Stellaris Dev Diary #181 : Threading and Loading Times

[klopkr]

Optimizations are good and all. I just wish you guys would start fixing the glaring unfun bugs in this game so I even bother launching the game.

 

[Apollo1784]

Is this an out of season paradox summer vacation joke?

 

[Axe99]

Cheers for the DD Mat, and great work on the optimisations I'm sure you're looking at these things all the time, but if there's any chance of continued work on optimising the pace of the game during play? I enjoy the first 150 years immensely, but after that it's so slow that it's literally "read a book while you play" because the ratio of waiting : playing has got too high to maintain engagement (and a watched galaxy never boils.....).

My PC isn't the fastest (a teensy bit below recommended specs with the CPU, a Ryzen 5 1600 - GPU and RAM are all above, but I've got no doubt it's the CPU holding it back), but I can happily play HoI4 through to the 1950s, and EU4 through to the end, and CK2 up until the last hundred years or so (that also got bogged down hard towards the end, at least about a year ago), and Imperator Rome is also quick enough to maintain interest. Vicky 2, of course, runs very well .

What's confused me particularly is that there's not a huge impact on galaxy size as to the speed and slowdown - I tried going small, and hit the same issues (it was slightly faster, but nowhere near as much as I'd have expected given the expected differences in terms of data processing requirements).

No biggy if not possible - I'll upgrade my PC in the next 12 months and get back into Stellaris then if it can't go any quicker - but from a hardware perspective it feels like the least accessible from a hardware perspective (ie, needs most expensive hardware) of any of PDS' games to enjoy.

 

[Hawawaa]

This will be most welcomed update. 86 mods and 5 minute start up going to shrinked down is a welcomed change. Thank you!

 

[AndrewT]

Who's Jeff?

Starbase Jeff

Starbase Jeff is a pipe-laying game where players maximize their profits by strategically placing and building onto sections of a starbase. Whoever makes the most money by the end of a given number of rounds wins the game. Any number of rounds can be played with each round lasting...

boardgamegeek.com

 

[Scottbert]

Thank you so much!

 

[Annihilat0r]

I don't understand people on this forum.

First, they complain that there hasn't been any DD or official communication in months (rightly so I think). People said, if PDX can't show any new content yet, they could at least write about something boring or technical, just to show that they are working on the game and engage with the community.

Now we have a DD doing just that, Mat even sticked around answering questions (thank you @MatRopert ), and now people complain that the topic of this DD was not important for them and the devs should have worked on something else.

Also, keep in mind that Mat said that they are still in home office and I can say from experience that the complete lack of physical meetings and interactions with your colleagues over such a long period of time definitely takes a toll on productivity.

I too was complaining about the lack of communication, so if any devs read this, plaese don't be discouraged by the always unhappy and keep DDs like this coming, until the marketing people give you the "go" for the juicy stuff.

 

[Troyen]

My experience with modding another game suggests startup times would help out modders as well. Don't know if you can change files in this game while it's running, but if not, it's a lot of game startups to play and test the results of your mod changes and time saved there is appreciated.

It sounded from the Dev Diary that only Windows users with DX11 would see the benefit, but from the blog it seems like all platforms would benefit from at least the changes to the file loading library. Will be interesting to see how the overall set of changes translates to mac - though with my computer I spend more time loading save games than the game itself.

 

[CrazyJ]

I don't understand people on this forum.

First, they complain that there hasn't been any DD or official communication in months (rightly so I think). People said, if PDX can't show any new content yet, they could at least write about something boring or technical, just to show that they are working on the game and engage with the community.

Now we have a DD doing just that, Mat even sticked around answering questions (thank you @MatRopert ), and now people complain that the topic of this DD was not important for them and the devs should have worked on something else.

Also, keep in mind that Mat said that they are still in home office and I can say from experience that the complete lack of physical meetings and interactions with your colleagues over such a long period of time definitely takes a toll on productivity.

I too was complaining about the lack of communication, so if any devs read this, plaese don't be discouraged by the always unhappy and keep DDs like this coming, until the marketing people give you the "go" for the juicy stuff.

you forget that stellaris fans are the kind of people who can spend a hundred hours on a game and then call it unplayable.

 

[LocutusvonBorg]

i am a bit curious regarding the production problem.
why do they even need to read the intial value ? just process the productions of the pops on the planet in order of the id of the pops and the planets are being calculated bei their id. or leave the planets out, the pops first of the whole empire in their id-order and then planets afterwards. this can be parallized. in the end each res is summed up and >>then<< the initial value is taken.
can there be negative values ? yeah, sure. but no problem, cause everything now effected cause of the negative values takes effect in the next cycle. but that is still all the pops of a planet, or the other method. this way its not getting rediculous to get the initial values for thousands of pops in a lategame and instead just one time per res, per empire, which should reduce it dramatically.

so, is this thought way to simple ??

 

[Thoughtful Gaming LP]

I really like that there are resources invested into rather less prominent parts of the game that are still very important.

