Yeah, something that takes microseconds to calculate and only needs to be calculated once per month is really taxing.
In 2.2.5. pop growth is still bonkers
- Thread starter Arutar
- Start date
-
We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
Yeah, something that takes microseconds to calculate and only needs to be calculated once per month is really taxing.
Yeah, something that takes microseconds to calculate and only needs to be calculated once per month is really taxing.
While true that a 4 GHz, Quad core CPU can do tens of billions of ops per second and this configuration was common 7 or more years ago [esp. with OC] I wouldn't be so quick to rule out the possibility that based on limitations of how Stellaris does things that we could be bottlenecking some subsystem on day 1 thus causing stutter.
For example if there was already long running population calculations that take on the order of N^2 [N-squared] time and you up the population by 5 then that already long running calculation could take 25X more time than it was before. If there exists any unforgivably bad algorithms [order of N! -- N Factorial] then upping population counts by 5x may result in a 120X increase in time.
While I would hope this by itself wouldn't necessarily cause problems this combined with other potential inefficiencies may cause problems.
Limited points of observation, but I will share this:
In my pre-patch game, every species seemed to be emigrating to every planet, regardless of habitability. In my current game, my main species, despite having full migration rights, has not migrated to any of the empire planets for which is has low habitability.
In my pre-patch game, every species seemed to be emigrating to every planet, regardless of habitability. In my current game, my main species, despite having full migration rights, has not migrated to any of the empire planets for which is has low habitability.
I suspect that based on the changelog that this beta patch was mostly to address machine empires, but they threw in the habitability fix because it was "easy." I think they need to entirely rethink how to approach pop growth and migration, which requires a lot more dedicated attention.
In the extremely unlikely case that it's done that sloppily, we're still talking about milliseconds once per month instead of microseconds.
Pretty much this. Slap a band-aid or two on it and then go back to square 1 and rethink the whole way you want this handled.
I assume that I missed it but I'll ask to be sure ... Do you have a theory as to why certain performance-related issues appear worse in 2.2.x??
If the reports of others can be trusted measured stutters [on high end systems] of hundreds of milliseconds are being reported. Different CPUs, different video cards, etc. all seem to be impacted.
Something appears to be going on. What that is hasn't been fully determined yet. It may be something simple like added pops [or something else] is causing data structure size to grow enough that cache misses are occurring now when they weren't earlier ; there could be IPC issues causing lockouts to certain critical data and wasting computer power ; ...
I have a 2.2.5 game where i set all other species to residence, main species is ar 1500 all other are below 150 and main species grows on 80% of my planets. Im fanatic egalitarian and no faction dislikes that every other species is on residence.
Correct me if I am wrong, but residence in itself should have no effect on pop growth at all?
That is what population controls are for (and these are ofc heavily disliked by egalitarians and outright impossible if you are xenophile)
Yeah, i would have thoughed so too, but i do have this game where almost none of the minor species are growing. They are all allowed to grow. So it is either 2.2.5 that actually did anything, or the fact that i put them all on residency (which i normally never do).
Paradox have a really tiresome ability to snatch defeat from the jaws of victory. They'll introduce some nice new gameplay concept but bundle it with a whole bunch of changes no one asked for or wanted.
"Here's a new awesome new pop system, but we're going to take away control of what your pops do, break robots and gene modding and introduce a nonsensical growth system. Enjoy!"
"Here's a new awesome new pop system, but we're going to take away control of what your pops do, break robots and gene modding and introduce a nonsensical growth system. Enjoy!"
Ok, so I just tried this - setting all my minor pops to residence, but have no population / migration controls enabled whatsoever (playing egalitarian / xenophile).
And indeed it appears to work!?!
I continued my first save (OP). Before setting to residence I had one single prime species pop grow in the previous 20 years. Now I have had 10 prime species POP grow within the next 4 years. About 80% growth in the last 4 years is from my prime species (non-residence).
While the sample size is ofc too small to say for certain, I think residence rights itself affects Pop growth selection!?
No offense, but nothing of this makes sense. Putting some errors in your math aside, nothing that has been proposed in this thread would require more then a linear increase in calculation time per pop. A small one on top and propably a lot less as this is something that's done per planet and species instead of per pop.
Writing this post has propably tken more time then every Stellaris running PC combined will ever require for the overhead.
Wait a minute. Talking about "a linear increase in calculation time per pop" isn't a sensible way to talk about calculations which are done per planet under the current system. ATM there's nothing to increase.
Sorry, I've propably lost something in translation here. Anyway, taking individual pops into account is not wrong. Let's say you have two species growing on a planet. One of them is a slow breeder and has the smaller number of individuals. The main species is a fast breeder. Right now, if it grows until both have equal size, the slow breeders almost complete negate the fast breeding of the main species. That's wrong. Fast breeders should outnumber everyone else in the long run.
It would be better to calculate the time to grow a new pop based on all pops present on the planet and then pick one at random weighted by the traits of all possible species.
I could explain to you some potential poor coding practices that would result in non-linear time increases but that's not the point of this thread. If you think it's impossible for someone to code so poorly [or have other restrictions that seem like poor coding] that even modern day machines have stuttering issues then you won't have to look far for an example.
You're right. Circling back to topic is best.
I agree. One "slow breeder" alien shouldn't significantly slow down the growth of an empire. There are a LOT of things that should be taken into consideration:
- Breeding speed [slow / fast / otherwise]
- Habitability
- Available jobs
- Available housing
- Suitability of jobs (?)
- Main species vs. "others"
Just out of curiosity, has it been other people's experience that the -20% growth rate penalty from specifying what species you want to grow on a planet is somehow crippling? Because at least in single-player, it hasn't been a problem for me.
Also, specifying pops to grow is very useful on at least a temporary basis after you've gene-modded your main species and you don't want to having to keep re-running the mod project because the game's insisting on growing the unmodded pops...
Also, specifying pops to grow is very useful on at least a temporary basis after you've gene-modded your main species and you don't want to having to keep re-running the mod project because the game's insisting on growing the unmodded pops...
I don't think it's optimal BUT the AI is often so brain dead that it doesn't matter in SP.