Imperial College London chooses Intel Xeon 6 for latest HPC Supercomputer

Intel has announced plans for the new HX2 supercomputer at Imperial College London, built on Lenovo servers and powered by Intel® Xeon® 6 processors with Performance cores (P-cores). The water-cooled system – designed as part of the co-investment ICICLE collaboration – will be deployed this year, delivering a performance uplift for the next generation of high performance computing (HPC) and artificial intelligence (AI).

Karin Eibschitz Segal, Intel corporate vice president and interim general manager of the Data Center Group, commented: “This is a great win for Intel due in part to Xeon 6’s strong performance for bandwidth-bound AI and HPC workloads. The deployment of the HX2 supercomputer at Imperial College London marks a significant milestone in our commitment to advancing scientific research and promoting sustainability. We are proud to help tackle global challenges like climate and healthcare, all while lowering the total cost of ownership and environmental impact.”

The HX2 supercomputer equips Imperial’s academics and students to accelerate their research and position themselves for careers in HPC/AI careers. The new system will be the United Kingdom’s first academic deployment of Lenovo’s ThinkSystem SC750 V4 Neptune servers with direct liquid cooling in a commercial co-location data center.

“The Imperial investment of 10 million pounds for HX2 has ensured that we can provide, in collaboration with Intel and Lenovo, a high-quality, future-proof compute platform to for our researchers,” said Andrew Richards, director of Research Computing Services (RCS) at Imperial. “We have been able to accelerate our HPC refresh using the latest sustainable platform and enable our academics to accelerate their research.”

Dugan Witherick, head of RCS Platforms at Imperial, said: “The new direct water-cooled system, based on Lenovo’s Neptune ThinkSystem SC750 V4 Neptune platform, will push forward Imperial’s commitment to sustainability, whilst ensuring the highest quality compute for our community.”

Kelly Zhang, lecturer in Statistics in the Department of Mathematics at Imperial, said: “My research group is using both CPUs and GPUs for jobs for developing decision-making algorithms for healthcare settings. Specifically, many are working with the ICU electronic health record dataset MIMIC. I’m excited to see what HX2 can offer for my group’s workloads. Having such a powerful yet sustainable cluster is such a privilege.”

The HX2 system will be available as part of Imperial College London’s central Research Computing Service. It will support the next generation of researchers across all faculties (engineering, natural sciences, medicine and business and others) and provide them access to long-term groundwork, emerging technology and gender-balance initiatives.

“The selection of Lenovo Neptune™ liquid cooling for the HX2 supercomputer marks a significant step forward in sustainable research infrastructure,” said Kate Steele, EMEA HPC lead at Lenovo Infrastructure Solutions Group. “By combining the high performance of Intel Xeon 6 processors with our advanced liquid-cooled ThinkSystem SC750 V4 servers, Imperial gains a powerful, efficient platform capable of supporting demanding AI and HPC workloads while dramatically reducing energy consumption. This project reflects a broader shift we’re seeing across Europe, where institutions are aligning cutting-edge computing with ambitious sustainability goals.”

Residing in a co-location data center in London, the new HX2 supercomputer was built with a strong focus on sustainability. Based on Lenovo’s ThinkSystem SC750 V4 Neptune servers with Neptune direct liquid cooling technology, the HX2 solution delivers significantly greater compute density and energy efficiency than traditional air-cooled systems.

Lenovo Neptune technology removes up to 98% of system heat using warm water, enabling up to 40% lower power consumption1 and reduced reliance on air conditioning, and dramatically improves Power Usage Effectiveness. It also supports a circular energy model by recapturing waste heat for reuse or district heating. In large-scale deployments, this can lead to energy savings of up to 1 million kilowatt-hours per year per data hall, reducing both operational costs and carbon emissions.