Just a quick and dirty readme file for this project:
There are two "main" branches: main and CPU. 

main was the original branch and features GPU training. This turned out to be pretty limiting 
due to trying to optimize CUDA code for a range of NN sizes. 

CPU was a derivative branch. It supports only CPU training, but is much more flexible in terms of what
kind of network you can train, and thus shows much better performance. The restrictions
that were imposed on the GPU branch were lifted for this branch (I think).



Build command for main (GPU) branch:
nvcc -rdc=true -lineinfo -o runner .\biology\Genome.cpp .\biology\Species.cpp .\biology\Specimen.cpp .\biology\Taxonomy.cpp .\Simulator.cu .\Runner.cu .\Kernels.cu .\simulations\BasicSimulation.cu .\simulations\TargetSimulation.cu .\simulations\MultibotSimulation.cu .\simulations\AirHockeySimulation.cu .\simulations\PongSimulation.cu .\simulations\PongSimulation2.cu .\simulations\MultiBallPong.cu -O3 -ccbin "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30133\bin\Hostx64\x64\cl.exe"
Note: you will likely have to replace the path at the end to match where you have cl.exe.

Build command for CPU branch:
g++ -pthread *.cpp .\simulations\*.cpp .\biology\*.cpp  -o runner -g -O3