Specs & Optimization

[Cri11e]

EU5 is looking great and I am excited to experience it. But I do have one concern...
How well low-medium performance computers can handle it.

It is well known to those who love paradox games that optimization is usually a problem.


Does anyone have any information or thoughts about this?

 

[JacceDON]

Why would a brand new game run perfectly on low-medium performance computers? I get that it isn't GTA VI, but the game is still complex and should require at least a medium-high performance machine to run.

 

[Dummkopf1]

Why would a brand new game run perfectly on low-medium performance computers? I get that it isn't GTA VI, but the game is still complex and should require at least a medium-high performance machine to run.

That's what most people have. Steam does hardware surveys so some information is available (I can't link it but a quick search will bring it up). A company has an incentive to make a game that can be run on the most hardware. Game design obviously takes precedence and will determine the floor, to a point, as there is no point in making a game only 10% of people can run.

For example, it would be ideal to optimize the game for 16gb of ram instead of 32gb, that way it covers 86% of potential players. If you look at Victoria 3 recommended specs, they're pretty much around the average. I don't see why they would deviate from the average for this release.

 

[Diamond2323]

EU5 is looking great and I am excited to experience it. But I do have one concern...
How well low-medium performance computers can handle it.

It is well known to those who love paradox games that optimization is usually a problem.


Does anyone have any information or thoughts about this?

Current performance is dismal as has been confirmed by content creators with access to the game, but that has no bearing on how the final result will be as optimisation is usually one of the last things to be worked on, as it makes no sense to optimise a feature that may or may not make it to the final product.

I wouldn't worry about it too much, Imperator runs just fine and that game is the most similar to EUV that we have seen. I think that with the added mechanics and complexity compared to EUIV, you'll be playing on slower speeds more frequently compared to long periods of speed 5, so you may feel hardware limitations less.

 

[Cri11e]

Current performance is dismal as has been confirmed by content creators with access to the game, but that has no bearing on how the final result will be as optimisation is usually one of the last things to be worked on, as it makes no sense to optimise a feature that may or may not make it to the final product.

I wouldn't worry about it too much, Imperator runs just fine and that game is the most similar to EUV that we have seen. I think that with the added mechanics and complexity compared to EUIV, you'll be playing on slower speeds more frequently compared to long periods of speed 5, so you may feel hardware limitations less.

In my limited understanding of game development, I've come to understand that important things need to be considered from the beginning to be well integrated as a core goal.

This would ring true for proactive performance considerations as well?

 

[Sirbab]

In my limited understanding of game development, I've come to understand that important things need to be considered from the beginning to be well integrated as a core goal.

This would ring true for proactive performance considerations as well?

I don't think that tracks, or at least, not in the end. Sure if you start out without planning to multithread on a project that needs it, yeah you're screwed. But from my limited understanding of game dev, how can you start optimizing in earnest before you know what all is even making it into the final release? There's also hardware that changes among consumers while the game is being developed, and that can change the goal posts.

 

[Sindai]

Obviously you do things like use faster algorithms instead of slower ones for known problems like pathfinding when you first put the game together, but just like it's impossible to balance a game on paper before having a bunch of people play it and seeing what breaks, it's impossible to optimize a game before being able to run it with profiling and seeing what's taking the most CPU cycles.

And of course, there's the fact that when dozens of people work on a game for years they make thousands of mistakes in the process, some of which will cause performance issues.

 

[solamyas]

In my limited understanding of game development, I've come to understand that important things need to be considered from the beginning to be well integrated as a core goal.

This would ring true for proactive performance considerations as well?

You can't optimize a game while actively developing it. Optimization is the last phase

 

[BI0ck]

In my limited understanding of game development, I've come to understand that important things need to be considered from the beginning to be well integrated as a core goal.

This would ring true for proactive performance considerations as well?

Imagine, lets say, a hypothetical new feature requires 10 hours to be created.

Lets assume that optimizing it while creating it takes about 1 hour, 10% of the time taken to create it. Once the feature is created, it has a chance of being scrapped, or significantly reworked. If either of those things happen, the time you spent optimizing the feature was wasted. Even if the feature isn't scrapped, and isn't reworked, the feature might not be performance-critical.

Performance, especially in GSGs which nowadays are multithreaded, tend to have bottlenecking problems, where performance is limited by the slowest feature. This means that if, for example, army calculations takes 50ms longer than anything else per tick, time spent optimizing navy calculations is wasted until the army is optimized, as both will be running separately on different threads. Even if optimization takes longer per feature at the end, due to having to revisit older code or whatnot, it is faster overall to save optimization towards the end, that way you can measure and actually optimize the features that need optimization, rather than optimize everything, even the features which won't run slow enough to matter.

 

[livebythecreed77]

EU5 is looking great and I am excited to experience it. But I do have one concern...
How well low-medium performance computers can handle it.

It is well known to those who love paradox games that optimization is usually a problem.


Does anyone have any information or thoughts about this?

Stop openning threads that have duplicates

Post Announcement Feedback MEGATHREAD

Hi All, I have seen lots of duplicate threads so I compiled a list of most if not all threads into categories so be good to check against this before you post a new thread. Its also good for Devs incase they missed some feedback, and ensures no...

forum.paradoxplaza.com

I have a megathread tracking all posts since the annoucement by category (I still need to break down general more) there are already 3 performance related threads. Post there.

Performance:
https://forum.paradoxplaza.com/forum/threads/processor-advice-for-eu5.1746917/

https://forum.paradoxplaza.com/forum/threads/performance-concerns.1742963/

https://forum.paradoxplaza.com/foru...layerbase-concerns-about-performance.1743486/

 

[Dennis [UA]]

Just hope for someone to redo the Toaster Universalis Mod for EU5

 

[GeneralNoob]

One of the devs once commented on that they are going for optimization on a medium range PC (about 1000-1500 Euro). So if you are in that range, you should be fine. Dont ask which thread it was, I dont remember.

 

[Cri11e]

Obviously you do things like use faster algorithms instead of slower ones for known problems like pathfinding when you first put the game together, but just like it's impossible to balance a game on paper before having a bunch of people play it and seeing what breaks, it's impossible to optimize a game before being able to run it with profiling and seeing what's taking the most CPU cycles.

And of course, there's the fact that when dozens of people work on a game for years they make thousands of mistakes in the process, some of which will cause performance issues.

It is neither possible nor feasible to be proactive about a game's performance, it has to be done primarily retroactively?

This seems at least so far to be the unofficial majority attitude of this thread.

I would like to think that the game industry has more evolved best practices, perhaps I am wrong.
There has to be some kind of reference points for structuring new core systems, especially in the Paradox ecosystem which has so much overlap.