Extreme slowdown from periodic boundary conditions

[Chelsea486MHz]

As per this revelation -> #9 (comment)

The way periodic boundary conditions are handled is beyond inefficient. Claiming I am surprised would be lying, but I am still speechless.

[raresraf]

Seems that this in not a Cataclysmic problem of cosmological proportions after all. 😄
Only when I ran
./senpai.bin --in examples/DES.mol --out render.xyz --density 0.001 --time 0.01 this issue appeared. If we run with --copy 5, periodic boundary conditions total time drops below 0.1%

[Chelsea486MHz]

Well, the force computations do scale at O(n²), while the PBCs scale at O(n).

I'll look into optimizing the PBCs anyway, since a lot of research went behind that.

Still, it's nice to know the impact isn't so bad.