The treachery of optimization
#2
I just couldn't leave it like that, I don't think I'd be able to fall asleep with such a negative result, so I tried bending the code a bit further.
I'm now down to 10.5 seconds, about 5 % less calls to CheckConnection. The initial assumption that the big hitbox would help turned out not to be true, most likely because the NetherFort passages twist and turn a lot and so the new pieces mostly don't go outside of the already-placed big hitbox. Also, the 3D space makes it a bit different from the 2D visualisations in my head.

I guess next up is some form of R-tree or kd-tree or something like that.
Reply
Thanks given by:


Messages In This Thread
The treachery of optimization - by xoft - 02-28-2017, 07:00 AM
RE: The treachery of optimization - by xoft - 02-28-2017, 08:19 AM
RE: The treachery of optimization - by xoft - 03-22-2017, 11:16 PM



Users browsing this thread: 3 Guest(s)