Skip Links

Network World

  • Social Web 
  • Email 
  • Close

(Comma separation for multiple addresses)
Your Message:

CSIRO beefs up research with GPU cluster

By Kathryn Edwards, Computerworld Australia
November 25, 2009 09:00 AM ET
  • Share/Email
  • Tweet This
  • Comment
  • Print

A supercomputer which combines the central processing units (CPUs) found in traditional PCs, with more powerful graphics processing units (GPUs) will be launched by the Commonwealth Scientific and Industrial Research Organisation (CSIRO) in Canberra today.

The first of its kind in Australia, the GPU cluster contains 61,440 compute cores and will help scientists to explore the use of GPU technology for parallel processing.

CSIRO Computational and Simulation Science leader, Dr John Taylor, said the Nvidia-based GPU technology gives the organisation the potential to deliver huge gains in computational efficiency, energy efficiency and scientific research.

In pictures: See the slideshow of the CSIRO's new CPU cluster

"CSIRO is investing heavily in what we can currently use for production computing, but the GPU cluster is investing in what we see as one of the next generation supercomputing resources," Taylor said.

Under the hood:

* 128 Intel Dual Xeon E5462 compute nodes (a total of 1024 2.8GHz compute cores) with 16GB or 32GB of RAM, 500GB SATA storage and DDR InfiniBand interconnect

* 64 Tesla S1070 (200 GPUs with a total of 48 000 streaming processor cores)

* 128 port DDR InfiniBand Switch expandable to 144 ports

* 80 Terabyte Hitachi NAS file system

The largest of its kind in Australia, Taylor said the use of GPUs for scientific research is becoming popular around the world as researchers look for ways to increase the speed of their applications.

A major barrier to be overcome, however, is moving the CSIRO scientists' code over onto the new GPU cluster, and having the approach to coding changed.

"The key issue at the moment is you can't take an application that's running on a multi-CPU core system and run it on the GPU," Taylor said. "You've got to change the code, which is very typical."

Not only will the scientists benefit from faster data processing, the GPU cluster also heralds major environmental benefits. It is more energy efficient and less expensive to run than a CPU computer per unit of processing power.

The GPU cluster will complement the recently launched NCI facility at the Australian National University.

Sign up for Computerworld's newsletters to stay up to date.

Got more on this story? Email Computerworld or follow @computerworldau on Twitter and let us know.

  • Share/Email
  • Tweet This
  • Comment
  • Print

Partner Content

Gartner 2009 Magic Quadrant for Job Scheduling

Gartner has positioned BMC CONTROL-M in the Leaders Quadrant of their "2009 Magic Quadrant for Job Scheduling." The report assesses the ability to execute and completeness of vision of key vendors in the marketplace. Read a full copy today, courtesy of BMC Software.

Download whitepaper

Dell's SMART Approach to Workload Automation

Read a compelling case study by EMA, Inc. to learn how Dell uses BMC CONTROL-M to cut cost and increase productivity with workload automation.

Download whitepaper

Workload Automation Cost Savings 2 Minute Video

A major computer manufacturer uses BMC CONTROL-M and just four people to schedule and run over 85,000 jobs every month. By switching to BMC CONTROL-M, they more than quadrupled the workload without adding a single staff member.  See how in this 2-minute video overview.

Go to video

Comment
Login
Forgot your account info?
Add comment
Anonymous comments subject to approval. Register here for member benefits.
Have a NetworkWorld account? Log in here. Register now for a free account.

Videos

rssRss Feed