University of Oxford Launches Top GPU-Based Supercomputer

The “Emerald” is the most powerful system of its kind in the UK

A consortium of UK academic institutions, led by University of Oxford, has launched “Emerald” – the most powerful GPU-accelerated computer in the UK.

The Emerald features 372 Nvidia Tesla M2090 GPUs, capable of more than 114 teraflops of performance. It will be mainly used for research, in the fields such as astrophysics, genomics and nanotechnology.

It was also announced that the University of Oxford has joined the elite club of GPU supercomputing, by being recognised by Nvidia as a centres of excellence for its CUDA architecture.

A job for Supercomputer!

The Emerald will be housed in the newly established Centre for Innovation in High Performance Computing at the Rutherford Appleton Laboratory in Didcot. It was created by the e-Infrastructure South Consortium, which includes the Universities of Oxford, Bristol, and Southampton and University College London.

The Centre will provide High Performance Computing (HPC) services for research purposes, as well as train scientists and engineers to use the Emerald.

“The Emerald supercomputer forms part of the government’s £145 million investment in e-infrastructure and will be an invaluable asset to business and universities,” said David Willetts, Minister for Universities and Science. “It will drive growth and innovation, encourage inward investment in the UK and keep us at the very leading edge of science.”

Nvidia has recognised  the  University of Oxford, which leads the project, as one of  18 centres of excellence worldwide for CUDA (Compute Unified Device Architecture), Nvidia’s parallel computing architecture that uses GPUs for computation instead of CPUs. The only other CUDA cente in the UK is the University of Cambridge.

As a CoE, University of Oxford will use equipment and grants provided by Nvidia to pursue some of its most ambitious projects. These will include real-time pulsar detection application for the forthcoming Square Kilometre Array – the world’s most powerful telescope, due to be finished in 2024.

“The CUDA Center of Excellence award reflects Oxford’s strength in scientific computing, as well as the success of OeRC [Oxford e-Research Centre] in developing and championing new approaches to computing, while working with application specialists across the university to bring these benefits to their research,” said Professor Anne Trefethen, chief information officer at the University of Oxford.

“With Nvidia’s support, we can continue to enhance our undergraduate projects and summer bursaries focused on GPU computing, and develop new programs to reach larger numbers of researchers and students,” she added.

Last week, we have reported on CORE, a partnership between University of Cambridge and Imperial College London that united their HPC infrastructure to create the fastest Intel-based system in the UK. Unlike Emerald, CORE has opened access to its supercomputer as a commercial service.

How much do you know about microprocessors? Take our quiz!