OpenCL / Cuda / etc Acceleration

OpenCL / Cuda / etc Acceleration

Is this something havok is ever likely to see? For example, in PhysX, you can accelerate the physics calculation by offloading some processing to a GPU or dedicated PhysX card. I read some articles mentioning the possibility of this for havok around 2008-09... now 2 years later and we still haven't heard anything about it?Or is this unlikely to ever happen because they are technologies of competitors like AMD and Nvidia...

2 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.
Kopavel's picture

Some facts...
Cuda is NVidia-specific. PhysX is NVidia product.
GPUs in games are almost always a bottleneck.
OpenCL is a bugfest.

Havok is cross-platform andhighly optimised for CPU/SPU.

Make a conclusion yourself...

Login to leave a comment.