Compile the project using make and just run the executable. Write stop
if you wish to end the generation.
I am not sure if control c fucks up the line endings so be careful.
Statistics that are reported are ALWAYS taken from a single thread.
In main.cpp you can specify the limits, as well as the engines name. In generation.cpp you can modify the format or the adjudication rules.
You can also specify a epd book in the config.
Currently it's not supported as I first need to support it in https://github.com/Disservin/chess-library
Pros:
- You dont have to do anything other than plugging your engine into this and recompiling.
- Cross OS support.
- epd book support
- Multithreaded
Cons:
- A builtin data generator into your engine is in my experience faster, unless I have a bottleneck somewhere in my code, please tell me if thats the case.
If you specify 30 threads to use, it will spawn 30 engine1's and 30 engine2's, and they will take turns when playing.