Impressed… today I decided to transfer the processing of the neural networks into a seperate thread. That was very easy, even a bit too easy. And since it did not appear to have any impact on the overall performance, I decided to stress test the brain-processing thread.
at allmost seven hundred brains with an everage of over 500 connections it only needed 3ms per frame to proces.
The GPU has much harder time, I might have to do some more optimizations there.
(for reference: I am working on a 6 core i5-8600K, 16GB ram and a NVIDIA GeForce GTX 1060 6GB)