ACADEMIA

As Earth Day 2005 approaches, more scientists than ever are working to understand our environment and mankind's impact on it. And today, Silicon Graphics announced that scientists at three respected institutions have joined the ranks of researchers worldwide who rely on SGI supercomputing, visualization and storage solutions to study and preserve the planet's resources and ecosystems. Most recently, the Massachusetts Institute of Technology (MIT), the University of Massachusetts-Dartmouth (UMass Dartmouth), and the China Meteorology Administration's Institute of Desert Meteorology Urumqi recently acquired powerful SGI supercomputing solutions to drive new discoveries in a range of environmental and atmospheric sciences. They join a long list of SGI customers pioneering the study of global warming and the impact of environmental changes on the Earth. These include Dalhousie University, Israel's Jacob Bluestein Institute for Desert Research, NASA Ames Research Center, the University of Minnesota Supercomputing Institute, Oak Ridge National Laboratory, and the Fleet Numerical Meteorology and Oceanography Center. At MIT's Department of Earth, Atmospheric and Planetary Sciences, researcher Chris Hill is working with fellow scientists at MIT and elsewhere to deploy advanced computing technology for studying phenomena including ocean circulation, weather systems, climate dynamics and seismic and mantle processes. "Studying the Earth system is important to society, but it also presents fascinating problems in fluid dynamics, physics, chemistry, geology, hydrology and computer science," said Hill, whose modeling work in MIT's Program of Atmospheres, Oceans and Science incorporates a 16-processor SGI Altix server with 80GB of memory, which was installed at MIT in the summer of 2004. Hill and his team also make extensive use of the 10,240-processor Columbia Altix supercomputer at NASA Ames Research Center. "To understand the practical problems the world faces, we first need to comprehend the underlying, purely scientific problems," Hill said. "Powerful systems such as SGI Altix allow us to accelerate our ability to apply basic science to these practical problems. By handling larger problems all at once, we can get to our results sooner." In particular, Hill said, Altix servers are "powerful tools for formally combining simulations and observations to produce giant brain-scans of the whole plant in unprecedented detail." In western China, researchers at the Institute of Desert Meteorology, Urumqi, part of the China Meteorology Administration, are turning to high-performance SGI technology to develop new insights in ecological regeneration, sustainable use of natural resources, and the image of desert areas on global climate change. SGI servers, visualization systems, and storage area network solutions were installed in March to give the institute's researchers dramatic new capabilities in their search for answers to climate changes and weather patterns in western China and beyond. By studying certain aspects of desert climates, institute researchers may discover new truths about global warming and even gases and glacial-snow meteorology. Achieving those insights requires numeric analysis and modeling of enormous amounts of weather data. This makes SGI's high-performance computing, visualization and storage technologies a vital part of the institute's work. For researchers at UMass Dartmouth's School for Marine Science and Technology in New Bedford, Mass., climate changes in the western North Atlantic Ocean may hold the answers to how oceanic circulation and biology impact the carbon cycle, the intricate web of interactions between the biosphere, atmosphere, oceans and geosphere -- all of them linked by the element that is the basis of all life on Earth. Until recently, biological data collected intermittently on ships lacked the duration and geographic breadth to provide a meaningful context for broad-based studies in this area. But by comparing circulation models with two new sources of data -- satellite-based observations and location-specific data collected within waters located off the coast of New England -- scientists have a chance to generate new and useful simulations. An eight-processor SGI Altix system, installed during June 2004, serves as the compute engine for the three-year project. "Our existing resources weren't adequate to produce the high-resolution regional modeling effort we had in mind," said Jim Bisagni, principal investigator on the project and associate professor at UMass Dartmouth. "This project involves assimilating and analyzing huge amounts of data from multiple sources. Only state-of-the-art technology like the kind provided by SGI could deliver the performance and architecture we needed to support our combined data analysis and modeling approach." SGI computing, visualization and storage solutions are particularly well suited to scientific applications, due in large part to SGI's third-generation NUMAflex architecture. This unique global shared-memory architecture enables researchers to hold entire data sets in memory, allowing for faster and more interactive data analysis, and resulting in more incisive conclusions.