ORION

Supercomputer to use new Nvidia 'Fermi' chip.

Oak Ridge National Laboratory announced plans today for a new supercomputer that will use Nvidia's next-generation GPU architecture, codenamed "Fermi".

The Oak Ridge and Fermi announcements were made at Nvidia's GPU Technology developer's conference, which kicked off Wednesday in San Jose, Calif. The Fermi chip integrates three billion transistors, about three times the number of transistors in Nvidia's most powerful graphics chip now on the market. In the future, the chip will also find its way into Nvidia's GeForce product line for PCs.

Oak Ridge's supercomputer will be used for research in energy and climate change and is expected to be 10 times more powerful than today's fastest supercomputer, according to a joint statement from Oak Ridge and Nvidia. The architecture would use both graphics processing units (GPUs) from Nvida and central processing units (CPUs), according to Nvidia. Intel and Advanced Micro Devices, among others, make the CPUs.

High-end GPUs today typically contain hundreds of processing cores, allowing them to accelerate certain types of computational tasks much more efficiently, and thereby much faster, than CPUs. The new Fermi chip will have a little more than twice as many cores as current high-end Nvidia GPUs, according to Nvidia CEO Jen-Hsun Huang, jumping from 240 cores to 512 in Fermi.

Fermi GPUs would enable "substantial scientific breakthroughs" that would be impossible without the new technology, said Jeff Nichols, Oak Ridge's associate lab director for Computing and Computational Sciences, in a statement.

"With the help of Nvidia technology, Oak Ridge proposes to create a computing platform that will deliver exascale computing within ten years," Nichols said. Exascale computing proposes to go beyond petaflop performance (a thousand trillion) to one million trillion calculations per second.

Oak Ridge also announced it will be creating the Hybrid Multicore Consortium focused on computing with different types of processor architectures. The goals of this consortium are to work with the developers to run applications on the next generation of supercomputers built with CPUs and GPUs.

Nvidia's rival AMD recently announced one of its fastest graphics chips yet that packs 2.15 billion transistors and supports Microsoft's upcoming DirectX 11 programming interface for accelerating graphics and general-purpose computing in Windows 7. The ATI Radeon HD 5870 has received consistently positive reviews, beating comparable Nvidia chips and, most importantly, it is available now.

news.cnet.com