration for AI, as part of our strategic Digital Inside and Out leap, and machine‑learning workloads such as deep learning, generative AI, modelling, simulation, and high‑throughput inference.
- Expanding the CSF3’s capabilities ensures the platform continues to support advanced, data‑intensive research across the University.
How do I request access to CSF3?
CSF3 is available to all researchers at the University, across all disciplines. To find out more about the service and how to request access, please visit: