Programmable Visual Processors On The Horizon

GPUs are engineered for high speed and feature highly parallel pipelines that exploit the natural parallelism inherent in vertex processing and pixel processing algorithms. However, their hardware implementations largely dictate that these algorithms are frozen in silicon. This limits the algorithmic complexity that can be supported and prevents these devices from being easily adapted to accommodate evolving standards.

From an architectural point of view, a graphics chip is increasingly becoming like a CPU because it's multithreaded, has virtual memory and intertask security, plus being programmable in a C-like language. The only difference is that graphics ICs are very parallel (as opposed to the main microprocessor), which is required so they can execute graphics more efficiently. CPUs will be highly parallelized in the next few years, but that's another topic being researched by the ROI Marketing Group.

3Dlabs Inc., Ltd. announced at the NAB (National Association of Broadcasters) 2004 trade show in Las Vegas , NV , a major breakthrough in graphics architecture entitled Wildcat Realizm technology. The architecture combines a next-generation Visual Processing Unit (VPU) and a unique Vertex/Scalability Unit (VSU). These work together to enable a software-compatible family of graphics accelerators ranging from a single-VPU AGP 8x solution to a unique dual-VPU configuration, which takes full advantage of the enhanced bandwidth of PCI Express.

Wildcat Realizm is specifically designed to provide unmatched graphics productivity for CAD, DCC and visualization professionals. It represents a significant advancement in accelerating industrial-strength shader programs written in high-level shading languages such as the OpenGL® Shading Language and Microsoft® DirectX® 9.0 HLSL. A range of products based on the Wildcat Realizm architecture is expected to be available mid-2004.