Really understand the people complaining about the bugs and odd ends that are there still in Stellaris, some things need at least a clean sweep, it's just little things but really a bunch of them, like the science ship curator technologies and/or the inability to upgrade / customize your own science ships, or constructors, that's just one bit. It doesn't make the game unfun or unplayable, but it's just annoying everytime you see this, salting the experience just that one bit, like taking 5% of your Stellaris happiness. These things are everywhere, and you notice them.
It's kind of a thankless, less flashy work, but would make many veterans happy, probably influencing them into recommending the game to other players.

That said thank you at PDX for the nice improvement, it will make many people glad, especially the mod lovers

All the best, Immanuel Can

 

[AlknicTeos]

oh, Victoria III confirmed!

 

[Foxosaur]

Reading between the lines the article is wrote as though it was just something on the side he found, experimented with and found some wildly unexpected success. Rather than Paradox strategically invested some time/money. So nice one on that front. I hope may take it to the next level, if its good use for the better part of the game.

 

[Tamwin5]

i am a bit curious regarding the production problem.
why do they even need to read the intial value ? just process the productions of the pops on the planet in order of the id of the pops and the planets are being calculated bei their id. or leave the planets out, the pops first of the whole empire in their id-order and then planets afterwards. this can be parallized. in the end each res is summed up and >>then<< the initial value is taken.
can there be negative values ? yeah, sure. but no problem, cause everything now effected cause of the negative values takes effect in the next cycle. but that is still all the pops of a planet, or the other method. this way its not getting rediculous to get the initial values for thousands of pops in a lategame and instead just one time per res, per empire, which should reduce it dramatically.

so, is this thought way to simple ??

You'd just go from having to calculate the changes to the main stockpile to calculating changes to the "buffer" of net resources gained/lost. All your suggestion would do is add an extra step. What needs to happen is basically just adding up a bunch of numbers. It doesn't matter if your initial value is added at the beginning or the end, it wouldn't reduce the amount of calculations that would need to be done.

What could theoretically work is some way of dividing the load, say calculating each planet or sector individually and then going through and adding the combined changes. Dividing based on jobs wouldn't work because jobs have a LOT of overlap in terms of resources used, If a miner and artisan and scientist all tried to process their production simultaneously it would break.

 

[Pellaken]

The dev diary seems to be missing something. it jumps right from "it is hard to fix the slowdown" to "we did this instead" without clarifying if they are continuing to work on fixing the slowdown, or have given up.

 

[Attitude Adjuster]

The dev diary seems to be missing something. it jumps right from "it is hard to fix the slowdown" to "we did this instead" without clarifying if they are continuing to work on fixing the slowdown, or have given up.

It has been mentioned that performance is an "ongoing project" that gets tinkered with every patch...

... To be made worse evidence would suggest.

 

[Pyzayt]

you forget that stellaris fans are the kind of people who can spend a hundred hours on a game and then call it unplayable.

This is far too true. I don't think Stellaris is unplayable, but I do feel 2.6 introduced a bug that made me not want to play (critical difference, the game is still playable, I just find that particular bug ruins my experience). With 800 hours into the game, I'm definitely rookie numbers compared to most, but it's long enough to notice this community is utterly unpleasable and will never be satisfied. Stellaris players always seem to need to find a reason to complain about what the devs are doing, and while I think that Stellaris could certainly be improved (and frankly does need more polish for its existing systems), the way people here go about it is... not okay.

We got radio silence, and we complained. So a member of PDX came and gave us radio contact, and we complained. They said they're looking at performance, and this little neat thing came out of it, and we complained. We demanded they fix systems, and they did, and we complained (specifically about it not being enough in this case, despite "enough" being subjective.) Frankly, every time this community asks for something, when we get it, people complain.

This is why fandom can't have nice things people. PDX has been absurdly tolerant of our crap, and seriously deserves a break from it all. I don't want to see the Stellaris dev team go the same way a lot of my favourite old things went; abandoned and despised by its creators for the people that came to call themselves fans.

Give them a bloody break; they're trying. Appreciate what we have, rather than always focusing on what we don't. Stop accentuating the negative, and just enjoy that we're getting improvements.

 

[CrazyJ]

i adore the game, and it is still playable, it's many glaring issues are the only reason i don't spend 10 hours a day on the game. the game is in a bad place at the moment. pretty much anyone will acknowledge it. honestly if the crisis actually worked and some of the worse bugs would be fixed i would be content, a game as vast as stellaris will never be bug free, that is a fact, out of every thousand bugs 999 of which are irritating at worst. it's the thousandth bug that gets you , there are more graphical glitches and minor issues than any sane man would care to learn about . those don't bother me, it's stuff like the awoken being broken since they were introduced that bugs me, seriously one of the coolest events in the game has been breathtakingly borked for years. and then there is the fleet manager which is a buggy mess(ghost fleet shenanigans ) . the ai is drunk, but since that is fixable via mods i really don't care. the crisis ai being borked however... that's a no-no that needs to be fixed. the sector ai is a mess and forces you to manage all worlds yourself but that can be mitigated via certain settings, and mods, if done correctly it is only a minor problem.
i try not to complain too much about minor problems but when major game features or important events are borked i am going to be annoyed. i think that is justified.

the biggest issues in my book are as follows: crisis ai, fleet manager problems, regular ai, major events being broken, micro issues( i shouldn't have to mod the hell out of the game to make micro bearable)
fix those and i will happily resume playing and buying dlc, but until then i can't bring myself to keep pouring money into a game that seems to only develop more problems with every patch

 

[Pellaken]

you forget that stellaris fans are the kind of people who can spend a hundred hours on a game and then call it unplayable.

let me tell you a story.

