AMD, CyberLink team up on DirectCompute acceleration

As much press as OpenCL has gotten, it's not the only player in town for vendor-agnostic GPU-computing applications. With Windows 7 now less than two weeks away from launch, eyes are starting to turn to the DirectCompute programming interface Microsoft has included in its DirectX 11 toolkit.

In their latest joint announcement, AMD and CyberLink have vowed to broaden their relationship "with a strategic focus on . . . DirectCompute." Their plans are pretty much what you'd expect:

The companies will work together to accelerate many computationally intensive tasks in CyberLink applications, such as video transcoding, automated facial recognition and tagging, video editing and processing applications with ATI Stream technology using the full specification capabilities of the new Windows 7 DirectCompute API.

AMD goes on to boast about having the only fully DirectX 11 compliant line of graphics processors, the Radeon HD 5800 series. However, going out and buying a fancy new Radeon may not be a prerequisite for running DirectCompute-enabled CyberLink apps.

Last month, AMD marketing chief Nigel Dessau wrote in his blog that DirectCompute will support "[the] previous generation of DirectX 10/10.1 GPUs." He nonetheless added that writing DirectCompute code for DX11 graphics processors will be "much more straightforward" for developers and will enable "new algorithms that were not possible previously," including "order independent transparency, ray tracing, better shadows, and depths of field."

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

This discussion is now closed.