Nvidia has released details of its upcoming GF100 gaming graphics processor that will compete with Advance Micro Device's most powerful ATI Radeon HD processors, such as the 5970.

To whet gamers' appetite for performance, Nvidia is touting the GF100's support for Microsoft Windows 7's DirectX 11 hardware APIs. The graphics processing unit is also the first based on Nvidia's new Fermi architecture.

The GF100's graphics performance is driven in part by its implementation of such DirectX 11 technologies as tessellation and DirectCompute. The former is a method used in representing 3-D objects that enables more realistic-looking animation. DirectCompute makes it possible for GF100 developers to boost overall PC performance by offloading more tasks, such as media processing, to the GPU, Nvidia says.

Nvidia's next-generation Fermi architecture features up to 512 CUDA cores. CUDA, or compute unified device architecture, is the computing engine in Nvidia's GPUs. The engine is accessible to software developers through industry standard programming languages.

Besides its many CUDA cores, Fermi also features error-correcting code that's expected to improve Nvidia GPUs for use in data centers. ECC protects data-sensitive applications, from medical imaging to financial options pricing, from memory errors.

Nvidia over the weekend released a white paper on GF100. The vendor has not said when the GF100 will be available and has not released pricing.

AMD's top-of-the-line GPUs for implementing DirectX 11 technology include the Radeon HD 5970. The vendor also supports Microsoft's latest graphics APIs in less expensive GPUs, such as the Radeon HD 5700 and 5600 series. Nvidia's GeForce products compete today with AMD's Radeon HD line.