bluegaspode wrote:I'm asking because the whole evening I started to optimize as well, because I had timeouts on the new cell_maze maps as well.
Wrote a testcase for myself:
200*200map (everything visible - so a full diffusion) with
100 enemy ants
100 friendly ants
100 food tiles
I'm getting 200diffusion-rounds in 330ms now.
(Unoptimized I took 1100ms)
Using Java, on a MacBookPro (2,4Ghz, Intel Core i5)
Wow, very nice! That's almost 5x faster than what I have. Do you do anything specific that I didn't mention in my list of tips? At the moment, I don't really see how I can optimize my code more, so all tips are welcome .
Edit: How do you create such a test map? Do you mind sharing it? It would be nice to do comparative benchmarks.