Energy and IBM construct a scientific computing grid

Energy and IBM construct a scientific computing grid

An Energy Department laboratory in Berkeley, Calif., is joining with IBM Corp. to develop a high-performance computing grid for research in global climate change and other areas that use large data sets.

The Science Grid will link the IBM SP supercomputer at the National Energy Research Scientific Computing Center in a WAN with other Energy labs and huge storage capacity, said Peter Ungaro, IBM's vice president of high-performance computing. NERSC's 3,328-processor system, named Seaborg, is the world's third-fastest computer and the fastest that is unclassified.

The partnership benefits both Energy and the company, NERSC director Horst Simon said. By leveraging the investment IBM has already made in grid infrastructure, Energy can have the initial version of the Science Grid up and running by the end of this year. IBM, which has commercial plans for grid computing, will gain experience with open-source grid middleware from a group called the Globus Project, based at Energy's Argonne National Laboratory in Illinois.

A key component of the Science Grid will be the High-Performance Storage System, a massive archival system with a capacity of 1.3 petabytes (1,300T), Ungaro said. The HPSS storage is now at NERSC, but the center is integrating Oak Ridge National Laboratory's HPSS subsystems into a WAN.

Eventually researchers can connect scientific instruments such as radio telescopes and subatomic particle accelerators to the grid and access large data sets without worrying where they are physically stored, Ungaro said.

inside gcn

  • artificial intelligence (ktsdesign/

    Machine learning with limited data

Reader Comments

Please post your comments here. Comments are moderated, so they may not appear immediately after submitting. We will not post comments that we consider abusive or off-topic.

Please type the letters/numbers you see above

More from 1105 Public Sector Media Group