OpenCL opens for business

OPEN SESAME! Or rather, Open CL. As the parallel processing programming language of choice releases its first formal spec, the major industry players are lining up to get behind the standard.

Originally conceived by Apple - and now handled by the Khronos industry group - OpenCL is a programming language that allows for programmes to be written for execution across different processing platforms, including both CPUs and GPUs.

The idea is simple: use the power of modern graphics cards to do chunky processing work, without the programmer having to learn an entirely new architecture, and without the user having to specifically make a point of doing it. OpenCL aims to make the GPGPU an entirely transparent experience for the user.

Lining up behind the 1.0 spec are companies including AMD, Nvidia, Apple and Intel. Apple will be integrating the spec into the next version of its OS, dubbed Snow Leopard, thus allowing apps like Final Cut to harness GPU processing power.

Nvidia has been keen to throw its weight behind the standard, even announcing that its own proprietary GPGPU language - CUDA - will now be 100% compatible with OpenCL. This ensures that the technology that powers GPGPU - the GPU, obviously - will surely become more prevalent, which benefits Nvidia regardless of the fact it doesn't own the OpenCL software standard itself.

With all the industry partners in agreement and compatible hardware and software now shipping, it seems that the GPGPU revolution is nigh, if not already upon us. µ