Living by numbers

28 Aug 2003

In recent years computers have become a key weapon in the armoury of scientists looking to solve scientific problems. But sometimes these problems are too complex to be solved using ordinary computers.

“What scientists and programmers do in this case is use a special computing system that is more powerful than a regular computer,” explains Audrey Crosbie (pictured) business development manager at the Trinity Centre for High Performance Computing (TCHPC), Ireland’s leading supercomputer facility at Trinity College Dublin (TCD). “Quite often these systems are just multiple computers linked together by a fast switch [electrical connection] that can transfer data from each of the computers in the ‘cluster’ to the others. This way, the job gets split into parts with each section of the system working on its own part and the job gets done a lot faster.”

In international terms, Ireland’s supercomputer facilities are quite modest. As well as the TCHPC, there are quite powerful machines at University College Cork and NUI Galway, as well as smaller ones at Dublin City University, University College Dublin and the Dublin Institute for Advanced Studies.

Founded in 1997, the TCHPC was the brainchild of Professor Jim Sexton, a mathematician, theoretical physicist and the centre’s director, in addition to his full-time role as a lecturer in TCD’s School of Mathematics. It was something of a painful birth, he recalls. “We made a number of abortive attempts to get a supercomputer up and running in Ireland but we were constrained by cost. Supercomputers were hugely expensive then and we just couldn’t afford one. In the mid-Nineties the total budget for research in Ireland was very, very small and if we had wanted a supercomputer we would have consumed that budget two or three times over,” he says.

In the end, an organisation called Atlantic Philanthropies, which in recent years has part-funded many infrastructural projects in Irish science along with the Higher Education Authority (HEA), donated the first machine. A second machine was later jointly purchased with Queens University and installed in Belfast.

The centre currently employs 13 staff and has an operating budget of more than €500k per annum, 20pc of which comes from the university. Much of the remaining funding is provided under the HEA’s Programme for Research in Third Level Institutes (PRTLI) scheme. The centre has been awarded €9m under the PRTLI Cycle 3 to fund a major interdisciplinary project known as IITAC 2 (The Institute for Information Technology and Advanced Computation) that links researchers in various scientific disciplines. Unfortunately, the freezing of the PRTLI fund earlier this year owing to the poor state of public finances has meant that only €4m of the funding has been released so far. The IITAC 2 project was to involve the installation of an advanced visualisation facility in the shape of a 3D booth that would allow scientists to view realistic simulations of molecular interactions. It would have been the first such facility in Ireland but unless the funding is reinstated, there will be no money to purchase the critically important visualisation suite.

The funding crisis also means there is no money to upgrade essential computer equipment. It is only thanks to a bridging loan from the college that the centre has been able to replace the original, donated supercomputer, which has reached the end of its useful life. The centre will shortly take delivery of a replacement machine following the completion of a full tendering process. The supercomputer, an IBM machine, features 80 dual Xeon processors as well as four terabytes of disk storage space. Of its purchase price of €400,000, €250,000 has come from the HEA’s PRTLI fund, the remainder from TCD funds.

“We are very grateful for the college funding that will tide us over for about a year but if we don’t get HEA funding the IITAC project, which forms a large part of what we do here, will start to crumble,” warns Crosbie.

While IITAC accounts for a major chunk of the work done at the facility, it has undertaken a range of other project work with both internal and external partners. The centre is participating in a growing number of EU projects. Although these projects take longer to happen, they tend to have larger budgets and involve more participants. One recent example was Monte Carlo, an EU-funded food risk project led by the Institute of European Food Studies. The role of the TCHPC was to create a website for the network of European food scientists and develop a software system, accessible via the internet, that uses high-performance computing to assess the risk to humans of exposure to food chemicals and nutrients.

“The role of the software is to analyse the risk to people of eating above the recommended intake of a certain chemical,” notes researcher Cronan McNamara. “Our high-performance computer meant we could give results back to scientists within 20 to 30 minutes as opposed to two days.”

This project has spawned another, Crème, which aims to commercially exploit the technology and has been given initial funding to do so by Enterprise Ireland. TCD has an impressive record of bringing ideas to market, notes Crosbie. “There have been about 40 campus companies that have arisen from the science and technology areas in the past 10 years or so and countless royalty agreements by the college for commercialisation of technology,” she says, giving examples such as Iona Technologies and computer games specialist Havok.

In addition to project work, the TCHPC has a training function. Three staff members — Cronan McNamara, Dermot Frost and Dr Geoff Bradley — lecture in High Performance Computing in the School of Computer Science.

Looking to the future, Sexton believes that the centre will continue to broaden its skill base through working on multidisciplinary projects. “Everyone says that multidisciplinary science is the future but we’ve been trying to make it work for a number of years now. We would like to see ourselves make contributions in supply chain management, telecoms networks modelling and in new materials design but all based on a couple of underlying ideas. What we’re trying to build up in the centre is capacity and expertise where people can jump across from one area to another,” he says.

Sexton is also keen to cultivate closer contacts with third parties who can help improve the standard of research done by the centre. As an example, Sexton has had a long-time association with IBM and in fact is spending the summer in upstate New York as a visiting scientist working on Blue Gene (, a supercomputer that will attempt to solve the mysteries of protein formation.

Based on a different type of architecture to standard computers, Blue Gene comprises 65,000 credit-card-sized computers linked together. “Blue Gene is a technology that allows you to develop a very high performance, special-purpose computer at a considerably lower cost than if you were to buy the same kind of performance using PCs,” he explains.

Sexton hopes that his involvement with Blue Gene will eventually allow him to apply this new and hugely powerful type of supercomputer to some of the projects the TCHPC will be tackling in the future.

Sexton is proud of what the centre has achieved in its first six years but cautions that unless public funds continue to be made available to fund research, the future of Ireland’s leading supercomputer facility as well as the others around the country will be threatened. “When you build up a programme and put six years into developing a team and it comes crashing down around you because that particular year there wasn’t any resources in the government budget, that just doesn’t work. You need long-term investment to continue and to grow and to give people confidence in their future,” he concludes.

By Brian Skelly