In the recent years Nvidia Corp. either acquired or developed a number of exclusive technologies, such as CUDA or PhysX in order to add value to its GeForce graphics processors. Perhaps, it could be logical for Nvidia to create its own graphics application programming interface (API) so to provide new graphics effects and ensure maximum performance. However, the firm has no such plans.
"Right now we don't have any plans for [exposing the hardware rasterizer to CUDA]. Graphics APIs are designed and tuned to be great at graphics. We have focused on good interoperability between CUDA C/C++ and the graphics APIs so you can easily switch back and forth without copying data, using the best tool for the job," said Ian Buck, the software director for GPU Computing at Nvidia.
3dfx, the graphics chip designer that Nvidia acquired in 2000 used to have its own Glide API that ensured high performance and timely introduction of new exclusive graphics features. However, as DirectX and OpenGL evolved, Glide became outdated and unpopular among developers of video games.
Still, Nvidia stresses that CUDA technology is the company's platform for innovation when it comes to GPU computing, which is an indicator that Nvidia is indeed looking forward exclusive software technologies and innovations.
"Nvidia’s goal is to support all GPU programming languages, but Nvidia's platform for innovation is CUDA C/C++. We add new features to our CUDA Toolkit based on new Nvidia GPU architectural capabilities and on customer requests, and deliver 3-4 releases per year. We then work with the Khronos OpenCL working group process to incorporate new features, and when they are accepted as part of the OpenCL standard we schedule them into Nvidia’s driver release roadmap," explained Mr. Buck.