Powered by SGI, GMU School of Computational Sciences Builds High- Performance Computing Center to Greatly Improve Time to Solution

MOUNTAIN VIEW, Calif. (October 10, 2005)—To run highly complex calculations in areas as diverse as bioagent dispersion, climate modeling and prediction, and hemodynamics (a branch of physiology that deals with the circulation of the blood), George Mason University's School of Computational Sciences has selected scalable computing technology and storage solutions from Silicon Graphics (NYSE: SGI) as the foundation of a three-year initiative to build a high-performance computing center. As previously announced in the customer highlights for the fourth quarter, the driving factor behind the SGI® Altix® and SGI® InfiniteStorage purchase is GMU's goal to teach and empower research as well as enhance their wide-spread collegiate collaboration with other leading universities and institutions and government laboratories, many of which have SGI Altix systems.

"Having an Altix system will allow us to get our young professors and our new ideas and innovation moving in exciting new directions," said Dr. Paul Schopf, assistant dean for research of the School of Computational Sciences and professor of oceanography, George Mason University. "We are a school of computational sciences, which means that we want answers to scientific questions. Time to solution from the concept, to the coding, to data management and the delivery of results, is our key metric. The SGI Altix and InfiniteStorage speeds up everything."

Headed by Dr. Menas Kafatos, Dean, the School of Computational Sciences collaborates with leading institutions including Lawrence Livermore National Laboratory, the National Institute of Standards and Technology, Los Alamos National Laboratory, the Naval Research Lab, and a number of Department of Defense labs. Faculty also use compute resources at NASA Ames' Project Columbia, a 10,240-processor SGI Altix system and NASA Goddard, which has a 1,240-processor SGI Altix with a 440TB SAN solution as well. All Altix systems run the Linux® operating environment on Intel® Itanium® 2 processors.

GMU Research Highlights
The new and ongoing research the SGI Altix will be running at GMU's School of Computational Sciences includes these, and many other areas.

• Bioagent Dispersion Simulation. To study contaminant transport requires intense grid calculations. Simulations are made of midtown/downtown areas to predict where a potential cloud of poisonous material goes in the first 10 minutes after release. The simulations are done to answer vital first responder questions and improve community preparedness. The calculations involve almost infinite variables: What are the actual physics of each environment? What is the difference in the physics of bioagent dispersion in a street canyon environment like New York as compared to an open city like Salt Lake City? Weather patterns and temperatures at different times of year, wind directions, turbulent content of the flow, whether the walls have been heated because it was a sunny day, what is the effect of air conditioning-all need to be mathematically calculated with different variables and different possible contaminants.

• Climate Dynamics and Climate Modeling. The faculty collaborates with NASA, National Oceanic and Atmospheric Administration (NOAA), and NSF on climate modeling and prediction, such as El Niño forecasting, and seasonal to decadal climate forecasting. The School is heavily involved in multi-model ensembling for climate forecasting and they will use the SGI Altix to pre-stage calculations on a variety of popular modeling programs. GMU faculty associated with the Center for Ocean-Land-Atmosphere Studies (COLA) have just recently been awarded a large grant from NASA's Modeling And Prediction (MAP) program, in which the climate dynamics faculty, using the SGI Altix, will collaborate with NASA Goddard on developing the new GEOS5 climate model.

• Hemodynamics. In an unusual application of Computational Fluid Dynamics (CFD), a patient's radiological image data (CAT scans, MRIs, angio, etc.) is used to understand the flow patterns in veins and arteries using fluid dynamics codes. These simulations serve as a basis for pre- and post-operative planning as well as a fundamental research environment to understand the clinical relevance of flow patterns in the pathogenesis of arterial diseases. This will allow surgeons to simulate clamps or stents or other surgical procedures and to check and evaluate blood flow and the supply of oxygen to the organ, other organs, or the entire body. Simulation can alert the surgeon to possible problems, and pre-planning can reduce operating time, which in turn reduces patient trauma and speeds recovery.

• Molecular Dynamics. The assembly of Alzheimer's polypeptides into fibrils is being simulated in the lab. Understanding how small proteins gather into clusters (fibrils) associated with diseases may lead to discoveries of how to slow or halt the formations.

