Performance Megathread

[Elkantar]

I remember when le guin came out how i said the game will end like this with perfomance and more micro then less. I got downvoted and disagreed on, its sad that i saw it right from the start. I think the whole Job System is a nice feature but don't work on the current engine on bigger galaxys. I won't even mind they rollback the whole stuff when it makes the game playable again. I know its a controverse but lets put this straight. The current system don't work for larger galaxys doesnt matter the pc and i have a high end one. Why we still stick to it? Cause the work for revamping the old system to work with le guin and some dlc's is a lot of work i agree. The current system can only have bandaid fixes unless the whole population job and ressource management gets ground up new designed. The only short bandaid fix is like limit pop to 25 max per planet (like the old system with visible tiles instead of hidden ones we have now). Then work all the other stuff balance wide around this. Like Districts, Housing, Buildings etc.

an Example:
Building Slots unlock like now but of course with the Maximum of 25 Pop to unlock but the same Number of slots we have now.
Pop works in districts, while buildings produce without pop but with bonuses from tech etc.

Problem:
Slave and Population which atm can'T work specialist jobs benefit more since they get production boosts on District stuff while they can't go specialist which is then obsolete as job choice.

Example 2:
All is centered around Pop % to max Planet Pop %. Districts work as now same as Buildings but lowered to 25 Pop including unlocking Building Slots. But all the things produce just % of the Planet max population to population ratio. Specialist and Jobs which cant be taken by specialist will adjust the % of the produceable by the populations.

 

[Enlonwhite]

I have read a lot of this thread, I do have some end game lag issues (not as bad as some post) its playable if only just to a point where I can be ok with it.

my main issues are job popping pops, who keep switching jobs daily......and some pops just not wanting to pick jobs for a LONG TIME, when jobs of that stratum are available
if i play on huge galaxies/large I have a bit more issue but not unplayable just annoying in a way.

I love what the devs have in store for the game feature wise. Just to point that out.

I do find that since we have a monthly income that pops should only change/promote jobs only every 4-7 days (even if its suppose to be only 25% at a time or whatever)
I still see the job popping pops to this day (and its unneeded calculations regardless of if its the cause of the lag or not)

having the AI function a little different improve performance may be a sacrifice that is needed to improve late game play, even if that defeats some style the devs want.

overall I hope this game is supported and maintained for at least a few more years, I really love its idea and design.

 

[Roobster88]

Wouldn't performance and micro become substantially better simply by reducing the amount of pops in the game?

-Reduce the size of planets to be between 1-10 districts and habitats/ringworlds to some other values that are suitably smaller.
-Reduce the amount of housing provided by districts and buildings and stop pop growth completely if there is no housing.
-Adjust all other values and stuff like building pop requirements to take the changes above in mind.

All in all this would be a quite low-effort way to improve performance and reduce micro and if it's not enough, then cut down on the pop amounts even more.
There is no law of physics that says there has to be dozens or hundreds of pops on a planet since a pop can represent any number of sentients.

 

[BlackUmbrellas]

Wouldn't performance and micro become substantially better simply by reducing the amount of pops in the game?

-Reduce the size of planets to be between 1-10 districts and habitats/ringworlds to some other values that are suitably smaller.
-Reduce the amount of housing provided by districts and buildings and stop pop growth completely if there is no housing.
-Adjust all other values and stuff like building pop requirements to take the changes above in mind.

All in all this would be a quite low-effort way to improve performance and reduce micro and if it's not enough, then cut down on the pop amounts even more.
There is no law of physics that says there has to be dozens or hundreds of pops on a planet since a pop can represent any number of sentients.

I'm honestly not opposed. It'd take some tweaking, but bringing population levels more in-line with pre-2.2 levels might be a decent solution. It'd also make species traits more impactful and planets feel more unique.

 

[Roobster88]

I'm honestly not opposed. It'd take some tweaking, but bringing population levels more in-line with pre-2.2 levels might be a decent solution. It'd also make species traits more impactful and planets feel more unique.

Yeah, that's actually what I was going for. I'm thinking the typical planet should have less than the typical pre 2.2 planet, to allow "performance-economy" for having a couple of Coruscant-planets also.

