PHYSICS

Powerful Linux OS-Based System Serves as Core Research Asset For Interdisciplinary Team of Computational Biologists -- Aided by high-performance computing technology from Silicon Graphics, Memorial Sloan-Kettering Cancer Center's (MSKCC) Computational Biology Center is using a 12-processor SGI Altix 3000 system to accelerate its computational disease research, including the modeling and simulation of genes involved in cancer pathways. Installed in September 2003, the SGI Altix system combines a 64-bit Linux operating environment, Intel Itanium 2 processors and a high-bandwidth system architecture to provide MSKCC researchers with a powerful and flexible platform for some of the world's most advanced computational biology research. The MSKCC team analyzes and simulates biological processes at different levels of organization, predicts the results of interventions in biological systems, and aims to improve the prevention, diagnosis, prognosis and therapy of cancer. "We have had excellent professional interactions with SGI over the years and are now using the new Altix system to speed computational biology research using human genome information as a foundation for the emerging field of systems biology," said Chris Sander, head of the Computational Biology Center (CBC) at MSKCC. "Our focus is to provide bridges between basic and clinical research, and computational technology of this type is a key enabling factor. We will be successful if the computational predictions are verified in the laboratory and become the basis for the design of highly efficient clinical trials for the development of new cancer therapies." The SGI Altix system serves as a core research machine within the center's Linux OS computing environment. The MSKCC team is using the Altix system in a "dry laboratory" that complements and strengthens traditional laboratory and clinical research. Researchers simulate what goes on in cells at the molecular level, and derive useful predictions from those simulations for disease states. "The computational biology team at MSKCC stretches the boundaries of their science, seeking to obtain essential medical knowledge from the human genome," said Dave Parry, senior vice president and general manager, Server and Platform Group, SGI. "We are proud to see MSKCC using SGI Altix 3000 as a core tool in this effort to advance the global fight against cancer." With 48GB of system memory and the SGI® NUMAflex™ shared-memory architecture, the Altix system makes it easy for MSKCC to achieve multiple types of parallelism needed for the sophisticated analysis of gene expression profiles and simulation of cellular control networks. "The Altix shared-memory architecture allows us to spread a computationally intensive task across many processors," said Dr. Anton Enright, biocomputing manager at MSKCC's Computational Biology Center. "We need to be able to provide new and existing laboratories with computational resources that can handle a variety of problems as efficiently as possible. Because the SGI Altix system runs Linux, it is relatively straightforward to port existing applications to Altix, and to interconnect the system with other machines and services within the center." Enabling Science and Research Worldwide Since their introduction in January 2003, SGI Altix servers and supercomputers have grown increasingly vital to scientists and researchers around the world. From cosmological studies at the University of Cambridge and atmospheric research at Harvard University's Department of Earth and Planetary Sciences to advanced ocean simulations at NASA Ames Research Center, some of the world's most important research projects depend on SGI Altix for its ability to handle large data sets without the I/O bottlenecks suffered by traditional PC clusters. Scalable SGI Altix 3000 systems are available today in server configurations of 4 to 256 processors, and supercluster configurations of 4 to 512 processors. For customers demanding even larger Altix configurations, SGI plans to support single Altix nodes of 512 processors and superclusters of 1,024 processors and larger over time.