Nvidia Corp. on Tuesday announced the launch of Cambridge-1, a supercomputer it has built in the U.K. at an estimated cost of $100 million to support life sciences research.

Nvidia says that the supercomputer is the fastest deployed in the U.K. to date. 

“Cambridge-1 will empower world-leading researchers in business and academia with the ability to perform their life’s work on the U.K.’s most powerful supercomputer, unlocking clues to disease and treatments at a scale and speed previously impossible in the U.K.,” said Nvidia founder and Chief Executive Jensen Huang. “The discoveries developed on Cambridge-1 will take shape in the U.K., but the impact will be global, driving groundbreaking research that has the potential to benefit millions around the world.”

Cambridge-1 provides performance equivalent to eight petaflops when measured using Linpack, a popular benchmarking tool that tests systems’ speed by having them solve a series of mathematical equations. When running artificial intelligence workloads, Nvidia says that Cambridge-1’s performance jumps to 400 petaflops. One petaflop equals a quadrillion processing operations per second. 

Cambridge-1 is powered entirely by renewable energy and runs at a facility managed by data center operator Kao Data Ltd. The supercomputer uses 80 of Nvidia’s DGX A100 computing modules, which feature its top-end A100 data center graphics card for running artificial intelligence software. The modules also feature BlueField-2 data processing units. DPUs are auxiliary chips that perform logistical chores such as network management to free up a system’s main processors for other tasks, which has the effect of speeding up applications.

Nvidia’s engineers built Cambridge-1 using the SuperPOD architecture that the chip giant introduced last year. The challenge SuperPOD addresses is that building supercomputers has historically taken months or years of work. The architecture makes it possible to assemble DGX A100 computing modules into a supercomputer relatively easily by linking them together using networking gear from Nvidia’s Mellanox subsidiary. The result, the company has stated, is that the hardware deployment process can be reduced from years or months to just a few weeks in some cases.

Nvidia is making Cambridge-1 available to life sciences researchers to help advance their work. The company said that the initial set of projects it supports are being carried out in partnership with researchers at five organizations: AstraZeneca plc, GSK plc, Guy’s and St Thomas’ NHS Foundation Trust, King’s College London and Oxford Nanopore Technologies Ltd.

Nvidia is helping AstraZeneca to develop an AI model for studying chemical structures that could be used to create new medicines. Additionally, the company plans to use Cambridge-1 for a project aimed at harnessing machine learning algorithms to automatically analyze scans of tissue samples.

GSK will leverage Cambridge-1 to support its drug discovery process, while King’s College London and Guy’s and St Thomas’ NHS Foundation Trust are teaching AI models to generate synthetic brain images. The hope is that the synthetic brain images will enable researchers to gain a better understanding of brain diseases and help them develop ways of facilitating earlier diagnosis and treatment. 

“Through this partnership, we will be able to use a scale of computational power that is unprecedented in healthcare research,” said professor Sebastien Ourselin, the head of the School of Biomedical Engineering & Imaging Sciences at King’s College London. “It will be truly transformational for the health and treatment of patients.”

Oxford Nanopore Technologies, in turn, will use Cambridge-1 to improve its scientific algorithms. Oxford Nanopore Technologies provides DNA and RNA sequencing products that are used in fields ranging from healthcare to environmental monitoring. Nvidia said that the company estimates Cambridge-1 will allow it to reduce the duration of certain algorithm improvement tasks from days to hours. 

Photo: Nvidia