Elimination of 40% of game lags for all PCs. There is a solution!

[PakhomEpifan]

Hello to all.

I am Russian, and I don’t know English, so I write through Google translate. Therefore, I apologize to everyone for my clumsy language.

3 weeks ago, on one Russian forum, we conducted a performance test of one mod on Stellaris. The goal is to test the effectiveness of the mod:
https://steamcommunity.com/sharedfiles/filedetails/?id=1677564824

The test procedure was as follows:
_______________________________________________________________________________________

1. Indicate the characteristics of the computer
2. Use Stellaris version 2.3.3.
3. Run the save:
https://drive.google.com/file/d/1pVQbs3FpNpYh21zYHg-EgxekaANHIi4O/view
4. Drive one year at maximum speed and use a stopwatch to measure the time for which a year will pass at maximum speed.
5 Then start the game with the mod of the respected forum member @HAD and repeat the procedure No. 4:
https://steamcommunity.com/sharedfiles/filedetails/?id=1677564824
6. Put everything here.
7. It is not necessary, but it is desirable to do the scrolling at the maximum removal of the card.

Saving DO NOT NEED TO UNARCH! IT'S SIMPLY THROWN THROUGH THE WAY (yours may also differ slightly):
C: \ Users \ UserName \ Documents \ Paradox Interactive \ Stellaris \ save games \ _2117361974 (BUT THERE MAY BE OTHER DIGITS)
________________________________________________________________________________________

The results stunned us. For clarity, we have reduced all the results to this graph.

A little explanation:
1) comrades with laptop versions are marked with an asterisk;
2) four / six-core configs are hatched, and with SSD - circled in bold;
3) time without mod and profit from the mod are marked additionally.
4) Vertical - usernames.
5) Horizontal - time in seconds
6) The far left side of each user's rectangle is time with a mod. The far right side of the rectangle is time without fashion.
The% value with a minus sign means "how much less time was needed for the game year to go with the mod"

 

[PakhomEpifan]

Yes, your eyes do not deceive you. ALL users received 40% faster playing time.

 

[PakhomEpifan]

Incidentally, a similar fix (reducing POP checks for the possibility of a job change) was in the Glavius modification. But he believed that the Paradoxes corrected this sin, and removed this correction from version 2.3.

 

[SectorsAreOkay]

1) comrades with laptop versions are marked with an asterisk;

Appropriate.

 

[Major Malfunction]

 

[Major Malfunction]

Okay heres a comparison of fleet movements at the same save
1. No mod

Spoiler

2. With the mod

Spoiler

Same speed, same save, same everything. The best part is that it seems this mod works with ironman mode

 

[mergele]

Everything works with ironman, the question is does it work with achievements?

 

[Major Malfunction]

1 month = 11 sec, from 1 month=14 sec

 

[Gromit]

Glad to see others were testing other stuff out too. I'm still using EDAI here but I'll give this one a look as well to see how it compares.

These guys used the same 2450 save file we were using in this thread so for those of you who did testing with it, you can quickly compare your results.
https://forum.paradoxplaza.com/forum/index.php?threads/unplayable-late-game-lag.1215903

 

[Larknok1]

Ran this for a year without the mod and with the mod in my end-game ironman save with 10,000 pops in the galaxy.

1 year without the mod: 1 minute, 59 seconds, 59 ms.

1 year with the mod: 1 minute, 58 seconds, 44 ms.

Almost no change for me, not sure why though.

 

[intently]

In the comments on the mod page it looks like the mod will prevent pops from promoting into new strata, is that right? Is that a big or small problem?

AlphaAsh Mar 9 @ 7:13am

*wave* If this does what I think it does, the only issue I've found with checking against whether pops have a certain job already, is that it can cause problems with the AI promoting workers into the specialist strata.

Other than that, I can confirm that short-circuiting the daily parse of a job's weightings does improve daily tick performance slightly. Ultimately it depends on whether the weight clause can in fact be short-circuited, or whether the entire clause is parsed regardless of whether an early condition is passed.

had [author] Mar 9 @ 7:19am

Yeah, you need to grow new pops to fill new jobs.
Not perfect but i dont see better options.​

 

[PakhomEpifan]

In the comments on the mod page it looks like the mod will prevent pops from promoting into new strata, is that right? Is that a big or small problem?

