INDUSTRY
Scottish supercomputer nominated for two BCS IT Industry Awards
A unique supercomputer called ‘Maxwell’ – built in Scotland by the FHPCA Alliance with the support of Scottish Enterprise – is in the running for two awards at next month’s prestigious British Computer Society IT Industry Awards 2007. The nominations recognise Maxwell’s unique combination of innovation and energy-efficiency. The categories are: Best Use of Green Technology Project Award and the BT Flagship Award for Innovation. The FHPCA was established in 2004 to promote the use of field programmable gate arrays (FPGAs) as an alternative to microprocessors. With traditional microprocessor-based solutions hitting performance limits, there is a growing need for new technologies that address the demand for ever greater processing capability without demanding large amounts of space and power. 
Maxwell uses FPGAs and requires much less space and cooling than a conventional microprocessor system. It is also over 100 times more energy efficient and up to 300 times faster. Several companies have been using Maxwell since its launch in March this year. Impressive results have already been achieved in the oil & gas and medical imaging sectors. One of the first companies to use the supercomputer, Aberdeen-based Offshore Hydrocarbon Mapping plc (OHM), found that its application ran significantly faster on Maxwell. OHM is the world’s leading provider of Controlled Source Electromagnetic Imaging (CSEMI) services to the offshore oil industry. Dr Lucy MacGregor, Chief Scientific Officer of OHM, said: “Improving the performance of our data processing and visualisation services is key to our continued success and we are very excited about the code speed-ups we’ve achieved with Maxwell.” Of course, many other sectors of industry could benefit too, particularly the financial sector. In order to demonstrate the power of FPGAs and the Maxwell system when handling Monte Carlo calculations for the investment banking sector, the Alliance decided to implement the Black Scholes algorithm which is commonly used to calculate future stock prices. Spectacular results were obtained. In particular, the algorithm ran 320 times faster per FPGA on Maxwell than when compared to the equivalent algorithm running on the host PC. This demonstration of the Black Scholes algorithm has shown the potential benefits of FPGAs to the financial sector and the Alliance is currently pursuing opportunities with several leading investment banks, some of whom have been conducting their own experiments with FPGAs. Dr Mark Parsons, Commercial Director of EPCC said: “Maxwell has been created so that businesses can easily investigate FPGAs. We’ve already seen it give companies a competitive advantage. We want more businesses to come and test their codes on Maxwell to see whether it will be useful for them too.” Maxwell was built by the FHPCA (FPGA High Performance Computing Alliance). The Alliance is led by EPCC at the University of Edinburgh and comprises Alpha Data, Nallatech, Xilinx, Algotronix, Scottish Enterprise and the iSLI. The BCS IT Industry Awards are the leading hallmark of success among practitioners in the IT industry today. Anna Duckworth, BCS Head of Corporate Marketing, said: “Year on year the standard of entries increases and this year is no exception with innovative projects that showcase excellence in IT. The next stage of judging will be very difficult as our expert panel endeavours to select the outright winners for each award.” The winners will be announced on Thursday 6th December.
Maxwell uses FPGAs and requires much less space and cooling than a conventional microprocessor system. It is also over 100 times more energy efficient and up to 300 times faster. Several companies have been using Maxwell since its launch in March this year. Impressive results have already been achieved in the oil & gas and medical imaging sectors. One of the first companies to use the supercomputer, Aberdeen-based Offshore Hydrocarbon Mapping plc (OHM), found that its application ran significantly faster on Maxwell. OHM is the world’s leading provider of Controlled Source Electromagnetic Imaging (CSEMI) services to the offshore oil industry. Dr Lucy MacGregor, Chief Scientific Officer of OHM, said: “Improving the performance of our data processing and visualisation services is key to our continued success and we are very excited about the code speed-ups we’ve achieved with Maxwell.” Of course, many other sectors of industry could benefit too, particularly the financial sector. In order to demonstrate the power of FPGAs and the Maxwell system when handling Monte Carlo calculations for the investment banking sector, the Alliance decided to implement the Black Scholes algorithm which is commonly used to calculate future stock prices. Spectacular results were obtained. In particular, the algorithm ran 320 times faster per FPGA on Maxwell than when compared to the equivalent algorithm running on the host PC. This demonstration of the Black Scholes algorithm has shown the potential benefits of FPGAs to the financial sector and the Alliance is currently pursuing opportunities with several leading investment banks, some of whom have been conducting their own experiments with FPGAs. Dr Mark Parsons, Commercial Director of EPCC said: “Maxwell has been created so that businesses can easily investigate FPGAs. We’ve already seen it give companies a competitive advantage. We want more businesses to come and test their codes on Maxwell to see whether it will be useful for them too.” Maxwell was built by the FHPCA (FPGA High Performance Computing Alliance). The Alliance is led by EPCC at the University of Edinburgh and comprises Alpha Data, Nallatech, Xilinx, Algotronix, Scottish Enterprise and the iSLI. The BCS IT Industry Awards are the leading hallmark of success among practitioners in the IT industry today. Anna Duckworth, BCS Head of Corporate Marketing, said: “Year on year the standard of entries increases and this year is no exception with innovative projects that showcase excellence in IT. The next stage of judging will be very difficult as our expert panel endeavours to select the outright winners for each award.” The winners will be announced on Thursday 6th December.