Stereo Tool

_________________
control point
control point2

good idea but i see more benefit sharing both GPU and CPU resources together especially if CPU has integrated GPU and the PC is dedicated to just sound processing.

very very interesting, who knows what does Hans think?

For what Stereo Tool currently does, there's no lack of processing power in the CPU (unless you have a very old or very slow CPU, but then you probably don't have a GPU either), so I don't really see the benefits of this. There are things for which this is definitely useful - a convolution as mentioned in this video is a prime example (but even a 1 second convolution doesn't use much CPU power if you do it right - a naive implementation would be extremely heavy but there are ways around that that are hundreds of times more efficient). The thing in Stereo Tool that would most benefit from this is the DeReverb filter, that gets heavy on the CPU very quickly.

For anything that needs to know what happened with the previous sample, such as compressors, limiters etc, it's not usable to my knowledge. So for anything like that, we would need to go back to the CPU. And going back and forth between CPU and GPU isn't free, so the benefits of using the GPU need to be pretty big for this to pay off.