About the GPU COE at the University of Oxford
The University of Oxford was awarded a GPU Center of Excellence in 2012 in recognition of its ongoing work in parallel computing research and education using NVIDIA GPUs and the NVIDIA CUDA parallel programming environment. The GPU COE at the University of Oxford is based within the Oxford e-Research Centre (OeRC) which promotes innovative computational methodologies in multidisciplinary collaborations across the university. GPU computing collaborations covering a very wide range of applications currently exist with the departments of Chemistry, Engineering Science, Physics and Statistics, as well as the Mathematical Institute. Oxford University and the GPU COE are also the lead in a consortium which has acquired a 372-GPU system, the largest in the UK.
As a GPU COE, Oxford will utilize equipment and grants provided by NVIDIA to support a number of research and academic programs across its mathematics, physical and life sciences divisions, including:
About the PIs
Mike Giles is a Professor of Scientiï¬c Computing in the Mathematical Institute, a founder member of the Oxford-Man Institute of Quantitative Finance and an Associate Director of OeRC. He studied mathematics at Cambridge before doing a PhD in Aeronautical Engineering at MIT, where he then taught for several years before moving to Oxford in 1992. He began working with CUDA when it was first released, and was made a CUDA Fellow in 2008 because of his work in random number generation and computational ï¬nance. He has also made contributions to the CUDA math and CURAND libraries. More recently, as well as working with others on applications in mathematical biology, statistics and astrophysics, he is leading an open-source initiative to develop OP2, a high level framework for the parallelization of unstructured grid applications in computational engineering and science on a variety of architectures including GPU clusters.
Anne Trefethen is Director of the OeRC and CIO of Oxford University. She has been active in the development of parallel software for many years, at Thinking Machines, the Cornell Theory Center, and the Numerical Algorithms Group, before joining Oxford in 2006 to establish the Oxford e-Research Center. OeRC has now grown into an internationally recognized group of over 50 multidisciplinary researchers committed to accelerating research through innovative technology, with a wide portfolio of projects in the sciences, social sciences and humanities, as well as in underlying technologies. Her own research interests currently are in energy-efficient computing and real-time astrophysics data processing in support of the Square Kilometre Array project.
Chris Holmes is a Professor of Biostatistics in the Department of Statistics, and a principal within the Wellcome Trust Centre for Human Genetics. After collaborating with Mike Giles in a CUDA implementation of sequential Monte Carlo methods, which attracted significant interest within the statistics community, he is now developing new GPU-based statistical tools for human genome sequencing analysis. The Wellcome Trust Centre is a world leader in the analysis and statistical modeling of whole-genome sequence data in population genetics and genetic epidemiology studies. A particular focus for the GPU research is the use of Hidden Markov Models in cancer genomics to detect copy-number aberrations in tumor cells.
Jonathan Doye is a Lecturer in Theoretical Chemistry. His expertise lies in the ï¬eld of molecular simulations, and his research currently focusses on the dynamics of self-assembly in DNA nanotechnology systems. Through a combination of original coarse-grained modelling and GPU implementations, his group is able to simulate the behavior of large DNA molecules for relatively long time scales, equivalent to tens of seconds in experiments. This makes it possible to undertake systematic exploration of the nature of the self-assembly processes in these systems.