chrisdorney - Fotolia

Oxford University gets deep learning supercomputer

Academics and the IT industry will be able to use university’s Atos-supplied supercomputer as a service

Oxford University has bought a supercomputer from Atos to support deep learning and artificial intelligence (AI) applications and will provide access to it as a cloud service.

Academics and the IT industry will be able to use the supercomputer to test deep learning applications and proof of concepts, as part of the university’s Joint Academic Data Science Endeavour (Jade) project. It will offer organisations super computing capabilities as a service.

The supercomputer, funded by the Engineering and Physical Sciences Research Council, is powered by Nvidia graphic processing units (GPUs) and Atos’ Extreme Factory technology. This will create a public deep learning cloud.

The Alan Turing Institute will be a primary user of the new service, helping industry and academia to access the latest cognitive analytics technologies.

Mike Giles, a professor at Oxford University, said: “The national GPU supercomputer will be used specifically to develop and support deep learning and AI applications in addition to GPU-based supercomputing.

The system comprises a cluster of 22 Nvidia DGX1 deep learning supercomputers, which makes it the largest GPU-based system in the UK, according to Atos. It will be hosted by Atos at the STFC Hartree Centre in Daresbury, near Warrington.

Andy Grant, head of big data and high-performance computing at Atos, said: “Deep learning technologies are set to revolutionise many industries. We believe UK industry and academia will benefit tremendously from the Jade facility and by offering it as a service, we are enabling our clients to develop proof of concepts and to run production AI workloads on the system without having to spend vast amounts of money.”

Read more on Artificial intelligence, automation and robotics