ACADEMIA

A new high-performance computing (HPC) cluster has the potential to revolutionize the creation of computer code for engineering design. The Spitfire Computer Cluster at the University of Southampton, UK running on Microsoft Windows Compute Cluster Server 2003, will enable engineers and academic researchers to more effectively perform complex computations. The project will demonstrate how HPC clusters can support complex engineering calculations such as geometry changes to aircraft features and aeroengine components using computational fluid. It also will enable multiple problems to be linked together in multi-stage and multi-physics simulations. Potential uses of the Spitfire Computer Cluster include research into internal and external air flows around aircraft jet engines, aeroplanes and Formula 1 race cars; water flows around yacht hulls; and the behaviour of biomedical implants in the body. The cluster is named in honour of the legendary Spitfire fighter plane designed by R.J. Mitchell in Southampton in the 1930s. The Spitfire’s engine was built by Rolls-Royce, which maintains strong research links with the University today. Future projects involving the Spitfire Computer Cluster include research in collaboration with Rolls-Royce. The project will comprise of two facilities. The first will focus primarily on applications for key industrial partners, while the second will provide a Windows-based, high-performance computing and data (HPC&D) platform for the broader academic community. Project leaders aim to build a community of users across a range of disciplines for the platform. Participants will explore ways to develop novel techniques for handling data and workflow from experimental tests, as well as develop and prototype new HPC frameworks. Users of the Spitfire Computer Cluster will comprise a range of engineers in the University’s School of Engineering Sciences, including members of the Computational Engineering and Design Group (CEDG). Industry researchers and other academics will also be able to log in to the Cluster over virtual private networks or Web-based interfaces. Professor Andy Keane, Professor of Computational Engineering and Professor Simon Cox, Professor of Computational Methods, are leading the project for the University of Southampton. "This Cluster will help us test out codes on realistic-scale jobs as well as engage in design improvements and research using approaches such as design of experiment methods and optimisation on full-scale problems," said Keane.