It’s not me, but i want to share here! It’s a mod that increase the performance based in ACE and some internal bugs!
Steam: Steam Workshop::Performance mod for ACE Stonehearth
Github: GitHub - fatsan1975/stonehearth_performance_mod: A stonehearth mod that aims to optimize simulation speed! · GitHub
Have you tried it? it would be a really nice weekend to start a new town
I’m using and helps a lot, but no much, because i have 32 pawns and now my bottleneck is Pathfinding (A*), but look… someone made a mod to fix this too, that one i didn’t tested and and don’t know if works with ACE
Reminder that people can (and did) fall into a super huge placebo effect.
They were commenting saying that it helped, but the reality is that the game was not even loading the mod!
It was a bunch of code that was not set properly and was completely ignored by the game, nothing was being done aside from just list itself as an active mod in the mod menu.
Did you tried the update in made in github?
but I wonder if there’s no way to increase performance by mods, that’s sad
Not with AI.
The best performance you can gain right now that is not placebo or can’t even tell the difference, is to simple pick a small map size with Extra Map Options mod. That will be a huge increase and very noticeable difference.
There’s theoretically ways to improve performance, I’d guess, but not with your run-off-the-mill lua/JS mod.
Replacing the GUI might actually be one of the things that might help, considering the huge advancements that have been made in the web world since whatever old Ember version Stonehearth uses. But its (likely very dated) CEF probably means you’re somewhat limited there, too, unless you polyfill.
Ultimately, I think you would only get decent performance gains if you either binary patch the game (like TTD) to improve systems, or - but that’s a big “if” - you somehow find the weak points in the current scripts and re-design the systems/re-think how they work.
For example, regarding pathfinding and the entire AI, I think it would not be impossible to yoink the existing implementation completely and replace it with something new. You could write a binary lua module that actually used multi-threading to e.g. do pathfinding, AI-stuff or similar, relieving the main thread from doing such things by using smarter things than the coroutines currently in use (as lua itself isn’t multithreadable). But saying that’s a rather big undertaking would be a vast understatement.
