Eviden deployment to boost DiRAC GPU supercomputing capacity up to 80 per cent

Eviden, the Atos business leading in advanced computing, and the University of Edinburgh have announced a three-year contract extension to increase the computing capacity of the BullSequana XH2000.

  • 11 months ago Posted in

This energy-efficient supercomputing system was delivered by Atos to facilitate the Extreme Scaling Service of DiRAC, allowing the UK science community to pursue cutting-edge research on a broad range of topics, from simulating the entire evolution of the universe to modelling the fundamental structure of matter.

As part of the contract, Eviden will expand the power of the existing system by providing additional hardware, including 256 NVIDIA A100 Tensor Core GPUs up to 80 per cent more capacity, allowing scientists across the Science and Technology Facilities Council (STFC) community to drive forward research in theoretical physics.

Additional DDN storage and NVIDIA Quantum InfiniBand networking will be provided with Eviden engineers installing, configuring and integrating the new GPU nodes into the state-of-the-art system, located in Scotland’s capital city.  

Professor Luigi Del Debbio, Professor of Theoretical Physics at the University of Edinburgh and Project Lead: “We are delighted to continue our strategic partnership with Eviden, an Atos business, which will enable our growing scientific community to harness increasingly powerful supercomputing resources and make potentially ground-breaking discoveries in all the research areas supported by DiRAC. The upgraded equipment will build on the success of the current system, further enhancing our HPC system offering and strengthening the University of Edinburgh’s position as a world-leading supercomputing hub.”

Emmanuel Le Roux, Group SVP, Head of Advanced Computing, HPC and AI, Eviden, at Atos Group: “The extension to our existing partnership with the University of Edinburgh is testament to the success of the BullSequana system with these upgrades providing a major boost to DiRAC’s high performance computing resources. Bolstering the existing system with additional GPU capability and storage will accelerate scientific research, helping to address real-world issues through the power of supercomputing.”

Realized with two next-generation large capacity recording technologies: HAMR and MAMR.
The first module of the exascale supercomputer JUPITER, named JEDI, is ranked first place in the...
New partnership reinforces Hitachi Vantara defense-in-depth strategic approach, combining company's...
Strategic partnership delivers easy-to-consume, highly cost-effective tape-based solution to...
Expanded portfolio serves Dell’s largest customers, cloud service providers (CSPs) and small...
Eviden expands global presence for customers’ core missions.
Veeam enhances Kubernetes-native data protection with Kasten V7.0, expanding VM and cloud-native...
Together, Portworx by Pure Storage and Red Hat OpenShift unify VMs and containers to meet...