I've played HoI4, as a number of countries.
I have a bad tendency to restart after a few days, weeks, or months of gameplay for various reasons. Some will be detailed below. Some are as simple as "I realized playing as country X would be more fun"
If I only count countries I've played for more than 6 months, then I'd say I've only played as a major (GER, JAP, SOV, ENG, USA, ITA, FRA) a grand total of 15-20 times.

I've played Liberia at least 25 times.

Canada is another nation I've played a lot of. Given it's where I was born and live IRL perhaps that is not surprising. I want to tell you the events of about 2-3 weeks that happened to me a few years ago.

Game 1: Everything is going great. I'm doing my thing, happy! Not sure what to do with my army, but I shuffle them around the country. Then I notice that France is out of the war... but something about the Italian army in Paris? I pan over, and sure enough, Italy gobbled up the majority of france again, meaning the regular events of the fall of France do not fire.
Sigh.
I really would like to try to replicate the war. I really would like that event to fire. So time for

Game 2: Can't have this happen 2 times in a row, right? This time I figure out what to do with my army; africa! I'm going to take Italian east africa! Sure enough, war begins, and my armies march forth. I happen to pan up to europe and... wtf, AGAIN Italy is plunging deep into france.
ffs.
Okay. So lets do this again.
This time I'm going to move MY army to the France-Italy border.
Great. I do that, the war is declared, and I sit there and without clicking any units, just watch that front. Looks good. No breakthroughs. Wonderful. It does not catch my attention that I don't seem to be losing men or equipment until I pan back over and have Italian units behind me! How!!
...I forgot to declare war.
I FORGOT TO DECLARE WAR.

Game 3: Okay. This time I'm gonna do it right. Army in place, I declare war; great! Italy is being beat back. Except this time, France, Belgium, and the Netherlands manage to over-run Germany. I don't have a single unit in that battle. Germany gets over-run and the war 'ends'.
the hell. This is not what I wanted at all.

Game 4: Okay okay okay. This time I'm gonna do it. I've been trying to play this dang game for days now. Everything seems great, I even remember to declare war...
...I forgot to move my army to europe

Game 5: Army in place, everything seems great... except the soviets are on some totally bizzare alternate history path (this was before you could force them not to). Needless to say the USSR gets over-run in no time.
Wonderful. All I want is to play WW2 as Canada. None of this totally unrealistic nonsense.

Game 6: I forget to declare war again...

Game 7: Set the wrong option, and my army attacks instead of defends...

Game 8: I try to be clever and only send half my army. Italy breaks through.

Game 9: Czechs beat the germans (more accurately, hold them off so long France invades and wins)

After this I took a few days off from playing. return.

Game 10: forget to move my army.

Game 11: forget to declare war.


This is where I gave up.

I had an old dual core computer at the time. Very slow. I think all of this added about 30-60 hours to my playtime.

And I dunno about you, but this sounds pretty unplayable to me.


I've only had one (1) game of Stellaris I couldn't finish because of stupidity (one of my stations became invincible)
There are very few games of Stellaris I've abandoned because they got too annoying. Contrast this with HoI4 where about 5/6ths of all my attempts end in frustration.

My playtime?

Steam Community :: pellaken :: Games

steamcommunity.com

700 hours in each. 714 in Stellaris, 686 in HoI4.

edited to add:

Know why I have so many HoI4 stories and so few Stellaris stories?

Stellaris is good game. I enjoy it.
HoI4 is deeply flawed. I want to enjoy it.

HoI4 probably isn't for me.
And by the same token.
Stellaris isn't for everyone.


edited the following day to clarify/add:
the whole reason I had to put my army there was the failure of the game to do what it was, by its own definition, supposed to do, and be at least somewhat realistic. The bug, that I don't know if is fixed yet, is that as soon as germany declared war on France, all of France's armies on the Italian border would rush away, only to see that undefended border be over-run as soon as Italy declared war. 100% of my efforts were spent in trying to solve a bug.

edited the day after the last edit:
this post made me fire up HoI4 again to see how bad it is (its bad. when I tell each nation to go historic and save that preset; I become unable to open the preset. even just after saving it. on account of me not owning the DLCs I own)
anyway after selecting liberia, it occured to me people might want to know why on earth I'd ever play liberia. the answer is simple. I want to see if I can "be the first to drop a nuke, but without ever inventing electricity" or, in other words, develop "nuclear bombs" tech, and drop a nuclear bomb - be the first to drop a nuclear bomb - and do so without ever inventing the "electronic mechanical engineering" tech. Thus dropping the world's first nuke, without ever 'inventing electricity'

 

[비잔티움1453]

United Nations of Earth
Let the apes(Hominidae) exist as pre-sapient soecies on Earth.