"We did an extensive comparison, a bidding process, and we found that the Altix had the flexibility to allow us to run all of the applications for all of our different disciplines," added Dr. Schopf. "And, we are finding that the ability to share codes with our colleagues, codes that may have come from a shared memory configuration, a message-passing paradigm, or whatever, all the codes are quite easy to bring to the Altix. And that's a big value to us and to our collaborative partners."

Other disciplines that will be using the SGI Altix system include: Hydrodynamics, focusing on designing naval ships and oil rigs to withstand highly turbulent seas; Cellular Modeling, to improve the understanding of heart disease; and Complex Explosion Analysis, to explore air flows and materials impacts in analysis of terrorist scenarios. Additional areas include space sciences, computer design of materials and fluids, including nanotechnology, and hazardous release modeling.

SGI Altix and InfiniteStorage Scalability is Key to GMU's HPC Center
"The important thing about the Altix that distinguishes it from other machines is the fact that it's a shared-memory machine," said Dr. Rainald Löhner, professor of computational sciences and informatics, and dean of fluids and materials research, School of Computational Sciences, George Mason University. "In my group, we do very complex multidisciplinary problems, such as fluids combined with structures, and it becomes very difficult to do them on a distributed memory machine. It was also important to have Linux because we all use Linux in the academic community at large."

GMU purchased a SGI® Altix® 3700 Bx2 supercomputer, powered by 64 Intel Itanium 2 processors with 128GB of memory and 2.3TB of disk storage, supported by a 10TB SGI® InfiniteStorage TP9300 with multiple 400GB Serial ATA drives for direct-attached storage.

"With research areas ranging from individual patient diagnostics to homeland security preparedness, the breadth of applications for which George Mason University uses their SGI Altix system is truly staggering, and is only possible because of SGI's shared-memory architecture. The ability of the Altix system to perform optimally with many different programming paradigms enables many projects to share the Altix's computational power; and to focus the entire capability of the system on a single application when it is being staged for larger runs on supercomputers such as NASA's Project Columbia," said Dave Parry, senior vice president and general manager, Server and Platform Group, SGI. "The goal of any university is to provide the best teaching tools possible to enable students and research faculty to excel. We are pleased that GMU has chosen SGI technology as one of those tools in its quest to continually elevate GMU and the School of Computational Sciences to new levels of multi-disciplinary learning and collaboration."

As part of the three-year initiative, Research 1, a new research building for the School of Computational Sciences at GMU is currently under construction. The new 64-processor SGI Altix and 10TB SGI InfiniteStorage TP9300, purchased through James River Technical, Inc., a value-added SGI Channel Partner, will be reinstalled at the new building. Also in the near future, GMU will join the National LambdaRail project, a high-speed fiber network sometimes referred to as Internet 3 and exclusive to participating research institutions and universities. The grand opening of Research 1 is expected next summer.

About George Mason University
George Mason University is an innovative, entrepreneurial institution with national distinction in a range of academic fields. Enrollment now tops 28,000, with students in 144 degree programs at campuses in Arlington, Fairfax and Prince William, Virginia. In fall 2006, George Mason opens a campus in Ras-Al-Khaimah (RAK), located in the United Arab Emirates (UAE). George Mason University can be found on the Web at www.gmu.edu.

SILICON GRAPHICS | The Source of Innovation and Discovery™
SGI, also known as Silicon Graphics, Inc. (NYSE: SGI), is a leader in high-performance computing, visualization and storage. SGI's vision is to provide technology that enables the most significant scientific and creative breakthroughs of the 21st century. Whether it's sharing images to aid in brain surgery, finding oil more efficiently, studying global climate, providing technologies for homeland security and defense or enabling the transition from analog to digital broadcasting, SGI is dedicated to addressing the next class of challenges for scientific, engineering and creative users. With offices worldwide, the company is headquartered in Mountain View, Calif., and can be found on the Web at www.sgi.com.

Silicon Graphics, SGI, Altix, the SGI cube and the SGI logo are registered trademarks and The Source of Innovation and Discovery is a trademark of Silicon Graphics, Inc., in the United States and/or other countries worldwide. Intel and Itanium are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. Linux is a registered trademark of Linus Torvalds in several countries. All other trademarks mentioned herein are the property of their respective owners.