The National Aeronautics and Space Administration (NASA), Intel Corp., and SGI on Thursday announced the signing of an agreement establishing intentions to collaborate on significantly increasing the space agency’s supercomputer performance and capacity. Under the terms of a “Space Act Agreement”, Intel and SGI will build a supercomputer with performance capacity that is two times higher compared to the fastest supercomputer today.

“Achieving such a monumental increase in performance will help fulfill NASA’s increasing need for additional computing capacity and will enable us to provide the computational performance and capacity needed for future missions. This additional computational performance is necessary to help us achieve breakthrough scientific discoveries,” said Ames Director S. Pete Worden.

The new supercomputer will be built in NASA Advanced Supercomputing (NAS) facility at NASA’s Ames Research Center in Moffett Field, California. NASA Ames, Intel and SGI will work together on a project called Pleiades to develop a computational system with a capacity of one Petaflops peak performance (1000 trillion operations per second) by 2009 and a system with a peak performance of 10Petaflops (10 000 trillion operations per second) by 2012. By comparison, today’s highest-performing IBM BlueGene supercomputer in Lawrence Livermore National Laboratory has peak performance of 478 200GFlops, or more than two times lower compared to projected performance of the system for NASA Ames.

This collaboration builds on the 2004 deployment of Columbia, which generated a tenfold increase in supercomputing capacity for the agency. Meeting NASA’s future mission challenges will require additional computational resources to handle increasingly higher fidelity modeling and simulation. In 2009, NASA expects to increase that computing capability 16 times with the Pleiades project, and by an additional tenfold in 2012.