British Computer Society

What does state of the art supercomputing looks like?

• 18:30: Networking & Refreshments
• 19:00: Presentation
• 20:30: Conclusions

Professor Martyn F. Guest is Director of Advanced Research Computing (ARCCA) at Cardiff University. He has led a variety of high performance and distributed computing initiatives over the past twenty years, from his time as coordinator of SERC’s Distributed Computing Initiative in the late 1980’s, to his position from 2002 as Leader of the Terascale Applications Team within the UK’s National Terascale Facility, HPCx. He spent three years as Senior Chief Scientist and Technical Group Leader of the High Performance Computational Chemistry Group at the Pacific Northwest National Laboratory (Richland, Washington, USA) before returning to the UK in 1995 as Associate Director of the Computational Science and Engineering Department (CSED) at the STFC Daresbury Laboratory. He was responsible for all activities in computational chemistry and High Performance and Distributed computing within CSED before taking up his present position at Cardiff University in Mach 2007.

Prof. Guest is an Honorary Professor in Physics at the University of Edinburgh and an Honorary Reader in Chemistry at the University of Manchester. He is a founder member of EPSRC’s Technology Watch Panel (TWP), a member of the corresponding Applications Panel (AP), and has provided ongoing advice to a wide variety of agencies and organisations in both the UK (RC’s, AWE, GCHQ) and abroad (NSF and DOE (USA), NSERC (Canada), SFI (Ireland), CHPC (South Africa)). His research interests cover a variety of topics in the development and application of computational chemistry methods on high performance computers. Prof. Guest is lead author of the GAMESS-UK electronic structure program, and has written or contributed to more than 230 articles in the areas of Theoretical and Computational Chemistry, and High Performance Computing.

Advanced Research Computing - A Vehicle for Research Enablement

There is no doubt that computational science and engineering (CS&E) has arrived firmly on the national and international scene and is here to stay. CS&E is concerned with predictive scientific modelling and simulation activity, understood in the widest sense to include fields outside the conventional scientific domain, and draws heavily on computational infrastructure. CS&E provides the third leg of modern scientific enquiry, alongside experiment and theory. Just as experiments cannot always provide the level of detail, information, or insight craved, so too the theoretical complexity of many systems denies us the ability to analyse their behaviour with pencil and paper alone. Computer-based simulation using high performance computing (HPC) is the only way forward in both of these situations. It is already an established approach for solving problems across most fields of science and engineering and indeed well beyond them – for example, in understanding complexity in financial, economic and human social systems.

This lecture will describe work underway in Advanced Research Computing and HPC at Cardiff University. ARCCA (http://www.cf.ac.uk/arcca), a new division within the University, will co-ordinate, support and develop ARC services for researchers. Staffed with experts in the field, ARCCA will help and support research needs through a variety of services. In providing a spectrum of technology offerings that includes one of the largest CONDOR pools in the UK, ARCCA has recently brought into service a range of dedicated high-end computing equipment in a new, state-of-the-art data centre. This equipment includes the third most powerful commodity cluster within the UK dedicated to University Research.

In describing this infrastructure and support framework, we outline the various fields of related research underway within the University in both the "established" HPC scientific domains - physics, chemistry, engineering, earth sciences, computer science - and in the "emerging" HPC disciplines - the arts, humanities and social sciences.