Nvidia, the Santa Clara, Calif.-based maker of graphic processors and cards for both PCs and game consoles like PlayStation and Xbox, announced Nov. 8 that it would begin to roll out new architecture for its GeForce graphics processors.
The company also announced that the new technology would be used in its professional line of Quadro graphics cards, which will hit the market in early 2007.
The company unveiled what it calls CUDA (Compute Unified Device Architecture), a new computing architecture for using thread computing on Nvidia GPUs (graphics processing units).
Thread computing allows hundreds of on-chip processor cores to simultaneously communicate and cooperate to solve complex computing problems.
In addition to its new architecture, Nvidia will offer standard C compiler programming language for developers, which the company said will offer better performance and capabilities for computation-intensive applications, such as data analysis and technical computing.
Nvidia will offer its new architecture first in its GeForce 8800 graphics processors, which are geared toward consumers and gamers.
The company also announced a new motherboard for use by gaming enthusiasts, called the nForce 680i SLI MCPs (media and communication processors).
These new graphics cards are specifically designed to work with Intel Core 2 Duo and the yet to be released Core 2 Quad processors.
In 2007, the company will incorporate the new technology in the Quadro graphic chips, which are used by professionals and those developers in the commercial space.
The company has been developing this new architecture for its graphics chips for the past three years, said Andy Keane, the general manager of GPU computing.
For consumers, the new architecture will allow better streaming video and images that will help with gaming.
For commercial users, the new processors will allow PCs to process more information faster.
"People who get a PC with this technology now have the capabilities of a large-scale computer," Keane said, referring to the commercial uses of the new graphic chip architecture.
For example, health care providers can use the PCs that have the new graphics processors to create medical imaging of a patients X-rays.
Since the new GPU architecture allows the chips to process information in a parallel format—as opposed to a sequentially—it allows highly complex information, such as images and video, to be processed faster.
"Medical imaging, with high-resolution images, is a perfect application for these chips," Keane said.