Wanted to try the 25x crisis modifier out. Set end game date to be 2300 and I can't even make it to see the crisis before it becomes mostly unplayable. 2340s and I have about a second of input lag and everything is a slow, stuttery mess.
- 1
-
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.
Our computers are crying to see Stellaris 2 , these performans issues are the tears 
- 1
- 1
Laptop
i5-5200cpu 2.2ghz
12GB Ram
geforce 920m
The largest map.
Default game settings (30AI)
Is my computer too weak?
You can't play in 2230. One year lasts 30 minutes or more.
It used to work. But the creators of the game blocked it.
fps 40 extra
Do you know any ways to accelerate?
i5-5200cpu 2.2ghz
12GB Ram
geforce 920m
The largest map.
Default game settings (30AI)
Is my computer too weak?
You can't play in 2230. One year lasts 30 minutes or more.
It used to work. But the creators of the game blocked it.
fps 40 extra
Do you know any ways to accelerate?
Faster Core speeds. My laptop from the GeForce 660M era has a faster processor. If you can upgrade, would recommend the CPU. Also, try a smaller map, fewer AI, fewer/different mods.
I propose a new DLC: Performance.
2D graphics and turn-based system. For owners of old laptops and computers.
Turn-based games have never had a performance problem.
2D graphics and turn-based system. For owners of old laptops and computers.
Turn-based games have never had a performance problem.
- 6
Civilization 5 & 6 have huge performance problems in my opinion as it takes very long for each turn on the biggest map & full Civs in late game. I can read a book while playing Civ. In Stellaris, i had huge performance problems in lategame on the biggest possible map before 2.6, but not to that extend.
- 1
I think the people here don't understand what was said in the dev blogs already when the devs described their old engine.
Stellaris has multithreading and is a multithreaded engine in the sense that several subsystems are split into seperate threads. Those systems need to complete and join together to produce the next frame/next sate of thegame. They never said that each subsystem, like planetary and modifiers evaluation is multithreaded. That's why all other cores in our CPUs are standing still, that's why you need a good performance CPU that does more per CUP core, that's why a 60$ old cpu has the same performance - new CPUs do less stuff per core that old ones (not alll of them)
What everyone here was asking from the start is just that: another level of multithreading.
Stellaris has multithreading and is a multithreaded engine in the sense that several subsystems are split into seperate threads. Those systems need to complete and join together to produce the next frame/next sate of thegame. They never said that each subsystem, like planetary and modifiers evaluation is multithreaded. That's why all other cores in our CPUs are standing still, that's why you need a good performance CPU that does more per CUP core, that's why a 60$ old cpu has the same performance - new CPUs do less stuff per core that old ones (not alll of them)
What everyone here was asking from the start is just that: another level of multithreading.
- 1
- 1
... and performance is still the elephant in the room and the limiting factor of the game.
- 4
- 1
I found an interesting large save file on other forum community. There are 719 colonies with ~28 thousand populations. If you wanted testing your hardware, checked this save file. May be this save file help developers for found critical place in performance of game.
This game was played with some gameplay mods, so after loading a save on vanilla, all resources fall down, do not pay attention.
This game was played with some gameplay mods, so after loading a save on vanilla, all resources fall down, do not pay attention.
I experience insane stuttering when the game is on fastest speed and not on pause. When i pause all the stuttering is gone. Normal speed also stutters, bet less. So the higher the gameplay speed the more stuttering. It's unplayable at fastest speed and normal speed gives me a headache.
My specs are:
AMD Ryzen 5 3600
Nvidia Geforce RTX 2070 Super
Crucial Ballistix 32 GB RAM 3600Mhz
Samsung Evo Plus 500 GB SSD
MSI Mortar Max motherboard
Windows 10 ofcourse
I just bought the 5 newest expansions including race packs etc.. So that's a total investment of 120 euro's minimum?
I lost count obviously. Really like the game, but i can't enjoy it with this amount of stuttering.
Does someone know a fix? Please let me know.
Greetz Bef
My specs are:
AMD Ryzen 5 3600
Nvidia Geforce RTX 2070 Super
Crucial Ballistix 32 GB RAM 3600Mhz
Samsung Evo Plus 500 GB SSD
MSI Mortar Max motherboard
Windows 10 ofcourse
I just bought the 5 newest expansions including race packs etc.. So that's a total investment of 120 euro's minimum?
