Top supercomputer race continues with Nasa/SGI box

Engineers at Nasa's Ames Research Center have staked their claim have built the world's fastest supercomputer.

Engineers at Nasa's Ames Research Center have staked their claim have built the world's fastest supercomputer.

At a ribbon-cutting ceremony at Nasa's supercomputing facility, the US space agency unveiled its 10,240 processor Linux-based supercomputer, named Columbia, and cited benchmark numbers that show it performing faster than any other system.

Built in cooperation with Silicon Graphics and Intel over the last four months, Columbia has achieved a sustained performance of 42.7 trillion calculations, or teraflops, as measured by the Linpack benchmark used to rank the bi-annual Top 500 list of supercomputers.

This result puts it ahead of both Japan's Earth Simulator, which was benchmarked at 35.86 teraflops in June, and IBM's Blue Gene system, which was benchmarked at 36.01 teraflops in September.

Nasa's 42.7 teraflop benchmark was achieved using only 16 of the 20 512-processor Altix systems that SGI built for the system, officials said. A benchmark number based on Columbia's full performance will be released in early November.

The system is already being used for a variety of tasks at Nasa, including hurricane prediction and analysing data from the more than 60 space probes that the agency currently operates throughout the solar system.

Columbia, which is able to simulate phenomena to a detail unmatched by Nasa's current systems, will also play a critical role in performing the simulations required to resume space shuttle missions, Nasa officials said.

"All of the simulations for the redesigned shuttles are being done on these systems," said G Scott Hubbard, director at Nasa Ames.

Robert McMillan writes for IDG News Service

Read more on Server hardware

Start the conversation

Send me notifications when other members comment.

By submitting you agree to receive email from TechTarget and its partners. If you reside outside of the United States, you consent to having your personal data transferred to and processed in the United States. Privacy

Please create a username to comment.