Folding Forum

nvidia-announces-cuda-40

I'm sure it will take some time before folding actually uses this new capability and I suspect it will not be a big productivity gain because it's main focus is on RAM and folding speed is not that RAM sensitive. However, perhaps it will allow for less CPU usage while GPU folding? But I'm just speculating, without any real knowledge. However, I'd be interested in how this will really affect GPU folding.

It might not affect anything at all. Last I knew they were switching the core over to OpenCL which Nvidia has separate drivers for. Then again their implementation of OpenCL could just plug into CUDA in the background or simply be CUDA with the OpenCL API on the front side.