Forgot password
Enter the email address you used when you joined and we'll send you instructions to reset your password.
If you used Apple or Google to create your account, this process will create a password for your existing account.
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Reset password instructions sent. If you have an account with us, you will receive an email within a few minutes.
Something went wrong. Try again or contact support if the problem persists.

PS3 Supercomputer Opened to University Reseachers

This article is over 13 years old and may contain outdated information
image

The U.S. Air Force has opened its PlayStation 3-based supercomputer for use in university research projects including the development of “artificial neural networks,” which sounds suspiciously like a fancy term for Skynet.

The USAF finished work on the Condor Cluster, its PS3-based supercomputer, back in December 2010. Ranked among the 40 fastest computers in the world, Condor is made up of 1716 PlayStation 3 consoles, 168 general-purpose GPUs and 78 compute servers powered by 2.67 GHz Intel Xenon processors. It was built for roughly one-tenth the cost of a traditional supercomputer and yet uses about one-tenth of the power as well, making it a truly remarkable machine.

Initially used exclusively by the Air Force to analyze spy plane images and other data, the machine has now been made available to various universities for other research purposes. Among them is the University of Dayton in Ohio, which is currently working on the creation of artificial neural networks. The project is focused primarily on two algorithms, one using traditional neural networks and the other using a more “biological” approach meant to model synapses in the human brain. That’s something the PlayStation 3 is particularly well-suited to handle, according to Tarek Taha, an associate professor of electrical and computer engineering at the university. “The PS3’s Cell processor handles a lot of parallelism, like neurons in a brain,” Taha said.

And where might this research ultimately lead? Taha said the technology could one day be incorporated into unammed aerial vehicles, allowing them to carry their own powerful, low-energy supercomputers rather than having to cope with bandwidth limitations to and from the aircraft. The unsaid but obvious implication is that this would allow UAVs to recognize, identify and engage targets without the need for human intervention, and of course what could possibly go wrong with that?

While we wait for that nightmare scenario to unfold, University of Massachusetts Assistant Professor of Physics Gaurav Khanna, one of the few researchers who has already been given access to the Condor Cluster, is assisting the Air Force Research Laboratory with the development of benchmarking software that it hopes will verify once and for all the Condor’s claim as one of the world’s fastest supercomputers. It’s not expected to take back the number-one spot claimed in November 2010 by China’s Tianhe-1A, but its innovative use of consumer electronics may prove to be “a catalyst in the world of high-performance computing.”

Khanna, who has used roughly 300 PS3s in the cluster in his research on gravitational waves, also praised the Sony-based system for its impressive durability. “They’ve been running almost continuously for four years now and it’s a non-ideal environment. It’s a lab, there are students,” he said, adding that none of the machines in the cluster have needed replacing.

The AFRL says it will put out a paper outline the results of its Condor Cluster research in June.

Source: GovTech.com

Recommended Videos

The Escapist is supported by our audience. When you purchase through links on our site, we may earn a small affiliate commission. Learn more about our Affiliate Policy