I remember a dev diary for 2.2 were Wiz said something along the lines of "we're keeping planet max size at 25 because we see no reason right now to change it" and I thought to myself "won't there be a lot more pops then and a lot more lag?".

 

[Shark7]

Wouldn't performance and micro become substantially better simply by reducing the amount of pops in the game?

-Reduce the size of planets to be between 1-10 districts and habitats/ringworlds to some other values that are suitably smaller.
-Reduce the amount of housing provided by districts and buildings and stop pop growth completely if there is no housing.
-Adjust all other values and stuff like building pop requirements to take the changes above in mind.

All in all this would be a quite low-effort way to improve performance and reduce micro and if it's not enough, then cut down on the pop amounts even more.
There is no law of physics that says there has to be dozens or hundreds of pops on a planet since a pop can represent any number of sentients.

This only helps if the AI doesn't spam Habitats at every planet in their empires. I do wonder if its the planets that are the problem, or the fact that the AI builds as many habitats as it can.

I also have noticed in my empire that when you do have more pops than jobs, the pops bounce around all over the place being assigned and reassigned to jobs. There must be a way to code the game so it doesn't do that every game tick.

 

[Roobster88]

This only helps if the AI doesn't spam Habitats at every planet in their empires. I do wonder if its the planets that are the problem, or the fact that the AI builds as many habitats as it can.

I also have noticed in my empire that when you do have more pops than jobs, the pops bounce around all over the place being assigned and reassigned to jobs. There must be a way to code the game so it doesn't do that every game tick.

I knew the moment habitats were announced that the implementation was a big mistake in that it would wreck performance to allow the AI to build new planets infinitely.

Personally I've used mods to disable them ever since they came, like you say something needs to be done about them. There are any number of ways of fixing them if the devs have the will.
Perhaps letting them count in the starbase cap could be a good start.

 

[KubiG37]

Honestly I don't understand how 10 years old Victoria 2 can get away so easily with its Population mechanics, which are probably even more complex (and number of pop groups is larger too) than in Stellaris, yet suffer no such performance drops, while Stellaris becomes literally unplayable past certain date (depending on galaxy size, and your CPU, but it WILL come 100% eventually)

 

[grommile]

Honestly I don't understand how 10 years old Victoria 2 can get away so easily with its Population mechanics, which are probably even more complex (and number of pop groups is larger too) than in Stellaris, yet suffer no such performance drops, while Stellaris becomes literally unplayable past certain date (depending on galaxy size, and your CPU, but it WILL come 100% eventually)

Victoria 2's mechanics slow down noticeably as the game progresses due to the way migration interacts with the demographics simulation. Ten million North German Protestant Farmers in the same province take approximately the same amount of time to process as ten North German Protestant Farmers in a province.

Stellaris's problems seem to stem from having a proliferation of O(NxM) calculations. The impression I get is that if a planet has ten empty job slots, then every day each of those slots examines each pop on the planet to see if it's eligible to take up the empty job.

Now imagine an ecumenopolis with ten empty job slots.

 

[KubiG37]

But not to an extent that I would really notice it, and mainly, at speeds I normally play (3 out of 5), it doesn't lag at all.

Anyway, for Vic 2 these mechanics are essential and desired. In Stellaris, even though they have added ton of complexity to each pop (happiness, ethics, crime, class, growth/decline, faction/issues, migration), it doesn't really affect the game. Unhappy Stellar pop won't incite a rebellion, or prompt foreign intervention if treated badly, or move to another empire as a refugee - therefore is of no concern to you no matter what empire type you play. All it can do AT MOST, is slightly lower its resource output, or trigger 'bad' events, that are a joke and can be ignored completely.

Therefore, all these new complex pop calculations are in my opinion unneccessarily complicated, while not providing anything relevant to gameplay. For sake of perfomance, I'd like the devs to perhaps take a little step back and simplify some things.

 

[GeorgieBest]

But not to an extent that I would really notice it, and mainly, at speeds I normally play (3 out of 5), it doesn't lag at all.

