Video game systems are no longer just for play; they are now being used by the U.S. military to build supercomputers.

Recently the U.S. Air Force built The Condor Cluster, a heterogeneous supercomputer made from off-the-shelf components. One of those pieces is the Sony Playstation 3. In total, the Air Force used 1,716 of them to help build the supercomputer.

In a conference call, Mark Barnell, the director of high-performance computing and the Condor Cluster project at the Air Force Research Laboratory, gave three reasons why the Air Force chose Playstations.

First, they were served as a general purpose graphics processor unit, which many general-purpose PCs won't have -- at least not one as powerful.

Secondly, they are environmentally efficient. Because of those GPUs, that technology, along with the Cell BE process in all the PlayStations, are extremely power-efficient processors, Barnell said. The supercomputer is the seventh-greenest in the world according to the Air Force.

The third reason was cost savings. According to Barnell, the cost of the Condor system was $2 million. The most advanced supercomputers cost $100 million, and one comparable to the Condor would be in the $50 million range.  

Even with off-the-shelf components, this supercomputer is no slouch, Barnell says. While it's not designed to compete with the best supercomputers in the world, it's still pretty good and could change the way supercomputers are built, focusing on using more computational resources and less energy.

This particular system is about half a petaflop, or capable of about 500 trillion calculations per second, Barnell said. In the current time that we can measure it, it's about the 35th- or 36th-fastest computer in the world, and with some things that are going to be changing in the next eight or nine months with some upgrades, we could boost it to maybe the 20th-fastest computer in the world, and at the same time make it, at that moment in time, the greenest computer.

The Air Force says the Condor Cluster will be used for neuromorphic computing, which is computational intelligence. Programmers will be able to write algorithims to teach the computer how to read symbols, letters, words and sentences. This will eventually allow the computer to think on its own, the Air Force says. Already, it can read 20 pages of information per second at nearly 100 percent accuracy.