High performance computing environments are a point of interest for several chip makers, and Nvidia in particular has been capitalizing on the huge potential of GPUs in such environments.   These chipsets are compromised of hundreds of smaller cores rather than just a few of the more standard kind, and this proved handy not only for processing advanced graphics but also in carrying out calculations on a massive scale.

There’s only one catch: porting applications from the more common-place x86 environments to infrastructure powered by these GPUs is a lengthy process, and actual implementation also remains fairly complex. This is where Intel has spotted a big opportunity.

The new Xeon Phi product family has been debuted today with the premise of a superior approach to achieving the same result. These microprocessors were designed to supplement existing Xeons, and they run their own OS. This allows them to run independently, according to Intel marketing chief John Hengeveld, who noted that GPUs cannot. The product also adds another layer of abstraction thanks to the fact software migration is reduced to a far lesser chore, says one expert who had a chance to examine the Phi.

A key selling point is the claim that existing software can be made to run on a Xeon Phi with little modification. One person who has witnessed that feature first-hand is Robert Harrison, director of the joint institute for computational sciences, which is affiliated with Oak Ridge National Laboratory and the University of Tennessee.

Intel will be showcasing exactly what its new technology can accomplish at the upcoming International Supercomputing Conference. The company assembled a system that ranks as the 150^th most powerful supercomputer in the world for demonstrating purposes.

This is not the only space that Intel is eyeing. It’s also actively working on growing its presence in mobile, another huge area of potential growth and increasing competition for the chip maker.