This account is pending registration confirmation. Please click on the link within the confirmation email previously sent you to complete registration. Need a new registration confirmation email? Click here
FREMONT, Calif., May 23, 2013 (GLOBE NEWSWIRE) -- SGI (Nasdaq:SGI), the trusted leader in technical computing, today announced that NASA's Ames Research Center has selected an SGI® UV™ 2000 shared memory system to support more than a thousand active users around the country who are doing research for earth, space and aeronautics missions.
Installed early this year at the NASA Advanced Supercomputing (NAS) facility at Ames, Moffett Field, Calif., Endeavour is a shared-memory system that took the place of the Columbia supercomputer. Named in honor of the Space Shuttle Endeavour, the last orbiter built during NASA's Space Shuttle Program, this new system is based on the latest Intel® Xeon® processor E5-4600 product family. This processing power, combined in a large, shared-memory cluster, allows Endeavour to provide more high-end computing resources for users while occupying just 10 percent of the previous Columbia system's floor space. Endeavour will provide large, shared memory capability and will enable solutions for many NASA science and engineering applications, including simulation and modeling of global ocean circulation, galaxy and planet formation, and aerodynamic design for air and space vehicles.
"A portion of our current code base requires either large memory within a node or utilizes Open MP as the communication software between tens to hundreds of processors," said William Thigpen, high-end computing project manager at the NAS facility. "The largest portion of Endeavour is able to meet the large shared memory requirement with 4 terabytes of addressable memory and can apply over 1,000 cores against an Open MP application."
The new Endeavour system includes a total of 1536 cores and 6TB of global shared memory. NASA Ames has an existing community of users who could not easily transition to MPI programming models, and the previous system needed to be replaced by a new platform to support this community. Today, user productivity has improved, and the machines are busy.