Game needs optimization

[sekelsenmat]

Wish I could upvote the OP twice. Having to play with minimal graphics settings is a hindrance since I can't see where forts are, among other things.

If changing the graphics settings improves things for you it indicates that your graphics card sucks. Graphics are run on the GPU which is an independent processor.

In my computers lowering graphics doesn't change anything, there is clearly too much CPU usage, not GPU.

 

[Zauberelefant]

Well, true that they probably did some improvements, but every new large feature (think Garrison or fuel) or refactoring has the potential of making software slower if the implementation is suboptimal. So it is not necessaryly getting faster, I'd say it is getting slower.

True that little constructive discussion is possible, but maybe Paradox will notice the complaints and assign resources to it, which is a gain for gamers isn't it?

Someone volunteered to check if less divisions helps, this can be used to mod the game as a workaround, but even if the test shows division count is partly to blame, it is still just a symptom, the real reason is N (unknown number) suboptimal data structures or lack of cache somewhere related to a feature involving divisions. I think that a properly optimized solution should be able to handle millons of divisions.

Also sometimes Devs answer to threads like this with interesting comments

I think optimization can only get you so far and at some point, you need brute calculating power.

 

[sekelsenmat]

I think optimization can only get you so far and at some point, you need brute calculating power.

It is true that it could theorically be the case, but I doubt that this game is close to the limit of optimization.

Also often optimizing leads to dramatic improvements in performance. An index in a database with a huge table for example can cause a query to go from taking days to taking nanoseconds.

Often you trade more RAM or HD usage for less CPU usage.

 

[GSP Jr]

One thing that always helps is to save and restart. Clears out things no longer needed and helps for a while.

 

[Zauberelefant]

One thing that always helps is to save and restart. Clears out things no longer needed and helps for a while.

Besides it not being healthy hanging around before the screen for hours on end.

 

[Zauberelefant]

It is true that it could theorically be the case, but I doubt that this game is close to the limit of optimization.

Also often optimizing leads to dramatic improvements in performance. An index in a database with a huge table for example can cause a query to go from taking days to taking nanoseconds.

Often you trade more RAM or HD usage for less CPU usage.

OK. Given the engine and the way the game is calculating a lot of stuff per tick, you feel there is a lot of potential speed available?

 

[Iskulya]

As I understand it from what the devs have said, one of the main bottlenecks is the way processor cores actually work. It's not really feasible, they say, to split a lot of the game's tasks between multiple cores as the latency issues would actually create a decrease in performance, thus the game's performance is strongly constrained by single core performance.

When you're dumping almost everything on just one core, there's only so fast it'll go. I'm not an engineer, I have no idea how difficult or feasible it would be to redesign the engine to be able to better take advantage of splitting tasks between multiple cores, but I can believe the devs at their word when they say it isn't easy.

 

[mursolini]

I'm just suggesting that the devs might want to consider improving performance since, although not as severe as Stellaris or CK2 pre-Reaper's Due, late game lag is real in HOI4. How am I supposed to know how Total War Warhammer 2 or CK2 managed to solve the problem.

CK2 doesn't really have a huge growth of character numbers and army numbers, because fundamentally it is limited to the number of counties, that is static.
Total war is absolutely a different game, but it tends to have far fewer pieces.

In HOI4, you start in world with very few units, and then see each major field x10-30 times the units in early game, and those have to be controlled due to having a war, so of course late game will be a lot slower than early game, no way around it.

more than a problem of the game as such it is an engine problem. The end game lag is classic of the Clausewitz Engine, it is the same in EU, Imperator, Stellaris. In HOI 4 it is probably more accentuated for some factors, but it is an endemic problem of their own engine that obviously cannot be solved in this generation, unless some technical bases are modified.

More like game design issue, with game constantly increasing the number of pieces it has to take care of.

As I understand it from what the devs have said, one of the main bottlenecks is the way processor cores actually work. It's not really feasible, they say, to split a lot of the game's tasks between multiple cores as the latency issues would actually create a decrease in performance, thus the game's performance is strongly constrained by single core performance.

When you're dumping almost everything on just one core, there's only so fast it'll go. I'm not an engineer, I have no idea how difficult or feasible it would be to redesign the engine to be able to better take advantage of splitting tasks between multiple cores, but I can believe the devs at their word when they say it isn't easy.

