Posted on 08/15/2003 6:43:45 AM PDT by bedolido
BLUFFDALE -- Linux Networx Inc., already the proud parent of the world's third-fastest supercomputer, on Thursday boasted it is building an even more powerful system for the Los Alamos National Laboratory.
The $10 million contract calls for delivery next month to the New Mexico lab's Metropolis Center for Modeling and Simulation of what will be one of the largest Linux-based clusters -- a supercomputer melding 2,816 AMD Opteron processors -- and among the fastest Linux Intel-based supercomputers in the world.
Dubbed "Lightning," Los Alamos' newest mega-brain will come with a theoretical performance of 11.26 teraflops (Tflops), or 11.26 trillion calculations per second. The lab plans to use the supercomputer to simulate nuclear explosions in the absence of actual underground testing of the nation's atomic weaponry.
Chief Executive Bernard Daines called Lightning's speed milestone "impressive," but insisted Linux Networx beat out big-time competitors such as IBM by building a reputation for "peak performance" and "system uptime" in its creations.
Still, in an interview, Daines couldn't help bragging a little about the new kid on the supercomputer block.
"It isn't just the big guys who can do these things," he said. "Los Alamos chose us because of our track record. . . . Here's this little, 100-person company in Bluffdale that got this big contract because we developed the technology to do this right here in Utah."People need to know that, they need to be supportive of that, so companies like us can continue to bring business here," Daines added.
John Morrison, leader of Los Alamos' Computing Communications and Networking Division, said in a statement that the new supercomputer "offers a cost-competitive way to meet our growing need to run large, important calculations and get results in a few days.
"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," Morrison added.
Compared with Lightning, the $6 million MCR Linux Cluster system the company made for Lawrence Livermore National Laboratory in 2002 -- currently ranked the globe's third-fastest on the Top 500 Supercomputers Sites list -- processes data at a 7.6 Tflop clip.
However, Lightning not only will supplant its slower cousin at No. 3, but give Linux Networx two of the top five systems; MCR moves, at least temporarily in a competitive industry, to No. 4, while IBM's currently third- and fourth-ranked 7.3 Tflop supercomputers -- ASCI White at Livermore and Seaborg, at Berkeley, Calif. -- slide down the scale.
Still atop the supercomputing heap are NEC's 35.86 Tflop Earth Simulator in Yokohama, Japan, and another Los Alamos system, the 13.88 Tflop Asci Q made by Hewlett-Packard.
For Daines that is fine, for now. After all, in conjunction with Thursday's Lightning announcement, Linux Networx also said it will provide Los Alamos with another system, a 256-node cluster brain christened "Orange."
Terms of that deal were not disclosed, but Daines is particularly proud of Orange, which will become part of Los Alamos' Institutional Computing Project -- an endeavor supporting cutting-edge AIDS and antibiotic research.
"The kinds of things these labs will use these computers for will impact everyone," Daines said. "Medical research, environmental issues and nuclear safety are all areas they have the opportunities to create a better life for all of us."
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.