I lost count obviously. Really like the game, but i can't enjoy it with this amount of stuttering.
Does someone know a fix? Please let me know.
Greetz Bef
- 1
To cut a long story short, it's the game engine itself. You could splash out £10k on the best AMD threadripper presently out there, and it won't make the slightest dent in performance. Better specs only delays when the slowdown happens, it doesn't stop it.
- 4
- 1
I have a far-fetched conspiracy theory that during the Megacorp's development, Wiz saw that Stellaris is fundamentally broken to realize his vision of the game, stayed as a game director to just finish the economic rework/release the DLC in whatever state it was currently in, and then went ahead to higher-ups at PDS and asked them to give him a green light to start conceptualizing Stellaris 2 while outsorcing the original Stellaris to grekulf.
I've heard a couple of times that Wiz is currently working on a "super-secret project" and I am willing to suggest that it's not Vicky 3 but actually Stellaris 2.
If it is indeed in the works, then all of the present issues with Stellaris can be explained by the fact that the project only has a skeletal crew consisting of new and inexperienced devs that are asked to release cash-grab DLCs because there is no point to make any drastic reworks of things that are already in the game. Myriad of bugs and general absence of testing with every update can be explained by inexperience and lack of manpower. Release of feature-bloat shiny new DLCs is required for Stellaris to stay afloat before Stellaris 2 is finished, and this even has a benefit of being able to test the waters and see what new features players like and dislike so that they can be better implemented in the next game. Finally, general lack of communication would be because lower-ranking devs just don't have the authority to do so, and the higher-ranking know that the game is on its way out - there is no vision for something that would be dead in a couple of years. The only reason anyone even looked at fixing performance in Stellaris is because the outrage was so big it would be unfeasible for them to just leave it as it is and not do anything about it.
I'll just make a wild guess and say that within the next year/year and a half PDS will announce Stellaris 2 and within the next 2-3 years we'll get to play the next game. Don't think there's any hope for any fundamental changes, Stellaris will keep getting cosmetic DLCs until second game is announced, then they'll just polish it up a bit and finally put it to rest.
I've heard a couple of times that Wiz is currently working on a "super-secret project" and I am willing to suggest that it's not Vicky 3 but actually Stellaris 2.
If it is indeed in the works, then all of the present issues with Stellaris can be explained by the fact that the project only has a skeletal crew consisting of new and inexperienced devs that are asked to release cash-grab DLCs because there is no point to make any drastic reworks of things that are already in the game. Myriad of bugs and general absence of testing with every update can be explained by inexperience and lack of manpower. Release of feature-bloat shiny new DLCs is required for Stellaris to stay afloat before Stellaris 2 is finished, and this even has a benefit of being able to test the waters and see what new features players like and dislike so that they can be better implemented in the next game. Finally, general lack of communication would be because lower-ranking devs just don't have the authority to do so, and the higher-ranking know that the game is on its way out - there is no vision for something that would be dead in a couple of years. The only reason anyone even looked at fixing performance in Stellaris is because the outrage was so big it would be unfeasible for them to just leave it as it is and not do anything about it.
I'll just make a wild guess and say that within the next year/year and a half PDS will announce Stellaris 2 and within the next 2-3 years we'll get to play the next game. Don't think there's any hope for any fundamental changes, Stellaris will keep getting cosmetic DLCs until second game is announced, then they'll just polish it up a bit and finally put it to rest.
- 1
- 1
- 1
- 1
Makes sense, but let me fix this slightly for you:
1. Wiz is not working on stellaris 2, he's working for a new engine for the next iteration of games, that will be able to handle 100's of thousands pops/entities.
2. Stellaris 2 will take 2 more years at least: If that engine was finished, we'd see it in CK, but of course no one knows....
3. You have to wait for CK to come out first before expecting any other major grand strategy/4x from paradox as a developer (not as a publisher)
- 1
Victoria 2 is already able to handle millions of pops as i have shown in an earlier post, so there is no need for a new engine. I do not think that changing to the vicky2 system would be easy, but i am quite sure that this would be far better than hoping for a fast Stellaris 2 with a completely new engine. I expect Stellaris 2 at the earliest in 6 years- but we might see a major update to Stellaris 3.0 in 1-2 years which tackles most of the current problems (AI, performance...) and remakes the game.
