Research Cluster on the use of novel hardware for real-time computing for the Digital Economy

Lead Research Organisation: University of Oxford
Department Name: Oxford e-Research Centre

Abstract

Field-programmable gate arrays (FPGAs) and graphical processing units (GPUs) are two kinds of computing hardware which offer the potential for factor 10-100 improvements in price/performance and enery efficiency compared to standard Intel and AMD processors. Although this hardware capability has existed for many years, it is only recently that the software tools have been developed to make it relatively easy for this power to be harnessed.This proposal is to fund a Research Cluster which will bring together computing experts who are specialists in working with FPGAs and GPUs, and application specialists working in key Digital Economy area, bioinformatics, medical image processing, digital media and computational finance. The interaction will be two-way, the computing specialists learning more about the computational needs and challenges of the applications, and the application specialists learning about the potential offered by these hardware platforms and the challenges in achieving this potential in practice.A number of different mechanisms will be used to build a community spirit amongst this group of researchers, most of whom have not previously known of each other's work. In addition to a kick-off meeting and a final meeting, both linked to a Multicore and Reconfigurable Supercomputing Conference, there will be 3 workshops to address different specific issues. There will be a number of scoping studies in which computing and application specialists will work together to assess the potential of FPGAs and/or GPUs for certain applications. There will be literature reviews which will collect and disseminate the state-of-the-art on certain key underlying issues. There will also be undergraduate student projects (funded via other EPSRC mechanisms) which are a cost effective way of tackling more application araeas and at the same time starting to train the next generation of computational scientists. Finally, there will be a website through which all of our results will be communicated to the wider UK research community.The main outcome from the project will be proposals for future collaborative research, building upon the foundations laid by this proposal. It is possible there may be one over-arching proposal covering all of the areas in this present proposal. Another possibility is that there may be three proposals, one on FPGAs, one on GPUs, and one on the common issues around accuracy and algorithms for finite precision arithmetic. A third possibility is that proposals will be constructed around an application theme, such as bio-informatics. Deciding on the best approach for future research is one of the challenges to be addressed in the project.

Publications

10 25 50
publication icon
Giles M (2011) Performance analysis of the OP2 framework on many-core architectures in ACM SIGMETRICS Performance Evaluation Review

publication icon
Giles MB (2014) Trends in high-performance computing for engineering calculations. in Philosophical transactions. Series A, Mathematical, physical, and engineering sciences

publication icon
Klingbeil G (2012) Fat versus Thin Threading Approach on GPUs: Application to Stochastic Simulation of Chemical Reactions in IEEE Transactions on Parallel and Distributed Systems

publication icon
Lee A (2010) On the utility of graphics cards to perform massively parallel simulation of advanced Monte Carlo methods. in Journal of computational and graphical statistics : a joint publication of American Statistical Association, Institute of Mathematical Statistics, Interface Foundation of North America

publication icon
Reguly I (2013) Finite Element Algorithms and Data Structures on Graphical Processing Units in International Journal of Parallel Programming

publication icon
Roldao A (2010) A High Throughput FPGA-Based Floating Point Conjugate Gradient Implementation for Dense Matrices in ACM Transactions on Reconfigurable Technology and Systems

publication icon
Serylak M (2013) Observations of transients and pulsars with LOFAR international stations and the ARTEMIS backend in Proceedings of the International Astronomical Union

 
Description GPUs and FPGAs are capable of providing very significant improvements in energy efficient computing compared to traditional CPUs.
Exploitation Route GPUs are being used increasingly by the finance industry for Monte Carlo simulation, and also for computational engineering in a variety of different industry areas. This network grant has led to a number of new research initiatives and collaborations, and also, indirectly, to the purchase of the UK's largest GPU supercomputer.
Sectors Aerospace/ Defence and Marine,Financial Services/ and Management Consultancy,Pharmaceuticals and Medical Biotechnology

URL http://www.oerc.ox.ac.uk/research/ccoe
 
Description Future-proof massively-parallel execution of multi-block applications
Amount £282,844 (GBP)
Funding ID EP/K038494/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Academic/University
Country United Kingdom
Start 09/2013 
End 08/2016
 
Description Multi-layered abstractions for PDEs
Amount £233,639 (GBP)
Funding ID EP/I006079/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Academic/University
Country United Kingdom
Start 09/2010 
End 03/2014
 
Description e-Infrastructure South Consortium - Centre for Innovation
Amount £2,820,000 (GBP)
Funding ID EP/K000136/1 
Organisation Engineering and Physical Sciences Research Council (EPSRC) 
Sector Academic/University
Country United Kingdom
Start 01/2011 
End 03/2012
 
Description CUDA Programming on NVIDIA GPUs 
Form Of Engagement Activity A talk or presentation
Part Of Official Scheme? No
Geographic Reach International
Primary Audience Postgraduate students
Results and Impact One week course on CUDA programming on NVIDIA GPUs, available to both academics and non-academics.

Lots of the students have since gone on to use CUDA programming in their research.
Year(s) Of Engagement Activity 2008,2009,2010,2011,2012
URL http://people.maths.ox.ac.uk/gilesm/cuda/