You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I had thought of calling worldengine from a script that would run through some fairly large amount of seeds to get a shape I liked. In my plan, I would generate a small (128x128?) version, and when I saw a thumbnail that looked good, I'd use the same seed but make it way bigger.
It turns out that changing the size changes the landmass entirely. I imagine that's because the tectonic simulation depends heavily on the size of the area.
I'm wondering though, is there a way to do what I described? Run many low fidelity simulations until I find a shape I like, then render it in full size? The alternative will take a whole lot of time.
If there is no way, can you think of a theoretical approach to doing this? Maybe I can take a crack at implementing it.
The text was updated successfully, but these errors were encountered:
I tried the script thing, just generating a bunch of land masses at full resolution. It's not feasible. I'm trying to generate quite large maps (8192x4905), and each one takes forever. It's not feasible to generate a bunch and choose the one I like.
I really am willing to take a stab at implementing low fidelity simulations, but I think it would help a lot if someone who actually wrote the thing could give me some cliff notes for what they might do if they were going to try, or at least what to watch out for?
Yeah, I just did a fermi estimate based on some smaller maps, and I think each map the size I want would take about 7.5 hours to generate. In my perfectly favorite ideal world, I'd generate a map 3 times that width, but that would take something like 68 hours of processing time on my machine. I could actually see doing that for this project if I knew for sure the map was what I wanted (by generating a low fidelity previous first).
I had thought of calling worldengine from a script that would run through some fairly large amount of seeds to get a shape I liked. In my plan, I would generate a small (128x128?) version, and when I saw a thumbnail that looked good, I'd use the same seed but make it way bigger.
It turns out that changing the size changes the landmass entirely. I imagine that's because the tectonic simulation depends heavily on the size of the area.
I'm wondering though, is there a way to do what I described? Run many low fidelity simulations until I find a shape I like, then render it in full size? The alternative will take a whole lot of time.
If there is no way, can you think of a theoretical approach to doing this? Maybe I can take a crack at implementing it.
The text was updated successfully, but these errors were encountered: