What is going to be the real solution to pops causing performance issues?

[Falesz]

I realize we have two campaign option sliders to reduce the number of pops in the late game, which is the current solution to extreme late-game lag. I'm not happy with this, this is a defensive workaround, and not an actual solution. If I just limit the amount of pops in the galaxy, I'll just end up with a galaxy spanning empire with a million free jobs and building my empire will just stagnate.
Is there ever going to be a real solution to late-game performance issues so that an actual huge galaxy can be dominated and a galaxy spanning utopia can be built for my citizens? I realize also that this is mostly just for the hardcore roleplayers and rarely do campaigns get to the point where the game becomes literally unplayable due to lag, but hey, if there are options in the game for 1000 star systems in the galaxy, it should be playable all the way through.

[ScreamCon]

Its not that the devs can't sit down and ponder up a new way they could mathematically represent pops in their game so as to reduce computation per pop. It that once you got something complex it requires even more time to make more complex. The mathematical code representation of pops. Meaning it costs more and more money to code.

This means it has to reach a point where it effects their sales and thus devs are motivated by potential gain to upgrade. Lets say the devs got all the money they would ever had. They would not be motivated to keep developing. The very next day the top executives would conclude this and cut the programming team from the game project.

Pop calculation never reaches 0 computation requirement. Computation is a result of feedback looping over circuit. This means all the devs can do is try to get closer to 0. And as your probably guessing by now it could be while until we get an upgrade.

[Daonar]

They should reduce pops "diversity" to mere coefficients instead of trying to calculate it per-pop as they do now. They fail at it (at automated pops management) anyways; AI would never assign pops at best jobs considering their individual bonuses and maluses.

So instead of having hundreds of pops and calculating it individually for each, just have a generic "+7% mining, +11% research, -3% housing usage, +2% consumer goods usage on average" pop profile and calculate it ONCE. Problem solved.

[ScreamCon]

Originally posted by Daonar:

They should reduce pops "diversity" to mere coefficients instead of trying to calculate it per-pop as they do now. They fail at it (at automated pops management) anyways; AI would never assign pops at best jobs considering their individual bonuses and maluses.

So instead of having hundreds of pops and calculating it individually for each, just have a generic "+7% mining, +11% research, -3% housing usage, +2% consumer goods usage on average" pop profile and calculate it ONCE. Problem solved.

They could get rid of pop jobs and move to having the population number multiply over a % representation of total world productivity. Total pop increase would increase total productivity number, and jobs would change the total ratio of productivity pull to specific resources.

Making planets 'stiff' up and not change only with pop increase decrease, while also reducing pop ethic changes/factions from changing things as much would save massive amounts of reprocessing.

Ideally also its best if the system does not have to calculate all 100 pops when it refreshes but only the next pop added. But thats beyond me

[Meewec]

unless there's a cap on growth or games are made shorter every strategy game will eventually suffer from slowdown.

[mss73055]

One can't simulate a K3 civilisation on K0 hardware.
At some points the game should switch strategies on job calculation.

But instead the game was optimised by removing all speed breaks. The situation still is like a train crashing into a mountain of gelatine, but now it's an express train.

[HappySack]

Besides the obvious solution of removing all pops, the next best thing is a total rework to pop growth and efficiency of pops in general.

Compare this to the Civilization games where there actually is a noticeable difference in playing wide vs tall as it costs pops to found new cities and there is a noticeable contrast in playstyles between having multiple low productive cities with small numbers of pops in a large area as opposed to a few high productive cities with many pops in a more secluded spot.

But in Stellaris pops grow like weeds to the point where they needed to limit pop growth by default (which is now optional) in order to curb their growth for the sake of performance, likewise pop growth is nearly identical from planet to planet, again to the point where it's actively better to colonize everything as long as you can afford it because you'll eventually have enough pops to make up for it.

A rework would both need to decrease all pop growth across the board, make habitability much more impactful, and make a jobs produce more resources to compensate for less pops in general to name a few other things but this is just off the top of my head.

[Kypamop]

I'd prefer population system like in Sword of the Stars I.
Makes perfect sense, also habitability is much more sci-fyish.

[DeadLy38]

I don't think that they will change pop system (hope they will). For now I am just using the "ticks_per_turn" command to speed up my games

[Peter34]

The current system with the individual POPs can simulate gradual change over time, e.g. from migration or modification.

A numerical tracking system can't do that. Therefore, the current system is superior to the proposed numerical tracking system.

[ScreamCon]

Originally posted by Peter34:

The current system with the individual POPs can simulate gradual change over time, e.g. from migration or modification.

A numerical tracking system can't do that. Therefore, the current system is superior to the proposed numerical tracking system.

Superior simulation quality, yes. Superior compute speed, no.

[Captain Rex]

Originally posted by Peter34:

The current system with the individual POPs can simulate gradual change over time, e.g. from migration or modification.

A numerical tracking system can't do that. Therefore, the current system is superior to the proposed numerical tracking system.

while it by itself would be superior/More fun, it is not the better option for me since i can't go past 2300 without serious slowdown.

[Peter34]

Without simulated change, including the simulation of the speed at which the change happens, Stellaris is just a map painter.

[Half Phased]

My first thought would be to adopt Victoria 2/3’s pop system, but then the developers would have to state how many of X a pop is. And, Victoria’s pop system is designed around a relatively fixed workload, meanwhile in stellaris the number of “states” or rather “planets” can go up insanely high and keep going if empires keep cranking out habitats, though whether that will rival the number of states that Victoria 3 will have… IDK.

But it would give granularity over pops and would eliminate the per pop calculations by condensing all pops.

[box]

1) Get rid of population entirely or allow for population and jobs to be "promoted" into larger groups so we don't have as many individual pops clogging the system
2) Rework the population code entirely from bottom up to remove the bloat
3a) Give incentives for players to "depopulate" the galaxy EDIT eg a Species war CB
3b) Create new mechanics that allow population of the galaxy to "disappear" eg send colonists to another galaxy, galactic pandemic, storms causing pop growths to be reduced, conflicts and dispute events that result in more pops being lost etc

Then again you the player could just get a better rig.

[ScreamCon]

Originally posted by box:

Then again you the player could just get a better rig.

Don't scream that aloud, the forumers saying no better rig could fix will hear you