Anyway, for Vic 2 these mechanics are essential and desired. In Stellaris, even though they have added ton of complexity to each pop (happiness, ethics, crime, class, growth/decline, faction/issues, migration), it doesn't really affect the game. Unhappy Stellar pop won't incite a rebellion, or prompt foreign intervention if treated badly, or move to another empire as a refugee - therefore is of no concern to you no matter what empire type you play. All it can do AT MOST, is slightly lower its resource output, or trigger 'bad' events, that are a joke and can be ignored completely.

Therefore, all these new complex pop calculations are in my opinion unneccessarily complicated, while not providing anything relevant to gameplay. For sake of perfomance, I'd like the devs to perhaps take a little step back and simplify some things.

I feel like there is plenty of suboptimal code which is causing more calculations to be performed than are necessary. A hell of a lot of performance could be gained just from streamlining this without actually having to simplify the underlying mechanics.

 

[GnoSIS]

Victoria 2's mechanics slow down noticeably as the game progresses due to the way migration interacts with the demographics simulation. Ten million North German Protestant Farmers in the same province take approximately the same amount of time to process as ten North German Protestant Farmers in a province.

Stellaris's problems seem to stem from having a proliferation of O(NxM) calculations. The impression I get is that if a planet has ten empty job slots, then every day each of those slots examines each pop on the planet to see if it's eligible to take up the empty job.

Now imagine an ecumenopolis with ten empty job slots.

Yes victoria 2 falls into that trap, hence why assimilation is important. I believe most mods should have stronger asimilation for that reason. It also depends on player actions and style of game.

 

[GnoSIS]

Hi all!!

Big Day tomorrow for Stellaris! I hope we'll have good news about performance! I feel optimistic!

 

[Blayze]

Hi all!!

Big Day tomorrow for Stellaris! I hope we'll have good news about performance! I feel optimistic!

This aged well.

 

[LilSquiggle]

Is de-sync in multiplayer considered to be something thats still part of performance that has to be discussed here or should I making a new thread?
Literally yesterday we've suddenly gone back to hitting continuous de-syncs. Playing with the exact same people as I was a week or two ago when it wasn't happening, same game settings, zero mods, all DLC.

Has something changed that we aren't aware of? Why is this suddenly happening?

 

[UltimateTobi]

Is de-sync in multiplayer considered to be something thats still part of performance that has to be discussed here or should I making a new thread?
Literally yesterday we've suddenly gone back to hitting continuous de-syncs. Playing with the exact same people as I was a week or two ago when it wasn't happening, same game settings, zero mods, all DLC.

Has something changed that we aren't aware of? Why is this suddenly happening?

I'd say it's considered stability, which they are working at.

 

[permeakra]

their DLCs usually contain a hell of a lot more than the average DLC for RunNGun XIII.

If you compare resources put into average PDX dlc and average plot (not cosmetic) RunAndGun DLC like Borderlands, I would question it. No doubt PDX games have higher replayability potential, but they don't spend on graphics and graphic assets nearly as much, especially when dlc's being considered, and are free from level design.

The current system don't work for larger galaxys doesnt matter the pc and i have a high end one.

The current model as it is presented to the player is definitely suited for high performance, especially with recent propagating modifier updates. It's the specific implementation and accompanying services that are a problem, and they are being improved.

 

[RevanOne]

Hi,

I've used to like this game and played it a lot in the past but I haven't been able to play it for the last 2-3 months now. This performance issue is game breaking, cant play a game for more then 100 years without slowing so much that becomes unplayable. Is there going to be a patch to fix this or not ? It is literally unplayable for a few months now and I see no direction from the developers address this. I won't be buying any products until they fix this !

 

[Rabid Ice Weasel]

Hi,

I've used to like this game and played it a lot in the past but I haven't been able to play it for the last 2-3 months now. This performance issue is game breaking, cant play a game for more then 100 years without slowing so much that becomes unplayable. Is there going to be a patch to fix this or not ? It is literally unplayable for a few months now and I see no direction from the developers address this. I won't be buying any products until they fix this !

The 2.6 patch their working on is supposedly going to include some performance and bug fixes, so we can hope for significant improvement. See a few Dev. diaries ago for more info.

 

[GnoSIS]

This aged well.

Tell me about it. They are not helping at all.