vacancies advertise contact news tip The Vault
facebook rss twitter

NVIDIA adds OpenCL support to CUDA

by Scott Bicheno on 9 December 2008, 10:26


Quick Link:

Add to My Vault: x

Opening up

CUDA is the NVIDIA technology that allows applications to be run on the GPU, rather than the CPU, and take advantage of the superior parallel processing capabilities of the GPU.

It already supported a number of programming languages such as C, Java, Fortran and Python, and the move to support OpenCL (open computing language), which is specifically designed for parallel programming, means applications can be written to run on either NVIDIA or AMD/ATI GPUs. CUDA is not currently supported by AMD.

"The arrival of OpenCL is fabulous news for the computing industry and NVIDIA is delighted to be playing a highly active role in the establishment of a new standard that promotes computing on the GPU," said Manju Hegde, GM of CUDA at NVIDIA.

It's reported that Apple will be including OpenCL in its OS X Snow Leopard release. "We are delighted that Apple has helped spearhead OpenCL," said Hegde. "Their recognition that the GPU will now play an essential role in consumer applications is a significant milestone in the history of computing."

What isn't clear from the announcement is if there remain any advantages to using CUDA, as opposed to OpenCL, and if so what they are. It would also be interesting to know if NVIDIA is going to produce a CUDA-to-OpenCL converter for existing and future CUDA based applications.

At an NVIDIA event yesterday, it was revealed that CUDA will also support Microsoft's impending API for this sort of thing, called DirectX 11 Compute. It also demonstrated its new 3D technology called GeForce 3D Vision, which looked promising, but does require special glasses and a ‘3D ready' display. It will be launched at CES in January.


HEXUS Forums :: 2 Comments

Login with Forum Account

Don't have an account? Register today!
Awesome. I was hoping this would happen.