[2020] A Question about worldgen ...

Originally posted by Dragonsheepling: Update: While loading a 132km² world does nothing great for the game (It froze and locked up my PC (had to restart completely)) a simple 8km² seems to run fairly smoothly until it crashes. (Like it did just now). The worlds are truely utterly massive ... so much so that you have a clear continent formations occouring. (Massive pieces of Land surrounded by even larger (and deep Oceans).

It does take quite a while to load even just a 8km² map and ... its size is ideal for a large server population. One with many active players. It would also likely happen that entire nations would form due to the sheer size of things. Curvature is still as stongly pronounced as it is with a smaller world meaning its truely just a graphical effect and has no impact on gameplay. (Besides day/Night Cycle).

2.56 km² is already ideal for a population of more than 1000 with 100+ people playing at the same time. Going big just removes any point of ecology and scarceness from the game.

Originally posted by Valck: Interesting side note: The server apparently only uses a single CPU core for world generation. I guess there's room for optimisation there...

No, the world generation is actually the only thing in eco that makes use of all CPU cores with 100%.

Originally posted by Valck:

Originally posted by SLG-Dennis: No, the world generation is actually the only thing in eco that makes use of all CPU cores with 100%.

It certainly only used one core on my Linux box. Granted I didn't use the dedicated server but generated a "hosted" world from within the regular client, although I'd have expected both to use the same code base...

That's very possible - I was talking of Windows. The linux server is still early development.

And no, the internal server in game uses a limited 32-bit server, so it's different in restrictions indeed.