Build linux computer

Los Alamos National Laboratory has selected Linux Networx to design, integrate and deliver one of the largest Linux clusters ever built, with a theoretical peak of 11.26 teraops.

Called Lightning, the computer supports the Advanced Simulation and Computing program, or ASCI, which helps ensure the safety and reliability of the nation's nuclear weapons stockpile in the absence of underground testing.

When delivered to the Metropolis Center for Modeling and Simulation at Los Alamos in September, the cluster is expected to rank among the most powerful supercomputers in the world. The system will also have a delivery time of less than two months from when the contract is signed to when the system will be delivered to the lab.

"While the 11 teraops peak performance of this system is impressive, Linux Networx places a higher importance on the real peak performance and the system uptime, so we deliver the most productive computer possible to our customers," said Bernard Daines, CEO of Linux Networx. "We feel confident the delivery of this system will help Linux Networx maintain a leadership position in the Linux cluster market."

ASCI supercomputers run three-dimensional codes that simulate all the physics involved in a nuclear detonation, allowing researchers to integrate past weapons test data, materials studies and current experiments in simulations of unprecedented size. ASCI is a collaboration among Los Alamos and its sister national laboratories--Lawrence Livermore and Sandia--and the Defense Programs office of the National Nuclear Security Administration that creates the leading-edge computational modeling and simulation capabilities essential for maintaining the safety, reliability and performance of the US nuclear stockpile and reducing the nuclear danger.

Los Alamos will use Lightning principally for smaller, more numerous computing jobs in the Stockpile Stewardship workload such as weapons code development, verification and validation.

"The Lightning system that Linux Networx is designing offers a cost-competitive way to meet our growing need to run large, important calculations and get results in a few days," said John Morrison, leader of Los Alamos' Computing Communications and Networking Division. "A system of this magnitude will provide a valuable proving ground for large-scale, practical cluster computing, building on the exciting development of open source tools by the larger high-performance computing community."

Lightning includes 2,816 AMD Opteron(TM) processors, making it the largest AMD Opteron processor-based system delivered in 2003 and the first 64-bit Linux supercomputer in the ASCI program. Linux Networx won the initial contract for 1,280 dual-processor nodes, and Los Alamos has agreed to exercise an option to boost the number of nodes to 1,408. The total cost of the initial contract, the option and another option for system maintenance is just under $10 million.

"Los Alamos Lab's Lightning cluster is a great example of how large-scale systems are benefiting from the AMD Opteron processor," said Marty Seyer, vice president and general manager of the Microprocessor Business Unit at AMD. "By combining industry-standard technology, outstanding 32- and 64-bit performance, AMD and Linux Networx are heightening the impact of supercomputing technology."

COPYRIGHT 2003 Millin Publishing, Inc.
COPYRIGHT 2003 Gale Group