Factorio Forums

noclaf wrote: Wed Apr 06, 2022 6:47 am I'm also aware of certain issues related to multithreading (parallelization)...but that's also reason why I just studied it but I work in another field.
And I didn't say I want to go unlimited. Just 5,4kSPM (where 60UPS starts to drop on my R7 5800X) seems to be pretty small and definitely below the computational power of the computer.

UPS is not related to graphics card. Important for UPS is the CPU ipc+clock as well as fast memory.
Edit: strike

Gummiente27 wrote: Sat Apr 09, 2022 2:27 pm UPS is not related to graphics card. Important for UPS is the CPU ipc+clock as well as fast memory.

R7 5800X is AMD CPU, not GPU. And memory at 3600MHz is also probably not a bottleneck.

There’s a while thread already on multithreading in which the devs have even given responses. Probably be best to talk about it there if their responses are insufficient?
viewtopic.php?f=5&t=39893

That was a whimper rather than a serious addition to ST/MT discussion. I think parallelization is not necessarily the only correct answer to this "problem" - but I really think the devs should look into it. If their customers are loosing interest in playing the game for (IMHO) understandable reasons...it's something they should address.
I mean other solutions could be to look at the most CPU/RAM hogging entities and help the players to reduce their numbers in the game. E.g.:

1) faster inserters
1a) new inserters
1b) more techs to research to increase their stats
1c) reintroduction of loaders as a late-game item?
1d) "double" inserters taking items from both sides of the belt w/o stupid dancing between the lanes

2) faster belts
2a) new belts
2b) introduction of tech to increase stats
2c) stackers if it can help? (see stacker in DSP)

3) trains
3a) again - more tech to increase their stats (both for locos AND wagon) thus reducing # of trains needed

4) overhaul of liquids
4a) I mean everybody just hates liquids anyway
4b) no nuclear power due to liquid UPS killing is IMO a huge drawback in the lategame - everybody HAS to go solar

5) production entities improvement
5a) again - new entities, new researches etc.
5b) tier 4 to X modules
5c) new recipes to be researched? (e.g. imagine how many entities/UPS/pain is going to be saved if there is a research to smelt steel directly from iron ore...Jesus...I hate steel smelting)
5d) ?allow "direct connection" between production facilities? i.e. copper smelter built directly next to copper wire assembler would be able to transfer the copper plates to the assembler really fast and w/o use of an inserter.

Most of the items / entities / researches above could be set-up as late game and expensive, but it might be the nextstep in the play where the players could leapfrog from 2-4kSPM bases to 10-50+kSPM. And I understand most (if not all) of the above is solved by various mods (Fluid must flow, Schall machines....) but we should use mods to improve vanilla, not to correct flaws in vanilla (+achievements for those who don't have them).

tl;dr
I think that current factory size limitation due to inevitable UPS drop is a mistake (I understand there always will be a limit, just that the limit is quite low currently) and there are IMHO two ways how to address it:
1) don't change the game mechanics but change how the game is programmed
2) keep the code design but introduce new improvements to the game mechanics through new researches and entities

noclaf wrote: Sun Apr 10, 2022 9:27 am That was a whimper rather than a serious addition to ST/MT discussion. I think parallelization is not necessarily the only correct answer to this "problem" - but I really think the devs should look into it. If their customers are loosing interest in playing the game for (IMHO) understandable reasons...it's something they should address.
I mean other solutions could be to look at the most CPU/RAM hogging entities and help the players to reduce their numbers in the game. E.g.:
[...]
Most of the items / entities / researches above could be set-up as late game and expensive, but it might be the nextstep in the play where the players could leapfrog from 2-4kSPM bases to 10-50+kSPM. And I understand most (if not all) of the above is solved by various mods (Fluid must flow, Schall machines....) but we should use mods to improve vanilla, not to correct flaws in vanilla (+achievements for those who don't have them).

tl;dr
I think that current factory size limitation due to inevitable UPS drop is a mistake (I understand there always will be a limit, just that the limit is quite low currently) and there are IMHO two ways how to address it:
1) don't change the game mechanics but change how the game is programmed
2) keep the code design but introduce new improvements to the game mechanics through new researches and entities

No. You have a fundamental misunderstand of what is good about Factorio. The production numbers shown on screen don't matter. The limit is not low. What matters is how many belts, inserters, trains and assemblers you have. Faster belts and inserters and assemblers reduce the amount of objects needed and is counter to what actually matters. Factorio is about the massive scale of the factory in search of higher production. Increasing the production numbers by upgrading the machines does not help for this goal.

