Cray Plans “Free-Cooling” Supercomputer for NERSC

New Cray energy-efficient supercomputer and next-generation storage system to be installed at NERSC.

Cray Inc. has announced that it will install a new energy efficient supercomputer named “Cascade” and a next-generation Cray Sonexion storage system at the Department of Energy’s National Energy Research Scientific Computing Center (NERSC).

According to the press release, the system will include the ability to run year-round using “free-cooling” at the NERSC site. “This approach utilizes water from cooling towers only, not mechanical chillers, to provide exceptional energy efficiency. The moderate Bay Area climate combined with Cray’s new design will allow us to keep power for cooling to less than 10 percent of the power used for computing,” said Jeff Broughton, head of NERSC’s System’s Department.

The Cascade supercomputer, which is expected to be available in the first half of 2013, will provide over two petaflops of peak performance for NERSC. The system will feature major advancements of the Cray Linux Environment, Cray’s HPC-optimized programming environment, and the next-generation Aries interconnect chipset. The Aries interconnect is integrated with PCI Express, an on-board bus that will support multiple processor types.

Additionally, the Cray Sonexion storage system will scale to more than six petabytes of usable storage and more 140 gigabytes per second of sustained aggregate IO performance. According to the announcement, “Sonexion brings together an integrated file system, software and storage offering that has been designed specifically for a wide range of HPC workloads, providing users with an integrated, scalable Lustre solution that is easy to install and maintain.”

The Cascade system will provide supercomputing resources to NERSC users working to advance open science research in climate modeling, biology, environmental sciences, combustion, materials science, chemistry, geosciences, fusion energy, astrophysics, nuclear and high-energy physics, and other disciplines, along with scientific visualization of massive data sets. The system is expected to go into production in 2013. You can learn more at: