So a running game is going to take a while to show a performance benefit. Once the AI gets the planets moved over to what the mod is looking for they'll start costing less. What happens is a "perfect" planet no longer checks for new buildings to build, no longer checks to move pops, and no longer does a number of other checks. It instead checks for a much more simple things such as the planet having a missing pop, building in need of upgrade, and the like.
Running one from the start has a performance benefit throughout in that the mod AI builds what it needs to and then stops checking the planet until the planet has unemployment.
If you are going to let it run overnight I wouldn't mind seeing a comparison video between the mod and the vanilla performance. It would be nice to see how far they get in 1 or 2 hours of running.
Alright, here are my findings after I've run an observer game for quite some time now with your mod enabled. Game version 2.2.3, Stellaris_Test version of the mod, large galaxy with default settings.
The good:
In early to mid game I definitely saw some overall performance improvements. Compared to the other games I've run without the mod (same galaxy settings) the days ticked by noticeably faster. Good job!
The bad:
The changes you've made don't seem to prevent the usual late game lag and constant stuttering I've suffered from since the Le Guin update. Once the game hit the year 2400, performance went downhill quickly. I had to call it quits in 2418, as the game became pretty much unplayable on my machine at this point (constant freezes, ui becoming unresponsive, etc.). Calculating a month takes 55 seconds (!!!) on the fastest speed setting, which means it would take me roughly another 15 hours to reach the default victory year of 2500...
Disabling the mod doesn't seem to make much of a difference performancewise at this point, so making a comparison video wouldn't be of much use, sorry.
I've attached the latest savegame to this comment. Anyone who wants a good laugh or who still believes that the game runs fine on his machine, feel free to give it a try.
And no, the abysmal performance isn't due to Glavius's AI mod. I've had the performance of unmodded games tank equally as bad.
Ok, for the sake of the "Am I missing something huge?" I downloaded your save and ran it.
A month at speed 3 took about 24 seconds, there is a distinct pause at the 1st of the month and again (predictably and nearly uniformly) at around the 20th. I watched about 4-5 months without touching anything to be sure.
So, I retract my earlier statements of "it's all in your head, bro" because with your save it was a noticeable slowdown for me, too.
Now, the question becomes more akin to "what should our expectations be for a good sized, fully painted map with huge fleets at mid-game trying to run at maximum speed?" and I don't have an answer for you, because my PC specs are well, well above the recommended line (and I'm sure yours are, too).
So move my vote from "It's fine, suck it up" to "Ok, maybe some optimization isn't a terrible idea."
Thanks for testing out my savegame.
I'm glad we're all on the same page now and are able to agree that the game needs some serious performance tweaks.
As to your question regarding what our expectations should be: I expect to be able to reach the victory year of a game on default settings with a pc that by far outperforms the recommended specs. It's as easy as that. Currently, I'm not.
Of course, a certain slowdown is to be expected the further the game progresses. I'm fine with that and that's not what I'm talking about. But right now the game becomes unplayable at some point in late game. The reason you haven't experienced this yet is probably because you're usually quitting your games after 150 years, as you said yourself in an earlier post.
