Over on the Discord, we just had a case where an user with a LOT of mods was getting a reduced framerate that started after a battle ended and persisted until the game was restarted.
I too am trying to figure out what's going on when this perf issues happens, but so far I've seen no indication that it's CA. That the problem resolved itself after removing the mod and restarting has everything to do with restarting and not the mod.
My personal experience with the mod is that the battle perf impact is negligible, you can verify this by looking at the timing information that is logged if you search for the string `total time in EveryFrameDamageDetector` in the log. The timing information should be pretty accurate since it uses wall-clock time at the start and end of the "advance" method, which is where all the work happens.
Regarding the size of tracked data outside of battle, it might be large (usually ~10MB), but it's all in a single string so there's not a huge amount of reflection calls to serialize a complex object hierarchy like you have with other objects. I'll test it, but I really doubt it's the problem.
Over on the Discord, we just had a case where an user with a LOT of mods was getting a reduced framerate that started after a battle ended and persisted until the game was restarted.
They fixed it by removing only Combat Analytics.
Here's the direct link to the resolution: https://discordapp.com/channels/187635036525166592/619635013201428481/722710987571986543
Taking a shot in the dark, is there logic that's looping over combat results every tick after the first battle happens for the session? Like a report in intel that's being calculated every tick or something?
From what I can tell from profiling when this is happening, there's something happening with economy scripts where it's thrashing around, but it's hard to tell with all the obfuscation and not running in a "true" profiler (since it's just attaching to a process).
CombatResults are created once when
is called at the end of battle.
So in short, from the data I have and my personal experience playing the game with lots of mods, CA isn't the culprit. At some level I would love for it to be the problem, because then I could fix my own play-throughs