Why do i think so? Paradox games are made for longevity. People do not only buy the basegame, but invest huge amounts of money in the DLCs. If paradox suddenly discontinues a well established game half in the lifecycle and replaces it with a new version, many people (includeing me) would be rightly pissed because the money spend on the DLCs would be lost. Sure, there are ways to alleviate the pain, e.g. by including most features of the old game in the new game and by giving owners of the old game & dlcs huge discounts.
Otherwise, the main target group for the new game might decide to nope the new game and buy the complete edition in a later sale for 10 bucks.
Why do i think so? Paradox games are made for longevity. People do not only buy the basegame, but invest huge amounts of money in the DLCs. If paradox suddenly discontinues a well established game half in the lifecycle and replaces it with a new version, many people (includeing me) would be rightly pissed because the money spend on the DLCs would be lost. Sure, there are ways to alleviate the pain, e.g. by including most features of the old game in the new game and by giving owners of the old game & dlcs huge discounts.
Otherwise, the main target group for the new game might decide to nope the new game and buy the complete edition in a later sale for 10 bucks.
- 2
- 1
- 1
Distant Worlds: Universe is able to handle billions of "Pops". You have realistic figures like 8 billion PER SPECIES on one planet (depending on the size of the planet).
Granted, I don't know if a single one-digit-pop has any semblance of "free will" (like a Pop in Stellaris has various modifiers to resemble a living being).
But they are getting shipped around in batches or individually, so. If we go by raw numbers. Billions, if not trillions, of people are possible. If we need to add depth to those Pops/people, that might be an issue. I can't imagine DWU simulating ethics for a SINGLE person (one among, say, 5 billion on one planet). That would be amazing, but insane as well.
Also: Every species grows seperately with their own +% (depending on various factors; amount already present, planet quality, developed culture, etc).
Granted, I don't know if a single one-digit-pop has any semblance of "free will" (like a Pop in Stellaris has various modifiers to resemble a living being).
But they are getting shipped around in batches or individually, so. If we go by raw numbers. Billions, if not trillions, of people are possible. If we need to add depth to those Pops/people, that might be an issue. I can't imagine DWU simulating ethics for a SINGLE person (one among, say, 5 billion on one planet). That would be amazing, but insane as well.
Also: Every species grows seperately with their own +% (depending on various factors; amount already present, planet quality, developed culture, etc).
- 1
I just want to say that this is a powerful and constructive conversation for 67 pages long! I like it.
It seems, even after the update, Factions still require way too much processing power.
So i run a pretty heavily modded game - By midgame im stuck with, at best, 20-30 fps and at worst 1fps depending on what i am looking at or what i have selected. Yet, despite how bloated my game is, it still never crashes. I could use the "Grow_pops" command to generate 2000 pops on the birch world i have in the center of the galaxy (Gigastructural Engeneering mod) and...well, it would take about 20-30 minutes of the game being stuck in a "not responding" state, but it would do it eventually. My point being that despite all the mods and despite there being litterally tens of thousands of pops in just my empire alone, it doesn't ever crash. Most events pop up and go by as normal, including megastructure events or even ending wars and gaining a looooad of territory isnt a problem, and usually only takes a few seconds, a minute at most, to load through it all.
However, as soon as i click on the factions screen, the game keeps freezing up intermittently and as soon as i click on a faction it instantly crashes the game. No big deal right? Not like factions are that much of a big game mechanic, its not a core thing. Well, it IS a big deal when another empire decides to embrace a faction and thus change its ethics etc - The game can freeze for upwards of an hour if this happens and its not even in my control so its not like i can avoid it.
My point is this - Factions still require way too much processing power and even when other major events like a 40 million fleet power contingency crisis spawning in DOESN'T crash my game, but factions DO, you know there is a problem.
So i run a pretty heavily modded game - By midgame im stuck with, at best, 20-30 fps and at worst 1fps depending on what i am looking at or what i have selected. Yet, despite how bloated my game is, it still never crashes. I could use the "Grow_pops" command to generate 2000 pops on the birch world i have in the center of the galaxy (Gigastructural Engeneering mod) and...well, it would take about 20-30 minutes of the game being stuck in a "not responding" state, but it would do it eventually. My point being that despite all the mods and despite there being litterally tens of thousands of pops in just my empire alone, it doesn't ever crash. Most events pop up and go by as normal, including megastructure events or even ending wars and gaining a looooad of territory isnt a problem, and usually only takes a few seconds, a minute at most, to load through it all.
