Nvidia demonstrated its RealityServer platform Tuesday at the Web 2.0 Summit in San Francisco. The new product comprises Nvidia's Tesla graphics processors running software from Mental Images, a German company Nvidia bought two years ago. Mental Images software is used in many 3D design and animation applications.
RealityServer, which is scheduled for release Nov. 30, could be deployed in-house in a large corporation. However, Nvidia is looking toward cloud-computing providers as the larger market.
Cloud providers could use the platform to host applications that would deliver 3D images of near photographic quality to any broadband-connected PC with a Web browser. Depending on its complexity, such an image created by computer graphics could take a couple of hours to render using current technology. Nvidia claims that remote computing centers using its technology can reduce the time to 15 to 20 seconds.
The rendered images could show engineering and architectural designs under different lighting conditions. Shoppers could change the furniture, fabrics and colors used in a home interior and see how it looks at different times of the day.
Because of the state of today's technology, the Nvidia platform has its limitations. The images are not interactive, meaning a person can't manipulate objects in the images.
Despite the limitations, e-commerce vendors will find more accurate product depictions valuable. Also, the RealityServer platform could someday find a home in online virtual worlds.
"This is one giant leap closer to the goal of real-time photo-realistic visual computing for the masses," Dan Vivoli, senior VP of Nvidia, said in a statement.
A developer edition of RealityServer software will be downloadable at no charge Nov. 30. Developers can also deploy non-commercial applications for free.
InformationWeek Analytics has published an independent analysis on application delivery. Download the report here (registration required).