AlphaAsh Mar 9 @ 7:13am

*wave* If this does what I think it does, the only issue I've found with checking against whether pops have a certain job already, is that it can cause problems with the AI promoting workers into the specialist strata.

Other than that, I can confirm that short-circuiting the daily parse of a job's weightings does improve daily tick performance slightly. Ultimately it depends on whether the weight clause can in fact be short-circuited, or whether the entire clause is parsed regardless of whether an early condition is passed.

had [author] Mar 9 @ 7:19am

Yeah, you need to grow new pops to fill new jobs.
Not perfect but i dont see better options.​

I played with the mod for 150 years. This is not a problem for AI.
AI has other problems)

The mod makes daily checks for changing jobs only for some slaves, unemployed, etc.

AI waits until the upper strata are filled, and then, as necessary, creates new jobs.

I have never had an AI fall into economic stagnation because of this mod. Economic stagnation was - well, purely the same as in the original Stellaris.

 

[PakhomEpifan]

Ran this for a year without the mod and with the mod in my end-game ironman save with 10,000 pops in the galaxy.

1 year without the mod: 1 minute, 59 seconds, 59 ms.

1 year with the mod: 1 minute, 58 seconds, 44 ms.

Almost no change for me, not sure why though.

Can you give me this save so I can test it for myself?

Which version? What are the mods?

 

[AlphaAsh]

Just a point of clarification: had's mod doesn't do anything EDAI doesn't already do. He came to the same conclusions I was already implementing in AlphaMod (and Stay Down Pit), which were later included in EDAI.

Yes, had's mod the OP is pimping will provide the performance improvement, sans all the other stuff EDAI does, like make AIs not tank their economies in early-game.

Choice. It's a good thing.

 

[Fourthspartan56]

Just a point of clarification: had's mod doesn't do anything EDAI doesn't already do. He came to the same conclusions I was already implementing in AlphaMod, which were later included in EDAI.

Yes, had's mod the OP is pimping will provide the performance improvement, sans all the other stuff EDAI does, like make AIs not tank their economies in early-game.

Choice. It's a good thing.

Huh, interesting.

And does EDAI have a similar performance increase?

 

[AlphaAsh]

Huh, interesting.

And does EDAI have a similar performance increase?

Yes. However, it does nab a tiny bit back for the monthly health checks (one planet per empire), which is where the economic fixings happen. Also had's mod won't reduce unemployment in end-game, which itself can lag up the game some, no matter what. There's plenty of hard-code processing of pops happening, other than just weight block evaluations for job suitability. Less pops means less of that as well, which is what EDAI provides some of for late-game.

 

[Fourthspartan56]

Yes.

Excellent, I will download it then

 

[AlphaAsh]

It’s not a matter of what fellow creators of mods are.

The fact is that a significant part of the lags is eliminated by a mod that can be done in an hour.

And the developers are not going to fix it.

It's not that simple. EDAI took a lot longer than an hour, and tackles AI competency as well as game performance, which a lot of players also care about.

Surely there must be some reason they haven't done it?

I don't find it plausible that no one involved in development sincerely cares about the player experience, for pragmatic or altruistic reasons.

It was a lot less work to integrate GAI than fix it themselves. Except GAI didn't fix it either (IMO, obviously).

 

[PakhomEpifan]

Just a point of clarification: had's mod doesn't do anything EDAI doesn't already do. He came to the same conclusions I was already implementing in AlphaMod (and Stay Down Pit), which were later included in EDAI.

Yes, had's mod the OP is pimping will provide the performance improvement, sans all the other stuff EDAI does, like make AIs not tank their economies in early-game.

Choice. It's a good thing.

Is your mod compatible with saves?

If so, then I will test it according to the described methodology.

 

[AlphaAsh]

Is your mod compatible with saves?

If so, then I will test it according to the described methodology.

Risky. The health-checker is probably going to have to do a lot more work initially on a save, depending on how far it's progressed and how broken AI economies are.