SGI
SGI competes with IBM, HP, Sun Microsystems and Cray in the growing market has sold the U.S. Department of Defense a supercomputer
to help the agency simulate aircraft, weapon systems and battlefield
scenarios more accurately than ever before.
The DoD’s U.S. General Services Administration and Federal Technology
Service group installed the machine at the Aeronautical Systems Center (ASC)
Major Shared Resource Center (MSRC) at Wright-Patterson Air Force Base in
Ohio, SGI said in statement.
Engineers at the base will use the supercomputer, called Eagle, to improve
weapon systems design and speed modification programs. Eagle
will also be used to upgrade the quality of weapons simulations.
Compared to many supercomputers on the Top500 list of the world’s most
powerful computing machines, Eagle runs Linux at a modest 11.63 teraflops
This speed comes courtesy of 2,048 1.6 GHz Intel Itanium 2 processors, along
with two terabytes
processor.
Leaders on the Top500 list include IBM’s BlueGene/L, which runs at
70.72 teraflops and another SGI Altix, which runs at 51.87 teraflops for
NASA. Were the Top500 list updated today, Eagle would rank ninth behind an
IBM BlueGene/L that runs in the company’s Rochester, Minn. lab.
Eagle is the DoD’s most powerful computer to date, building on the ASC
MSRC’s existing SGI supercomputing power, which now employs more than 4,100
processors from five separate SGI systems.
for machines that can scale to massively parallel computations. Such
computers can often throw inordinate amounts of computational resources to
address crucial challenges.
Like products from its rivals, SGI machines have been used in several
important areas. Supercomputers are used in the medical field to help brain
surgery, as well as to find oil more efficiently, study global climate and
shore up homeland security and national defense.
SGI, based in Mountain View, Calif., said it believes it has an advantage in helping the government’s Linux computing gear expand without tradeoffs between scaling up and scaling out.
This is because the Altix architecture allows clustering, or scaling out, and the addition of more nodes of processors, commonly referred to as scaling up. Each node scales from four to 512 processors, sharing up to 4TB of memory.
SGI Has ‘Eagle’ Eye for DoD’s Defense
