Hewlett Packard Enterprise (HPE) and Dell Technologies have both announced UK supercomputer projects as part of a UK government initiative on artificial intelligence (AI) and exascale computing.
The UK government will invest £225 million in a joint project between HPE and the University of Bristol called Isambard-AI, which will become the most powerful supercomputer in the UK when it’s turned on in 2024.
Isambard-AI will be capable of 200 petaflops of performance under the Top500’s Linpack benchmark – around ten times that of the current fastest UK supercomputer, ARCHER2. It will also be capable of 21 exaflops of AI performance for training frontier generative AI models, such as new and advanced large language models (LLMs).
A further £75 million has been awarded to Dell Technologies, which has collaborated with Intel and the University of Cambridge to produce the Dawn supercomputer, which has been deployed and will be fully running in the next two months. The trio claim that it’s the UK’s fastest AI supercomputer at present.
Isambard-AI and Dawn combine to form a supercomputer cluster as part of the AI Research Resource (AIRR), a facility for AI research and innovation created through a £300 million government investment.
Once the AIRR is fully operational, the UK will host some of the most powerful supercomputers not only in Europe but the world. Through this funding and the UK’s AI Safety Summit, the government has attempted to legitimize the UK’s place as an international research hub.
The Universities of Bristol and Cambridge will each host part of the cluster, with a phase 2 for Dawn expected in 2024 that will increase its performance tenfold.
"Today's announcement of the UK’s major investment in AI supercomputing underscores its commitment to taking a global leadership position in AI,” said Justin Hotard, executive vice president and general manager, HPC, AI and Labs at HPE.
“The Isambard-AI system will harness world-leading supercomputing, including high-performance networking co-developed at HPE's Bristol labs, to provide the performance and scale required for compute-intensive AI projects. We are proud to partner with the UK Government and the University of Bristol to give UK researchers and industry access to Europe's largest AI system for open science."
More than a number: Your risk score explained
Learn about the benefits of continuous cyber risk scoring
DOWNLOAD NOW
Unlike HPE with Isambard-AI, Dell has not yet released expected performance numbers for Dawn, but will do so at the Supercomputing 23 conference, held from 12-17 November 2023 in Denver, Colorado.
The supercomputer has been formed from PowerEdge XE9640 servers, which contain Intel’s Xeon Scalable processors and Intel Data Center GPU Max accelerators. In Dawn, Intel has a chance to prove the power of its GPUs in comparison to Nvidia’s sought-after AI chips, in its goal to achieve AI hardware dominance by 2025.
“Collaborations like the one between the University of Cambridge, Dell Technologies and Intel, alongside strong inward investment, are vital if we want compute to unlock the high-growth AI potential of the U.K,” said Tariq Hussain, head of UK public sector at Dell Technologies.
“It is paramount that the government invests in the right technologies and infrastructure to ensure the UK leads in AI and exascale-class simulation capability. It’s also important to embrace the full spectrum of the technology ecosystem, including GPU diversity, to ensure customers can tackle the growing demands of generative AI, industrial simulation modeling and ground-breaking scientific research.”
Isambard-AI is being built using HPE’s Cray EX supercomputer, hosting 5,448 of Nvidia’s GH200 Grace Hopper Superchips connected using HPE Slingshot 11 interconnect technology. It will also be capable of almost 25 petabytes of storage using the Cray Clusterstor E1000.
HPE says that Isambard-AI will be optimized for AI data and will be housed in a self-cooled data center at the National Composites Center at Bristol and Bath Science Park. Excess heat will eventually be used to heat nearby buildings.
Once fully operational, the AIRR cluster will push the UK closer to its exascale goal and drive research in fields including climate modeling, AI, future energy systems such as fusion, and frontier science.