The logistics puzzle of lots of materials is exciting. Seeing all the things you built is awesome.
Performance updates are good, they allow you to build more assemblers and belts. So that YOU can achieve more factory. Not so that the game can achieve higher numbers. Yes that also happens, but don't mistake that side effect with the actual source of joy just because they happens simultaneously.

noclaf wrote: Sun Apr 10, 2022 9:27 am 1) faster inserters
1a) new inserters
1b) more techs to research to increase their stats
1c) reintroduction of loaders as a late-game item?
1d) "double" inserters taking items from both sides of the belt w/o stupid dancing between the lanes

2) faster belts
2a) new belts
2b) introduction of tech to increase stats
2c) stackers if it can help? (see stacker in DSP)

3) trains
3a) again - more tech to increase their stats (both for locos AND wagon) thus reducing # of trains needed

4) overhaul of liquids
4a) I mean everybody just hates liquids anyway
4b) no nuclear power due to liquid UPS killing is IMO a huge drawback in the lategame - everybody HAS to go solar

5) production entities improvement
5a) again - new entities, new researches etc.
5b) tier 4 to X modules
5c) new recipes to be researched? (e.g. imagine how many entities/UPS/pain is going to be saved if there is a research to smelt steel directly from iron ore...Jesus...I hate steel smelting)
5d) ?allow "direct connection" between production facilities? i.e. copper smelter built directly next to copper wire assembler would be able to transfer the copper plates to the assembler really fast and w/o use of an inserter.

Most of these are modable and you can find mods for them on the mod portal. Try them (or write your own) and see if it helps. (It won't, production numbers may be higher, but the factory itself didn't grow.)

To my knowledge, modding 2b is currently impossible as there are currently no technologies affecting belts and 5d can only be faked by using e.g. a hidden inserter or maybe a mining drill, so you actually don't save on entity count.

I think parallelization is not necessarily the only correct answer to this "problem" - but I really think the devs should look into it.

Have you actually read the various FFF and dev posts about game optimization?

Qon wrote: Sun Apr 10, 2022 11:08 am No. You have a fundamental misunderstand of what is good about Factorio. The production numbers shown on screen don't matter. The limit is not low. What matters is how many belts, inserters, trains and assemblers you have. Faster belts and inserters and assemblers reduce the amount of objects needed and is counter to what actually matters. Factorio is about the massive scale of the factory in search of higher production. Increasing the production numbers by upgrading the machines does not help for this goal.

The logistics puzzle of lots of materials is exciting. Seeing all the things you built is awesome.
Performance updates are good, they allow you to build more assemblers and belts. So that YOU can achieve more factory. Not so that the game can achieve higher numbers. Yes that also happens, but don't mistake that side effect with the actual source of joy just because they happens simultaneously.

Thank you for telling me what's good about the game I like. I was never aware I like it for wrong reasons!!!
I have a neat factory with stable 5,4k; 190 mining efficiency; homemade blueprints as efficient as I was able to make them in order to decrease number of entities (12 beacons, consuming/producing as close to a full belt where possible etc.); no crazy main buses or ingredient drone transport. But I'm stuck - sure, here or there I might be able to do small improvements, but w/o total overdo of the whole factory I'm not able to improve it significantly.
What's in there for me? Spend 100 hours thinking, how belt weaving might be able to save me 1 out of 20 smelters and 2 out of 40 inserters? Going bigger is the only "fun" target ahead of me. And you are right more efficient entities won't make my factory larger - but at least I will produce more. Currently I'm not able to do any of the two.
And I know there are 30-40kSPM bases at 60UPS but...let's be honest they have nothing in common with regular gameplay and they are designed from the scratch for this one purpose (including stuff like not having the same item on both sides of belt to improve performance, setting up circuits to turn on certain entities only once in X ticks...). That's really cool and kudos to the creators, but that's no longer a fun for me.

Nidan wrote: Sun Apr 10, 2022 12:06 pm Most of these are modable and you can find mods for them on the mod portal. Try them (or write your own) and see if it helps. (It won't, production numbers may be higher, but the factory itself didn't grow.)

To my knowledge, modding 2b is currently impossible as there are currently no technologies affecting belts and 5d can only be faked by using e.g. a hidden inserter or maybe a mining drill, so you actually don't save on entity count.

I think parallelization is not necessarily the only correct answer to this "problem" - but I really think the devs should look into it.

Have you actually read the various FFF and dev posts about game optimization?

I know. But modding breaks your achievements (if you don't have them yet). Modding increases risk of cross-incompatible mods. Mods have to maintained/updated. Etc.
Yes, I know Factorio is already highly optimized and that any movement in other direction would require rewriting a huge portion of the code, bring new challenges, certain tasks are not possible to parallelize etc. That's why also said code-changes might not be the right call.

