Nvidia's CUDA architecture examined


— 3:02 PM on February 16, 2007

Nvidia released the beta version of its Compute Unified Device Architecture (CUDA) software development kit today, and the folks at Beyond3D have put together a couple of articles on the architecture for the occasion. One of the articles briefly sums up what actually CUDA is, while the other goes into more detail about how CUDA works and how it's implemented. In short, CUDA is an architecture that opens up Nvidia's GeForce 8-series graphics processors (and their flurry of stream processors) through a high-level, C-like application programming interface. With the CUDA API, developers can code non-graphical applications that rely on Nvidia's new graphics processors to perform computations. If you're interested in the nitty-gritty details behind Nvidia's general-purpose computing implementation or even want to take up writing GPGPU apps yourself, Beyond3D's articles should be a good primer on the subject.

   
Register
Tip: You can use the A/Z keys to walk threads.
View options

This discussion is now closed.