In the battle of GPGPU API's, a lot of acronyms and words are being thrown around. nVidia pitches its set of robust tools such as CUDA language, Nexus development environment, GPU acceleration for open-source Bullet and proprietary PhysX physics engines and many more. From other side, AMD speaks OpenCL through every vocal cord and pore of company's skin, Apple is also the forbearer of OpenCL and its "openness". Intel was supposed to be the third [fourth?] player with Larrabee and its native set of compiler tools, but as we wrote back in early October, the project's first silicon ended up as a dud.
Full story at
BSN
