Electronic communication and sharing of models in a virtual environment is an important part of the productivity of modern global businesses. Virtual working is a key enabler of an advanced nation.
Dr Lesley Thompson
The e-Infrastructure South consortium, comprised of the Universities of Bristol, Oxford and Southampton and University College London (UCL), was formed in 2011 with the aim of co-developing and sharing e-infrastructure capabilities. With £3.8 million of funding from the Engineering and Physical Sciences Research Council (EPSRC), the consortium has established a Centre for Innovation at the Science and Technology Facilities Council’s (STFC) Rutherford Appleton Laboratory (RAL).
The funding was awarded as part of the UK government’s £145 million programme of investment in e-infrastructure, a term which encompasses hardware, applications software, support services for users, and training and skills for computational sciences.
Two supercomputers, Emerald and Iridis 3, have been procured by the consortium with the use of this funding. The latter is hosted by the University of Southampton while Emerald – the largest and most powerful GPU-based supercomputer in the UK and one of the largest in Europe – is housed at RAL. These high performance computing (HPC) systems will aid researchers across disciplines as diverse as bioinformatics, climate modelling, healthcare, astrophysics, and even financial modelling.
"They [the supercomputers] will drive growth and innovation, encourage inward investment in the UK and keep us at the very leading edge of science," commented Minister for Universities and Science David Willetts MP. Partners from business and industry will also be engaged, making use of the facilities and collaborating with academic partners to develop data-driven applications, software and simulation techniques.
I interviewed Dr Lesley Thompson, Director of EPSRC’s Research Base, who spoke at the official launch of the e-Infrastructure South Consortium and unveiling of the two supercomputers earlier this week. She began by describing the rationale behind and importance of the government’s investment in this area using the manufacturing industry as an example.
"The issues facing the UK manufacturing industry centre on how we can maintain our position as a high value-add manufacturing nation. In order to remain among the top ten nations in global manufacturing, it is no longer possible to hand-make and craft things. You need to be able to model products before taking them into production and the only way to do that is through access to high performance computing (HPC) and e-infrastructure."
L-R: Dai Jenkins, David De Roure, Anne Trevethen and Lesley Thompson at the launch seminar
Dr Thompson cited Jaguar Land Rover – who have a production plant in Coventry and a research lab in Gaydon, Warwickshire – as an example of how business is becoming ever more global. This has some obvious repercussions for working practices and the kind of infrastructure that is needed to support them.
"Electronic communication and sharing of models in a virtual environment is an important part of the productivity of modern global businesses," Dr Thompson pointed out. "Research and production teams can be geographically distributed, which cuts down time-to-market. Virtual working is a key enabler of an advanced nation."
"It’s about connectivity and access to the right machines and software across the ecosystem. It’s about using leading edge researchers and academics in universities and coupling them with businesses that are interested in pursuing ways to make better products, so that we can rebalance the economy of the UK."
The global nature of modern business – and indeed of modern science – makes connections with Europe and the rest of the world critically important and highlights the advantages of cooperation. I asked Dr Thompson how the UK can be said to measure up, with regard to e-infrastructure, as a competitor on the world stage.
"We don’t have the most powerful machine," she admitted. "HECToR is 32nd in the world at the moment, according to the biannual Top 500 list, but our strength lies in the support structures that are in place around it. We have an edge in software development. If you have better software you can get more out of the machine.
"However, it is no accident that in the UK we have Microsoft research labs at Cambridge, IBM at Hursley, and Hewlett Packard in Bristol. Many of the large IT companies, as well as global outfits like BAE or Rolls-Royce, interact with our research scientists on e-infrastructure issues because they have an international science lead."
Emerald is the largest and most powerful GPU-based supercomputer in the UK, with 372 NVIDIA Tesla processors giving it a sustained capability of 114 teraflops. It is operated by the STFC’s e-Science department at RAL in Didcot, Oxfordshire. Iridis 3, the next generation of supercomputer at Southampton, utilises 12,000 cores
In the talk she gave at the launch gathering, Dr Thompson laid emphasis on the role of people
in the e-infrastructure vision of the research councils, highlighting the sizeable amount of money that has been poured into training and support systems. I asked her to explain in more detail the importance of this in future.
"The way that researchers operate these days, they very rarely go to the library," she said. "They access publications online, and much of the data they collect is gathered online. The challenge for us is to ensure that the older individuals teaching in universities are able to access and instil into younger people the use of up-to-date digital research methodologies. We shouldn’t let anyone graduate without understanding what e-infrastructure is and what it will mean for their working life."
To keep talented computational scientists in UK universities and research institutes may require a paradigm shift in an academic reward structure which is based around producing influential research papers and obtaining funding, however.
"Computational scientists enable other scientists to do their research, and they have a unique skillset, but the existing reward system in most universities doesn’t recognise them," Dr Thompson remarked. "One of the things we have to think about in the e-Leadership Council (eLC) is how the reward structure can be modified so there is an academic scientific career path open for computational scientists, because they are needed both in universities and in business. The message from the business community is that they are finding it very hard to recruit really good computational scientists.
"It is part of EPSRC’s job to train people to meet the long-term needs of the UK. If we hear computational scientists are needed, we work on how we can develop more but set against the backdrop of reward structures sometimes not being quite right. At the end of the day, we are a government agency with £800 million to spend each year and not without influence, so we can work to change these systems."