CAMBRIDGE UNIVERSITY COSMOS PROJECT LED BY PROFESSOR STEPHEN HAWKING CHOOSES SGI ALTIX 3000 SUPERCLUSTER
New System Allows Extension of Cosmology Grid; Consortium Members Collaborate Using Visual Area Networking Technology
Dubai, UAE May 5, 2003 - SGI has today announced a contract to provide an SGI® Altix™ 3000 supercluster to the U.K. COSMOS consortium led by Professor Stephen Hawking of the University of Cambridge as part of a long-term agreement between the two organisations. The latest addition to the consortium’s SGI® infrastructure, based on the Intel® Itanium® 2 processor, will form the next phase of a British computational and visualisation grid installed by SGI to support the COSMOS project. This initiative enables experts to collaborate on research to model the history of the universe from the first fractions of a second after the Big Bang to the present day, about 14 billion years later.
The Altix 3000 system, powered by128 Itanium 2 processors, will form the core of the U.K. cosmology grid (CosmoGrid), supporting collaboration between consortium members through the continuation of integrated high-performance computing (HPC) and complex data management, as well as remote collaborative visualisation using SGI Visual Area Networking technologies. Members of the COSMOS consortium include the Department of Applied Mathematics and Theoretical Physics, Institute of Astronomy, and Cavendish Laboratory at Cambridge University, as well as Imperial College London, the University of Portsmouth and the University of Sussex.
This 64-bit system, featuring Intel Itanium 2 processors running the Linux® operating system, supports SGI’s close working relationship with Intel, which has also provided two small servers to the consortium to support development work. The decision to purchase the Altix 3000 supercluster was based on the combination of highly scalable Itanium 2 performance, cache-coherent
global shared memory, integrated visualisation and hierarchical storage. The system allows for fast development time scales and high productivity using shared memory programming with OpenMPÔ, while also producing excellent benchmark results for distributed-memory programming codes based on MPI.
The contract is part of the long-term collaboration between Professor Hawking’s COSMOS consortium and SGI; the two entities have worked together since 1997. The original COSMOS platform was a 32-processor SGI® Origin® 2000 server, later upgraded to a 64-processor SGI® Onyx® family server, including the CXFSÔ storage system. The Altix 128-processor system is a further indication of SGI’s long-term commitment to the COSMOS project.
Professor Stephen Hawking, principal investigator of the COSMOS consortium, commented: “The new COSMOS computer¾an SGI Altix 3000¾will enable us to keep up with the dramatic data about our universe which is now coming in. We are pleased that our collaboration with both SGI and Intel will ensure that U.K. scientists remain at the leading edge of cosmological research.”
Paul Shellard, Ph.D., director of the COSMOS supercomputer, added, “The SGI Altix 3000 system with Itanium 2 chips makes available an entirely new threshold in scalable shared-memory performance for the U.K. COSMOS consortium. The rapid time to solution possible using OpenMP shared-memory programming offers a key competitive advantage in this fast advancing field driven by new satellite observations and other experiments. Together with integrated shared filesystems and remote visualisation, the state-of-the-art Altix supercluster will give us the ability to test our mathematical models of the universe seamlessly against the huge data sets being produced by cosmological observations.”
Professor Roy Maartens, head of the cosmology group at the University of Portsmouth, noted,
“With the new Altix system, cosmologists will have extra computing power to extract ever more accurate information from the data about the size and shape and ingredients of the universe.”
“Scientific research is constantly at the forefront of technology, pushing the boundaries of high-performance computing and demanding higher processing power with greater scalability,” said John Woodget, Intel marketing director, EMEA. “The Itanium Processor Family delivers new levels of performance to support this kind of complex, data-intensive operation, in both academic and enterprise environments. By combining the high-throughput functionality of the Itanium 2 architecture with SGI hardware, the COSMOS team has dramatically upgraded the capability of its computer infrastructure. Intel’s own commitment to development ensures that the Itanium processor family will continue to support the needs of projects such as COSMOS well into the future.”
SGI’s Altix 3000 systems, launched in January of this year, combine SGI supercomputing architecture with Intel Itanium 2 processors and the Linux operating system, and since its launch it has shattered scalability and performance records. The new installation at Cambridge University also includes additional graphics equipment and an upgrade to a CXFS storage system featuring 2GB storage area network and 6TB SGI® TP9400.
Steve Coggins, senior VP, EMEA at SGI, commented, “With the help of SGI Managed Services, the COSMOS project is able to exploit the full potential of the combination of visualisation, high-performance computing and data management. We value the continued challenge of meeting the needs of a world-class organisation that is pushing the limits of computational science.”
For further details about SGI’s involvement with universities and research labs around the world, visit www.sgi.com/go/research.
AI