However, as soon as i click on the factions screen, the game keeps freezing up intermittently and as soon as i click on a faction it instantly crashes the game. No big deal right? Not like factions are that much of a big game mechanic, its not a core thing. Well, it IS a big deal when another empire decides to embrace a faction and thus change its ethics etc - The game can freeze for upwards of an hour if this happens and its not even in my control so its not like i can avoid it.
My point is this - Factions still require way too much processing power and even when other major events like a 40 million fleet power contingency crisis spawning in DOESN'T crash my game, but factions DO, you know there is a problem.
- 2
- 1
I'm a big fan of Stellaris and I think the game is simply ingenious. For me it surpasses the old classics like Ascendency and Master of Orion.
I don't think your theory @Rosthun is far-fetched. I've played Stellaris for almost 1000 hours now and I've never seen a game with so many bugs and such central performance problems. From a software development point of view, without knowing the insights, it seems that the craftmanship of software development is not at a common level (No unit test, no Performance Test, etc., missing QA standards). And also the strategy of how they got forward until now, supports your theory very much.
I just think it's a pity, I would be great looking forward to a second part. But if the quality of the game stays like this, I will think very carefully if I will buy a 2nd part. My wish would rather be, to give the version 1 a real boost again. Fix the bugs and solve the performance problems and tell openly that you are willing to produce a second version. With a well working game I can sweeten the waiting time and collect anticipation. Otherwise it would be logical that the community would play something else and if we return is not sure.
For my part I hope that our wishes for a stable game will be taken up and that they will be given a high priority.
- 1
- 1
Think about what's happening with a pop and factions.
A pop is just a row in a spreadsheet with a series of datapoints appended to its ID as an example:
When a pop is just sitting on a colony not doing much it consumes next to no processing power.
It's only when the game has to update pop-related equations (like planetary happiness) or calls on all/some of the pops within a given scope that it needs to cycle through them. And it seems to do this pretty often.
E.g. if you abolish slavery it will alter the status of all enslaved pops in your empire. so the game sets a nation wide scope (search filter) and runs a filter for enslaved pops then alters that entry. if you've got 10000 pops, that means it's going to need to filter out the unenslaved ones (e.g. 4000) then run through the remaining 6000 and set them to be free citizens.
Then it needs to re-calculate their happiness by removing the Enslaved = yes and adding Was enslaved=yes modifiers in the happiness equation before finally running the planet-wide average happiness calculation, which feeds into planet-wide stability, and probably a bunch of other blocks of checks too.
That's an awful lot of crap to crunch through and it cannot be easily multithreaded as it's a serial operation which relies on a very strict order of events that all need to be done in order before the next step in the web can be computed.
Most of the game's economic simulation seems to be built like this, with lots of bottlenecks, and so it requires not many CPU cores, but high frequency CPU cores as the name of the game is speed, not parallelisation.
Someone above mentioned an AMD threadripper probably not doing much. He's right, its like $3k for a 64 core CPU running at 4.3ghz (AT BOOST), most of the time 60 of those cores will be sitting around doing bugger all, whilst they wait for cores 0-3 to finish their current operations. Whereas my old i7-4790k which has 4 physical cores and runs at 4.5ghz will probably perform as well if not better (there are other factors like CPU Lx cache sizes and bus throughput that make this a less clean-cut comparison than just clock frequency, but you get the idea).
This is also why programs like Excel can chug hard and lock up when you perform concurrent operations, e.g. calculating sharpe ratios (a financial measure of risk-adjusted returns) for a portfolio of 10000 stocks in excel can literally grind the thing to a halt whilst it crunches numbers - but it has excellent multi-threading so this is less common (and it is common practice to design your macros efficiently and only update them on-demand where possible, to further mitigate this).
When your pc locks up for 20 mins as it generates 2k pops, it's doing something similar, it's generating 2000 more rows, writing them out and populating them serially (likely by comparing them against your existing species/governing ethics etc. via in-built filters/checks one at a time before it records any specific datapoint for the pop in question)
Last edited:
- 2
- 1