updated 09:25 pm EST, Tue December 13, 2011
NVIDIA drops proprietary lock on CUDA tech
NVIDIA took a major step towards spurring growth of its CUDA general-purpose code technology for video cards on Tuesday by posting the CUDA source code. Developers and education now have access to a variant on the LLVM compiler that will let them add new processor types and languages. The extension could see CUDA run on AMD's Radeon hardware, Intel's integrated graphics, and even use relatively old code like Fortran.
Early access is available through a screened sign-up program (above), with work on going beyond NVIDIA hardware already started by the Ocelot Project. NVIDIA hasn't said how soon it plans to widen access.
The step isn't completely necessary for general-purpose uses on rival video hardware following the existences of DirectCompute and OpenCL, but could widen support for apps that were written to use CUDA. Adobe's Creative Suite, numerous games from EA, and some pro modelling and video apps both on Macs and Windows PCs use it to accelerate tasks that would normally lean on the main processor.