IDK, I got quite a good increase in performance, especially late game from migration from 4 physical core CPU to 8core CPU. I doubt there was that large of an increase in single core performance of Intel processors over the period.

 

[Hoi Neuling]

I play with Mods to. But not with such Mods which destroy implemented cool new Stuff the Devs or her Helpers bring in / brought in. Only Mods for better Historical Feeling, missing Speeches as well as News and such.

with the last complete Rework for Garissions and the World Map about 150 to 300 Devisions lesser are on the Map, which reduces the Lagg-Problem.

And if there is a Devisions-Limiter for the USA, Sowjetunion and such big country’s you reduce the Lagg-Problem as a second Step. That‘s an nice Feature from Expert-AI.

 

[STABBY5]

As I understand it from what the devs have said, one of the main bottlenecks is the way processor cores actually work. It's not really feasible, they say, to split a lot of the game's tasks between multiple cores as the latency issues would actually create a decrease in performance, thus the game's performance is strongly constrained by single core performance.

They say a lot of things. Other companies manage to do this effectively. At the end of the day its really just about them not wanting to invest the money.

 

[Zauberelefant]

They say a lot of things. Other companies manage to do this effectively. At the end of the day its really just about them not wanting to invest the money.

Because money would do what exactly?

 

[SophieX]

I just created a small mod to double the amount of infantry-equipment needed for all infantry and infantry-sub-units...
I'm curious to see what will happen...

Keep us appraised, will you?

I will play 3 or 4 games with different countries and then I'll give a report.

Sadly, due to personal reasons I will not be able to make this 3 or 4 game-runs in the near future.

For unknown rerasons I couldn't upload the the mod
So if anyone will try it, here are the 2 files and please read carefully:


Open C:\...\User\Documents\Paradox Interactive\Hearts of Iron\mod
Create a folder named: Inf_Equip_x2
Open the new Inf_Equip_x2-folder and create the folder: common
Open the new common-folder and create the folder: units
Now copy the file: infantry.txt in the units-folder.
Copy the file Inf_Equip_x2.txt in the mod-folder AND rename it as: Inf_Equip_x2.mod ( because here you are not allowed to attache a xyz.mod - file )

 

[Mousetick]

For unknown rerasons I couldn't upload the the mod

Only certain types of files are allowed to be uploaded to the forum, and .mod is not one of them.
You could pack your mod into a .zip file and upload that instead. Zip is allowed.
That should make installation of your experimental mod easier and less error-prone.

 

[FindFloppies]

Hmm. Try playing HOI4 on a computer with 32 GB of RAM and an SSD, and you can say 'what lag?'.

 

[STABBY5]

Because money would do what exactly?

Obtain goods or services? Even if they aren't hiring/contracting additional skill programmers, there's an opportunity cost to dedicating people to anything that is not directly making games to sell.

 

[Zauberelefant]

Obtain goods or services? Even if they aren't hiring/contracting additional skill programmers, there's an opportunity cost to dedicating people to anything that is not directly making games to sell.

Let's get real here, assume PDX put up money for the hoi4 Team - what's the money used for and how would that impact Performance exactly?
Because one thing i've learned is that money does not always solve issues.

 

[Hoi Neuling]

Hmm. Try playing HOI4 on a computer with 32 GB of RAM and an SSD, and you can say 'what lag?'.

Correct. Better you can´t say it. A medium high End or High End PC is the Stage of the Art. For about 1.500 Bucks to 1.800 Bucks you can get a very good Computer incl. Graphics Card from the prevouis Generation for the next 10 to 12 Years or longer incl. Win 10.

Not the Games are the Problem, it´s the overolded Laptop / PC with 20 Years Lifetime and Windows XP etc. which is doing the Laggs.

 

[kiamhiung]

is SSD really improve the slowdown ? and RAM ?
i thought CPU is the bottleneck ?

 

[Mousetick]

is SSD really improve the slowdown ? and RAM ?

Not really.

i thought CPU is the bottleneck ?

Yes primarily. In some cases, the GPU may be the bottleneck, depending on system's specs.

 

[Crispin]

is SSD really improve the slowdown ? and RAM ?
i thought CPU is the bottleneck ?

Not much, it's mainly the cpu. But it happens regardless of cpu in the lategame.