Anyway - I have a funny feeling. I'm a customer that paid for the game. And I HAVE a problem. While you are trying to convince me that I don't have a problem. This is rarely a working strategy. Sure, you might disregard my problem as irrelevant or call me stupid, because I don't "understand" why I should like Factorio...but the result will be a lost customer.

noclaf wrote: Sun Apr 10, 2022 12:46 pm Thank you for telling me what's good about the game I like. I was never aware I like it for wrong reasons!!!

Np, glad I could help.

This is also the official position, this is what kovarex and Wube think is great about Factorio.

noclaf wrote: Sun Apr 10, 2022 12:46 pm I know. But modding breaks your achievements (if you don't have them yet). Modding increases risk of cross-incompatible mods. Mods have to maintained/updated. Etc.
Yes, I know Factorio is already highly optimized and that any movement in other direction would require rewriting a huge portion of the code, bring new challenges, certain tasks are not possible to parallelize etc. That's why also said code-changes might not be the right call.

Install mods that do it anyways. Let go of your fears and enjoy Factorio the way you want to. The mods are there for you to have it your way. It is part of the game to mod it so it fits your playstyle. You can enjoy big numbers, do it.

noclaf wrote: Sun Apr 10, 2022 12:46 pm Anyway - I have a funny feeling. I'm a customer that paid for the game. And I HAVE a problem. While you are trying to convince me that I don't have a problem. This is rarely a working strategy. Sure, you might disregard my problem as irrelevant or call me stupid, because I don't "understand" why I should like Factorio...but the result will be a lost customer.

You have already paid for it. They can't abandon their whole plan and change the whole game to your specific whims. You can decide if you want to be a lost customer or if you want to play with mods. You can't hold Wube hostage with threats of not playing until they make it in your image.

I would like optimisations that group entities into bigger super-entities that require less CPU power, like they do now with belts and solar and accumulators. Would be cool if that could happen with whole ore mining outposts, smelting columns and nuclear power facilities etc if it could be done without losing simulation detail. That would unlock a lot more than parallelization ever could. And we could build bigger, with more production also.

noclaf wrote: Sun Apr 10, 2022 9:27 am I mean other solutions could be to look at the most CPU/RAM hogging entities and help the players to reduce their numbers in the game.

Yeah, I gotta agree with Qon here. To me, Factorio is not about the numbers themselves, and as such adding tier 4, 5, 6, etc of entities feels like the wrong way to go.

W00t, might be just a coincidence, but I have been recently getting very annoying lag when zooming in/out of the map view, and this issue seems to be gone now !

noclaf wrote: Mon Apr 04, 2022 5:12 pm [...]

Frankly for me...I feel kind of stuck. I have 5,4kSPM 60UPS factory, but I cannot enlarger it, because the UPS would fall (ok, I can flip 100% to direct train mining and smelting, but that's just postponing the inevitable). Or go for some mods - where I'm currently rather annoyed by Krastorio 2 complex production chain where one new product requires exorbitantly (IMHO) complex input set-up.

[...]

YMMV, but Factorio is quite playable in the 30-60 UPS range !

And consider yourself lucky that you're not in the situation where you have to resort to modded modules and beacons just to be able to finish the game before end up way lower than 30 UPS :
"I didn't want to do this | Path to Int Units | Pyanodon's Factorio" :
https://youtu.be/JK85mFlybFw?t=434
(I am myself still lower in the pY Alien Life tech tree (just before circuits3), and my base is way smaller, but I have already seen down to sub-30 UPS drops, and am already starting to consider what megabase-style mitigations I might have to take...)

noclaf wrote: Sun Apr 10, 2022 12:46 pm [...]

What's in there for me? Spend 100 hours thinking, how belt weaving might be able to save me 1 out of 20 smelters and 2 out of 40 inserters? Going bigger is the only "fun" target ahead of me. And you are right more efficient entities won't make my factory larger - but at least I will produce more. Currently I'm not able to do any of the two.
And I know there are 30-40kSPM bases at 60UPS but...let's be honest they have nothing in common with regular gameplay and they are designed from the scratch for this one purpose (including stuff like not having the same item on both sides of belt to improve performance, setting up circuits to turn on certain entities only once in X ticks...). That's really cool and kudos to the creators, but that's no longer a fun for me.

[...]

Isn't the UPS-minimizing approach the *definition* of megabases ?? What *else* is there to do (strategy-wise) once you finish the game ?

not having the same item on both sides of belt

How is this *not* regular gameplay ? IMHO it's the opposite that isn't "regular gameplay" : with rare exceptions like iron and copper.

BlueTemplar wrote: Sat Apr 16, 2022 12:03 pm YMMV, but Factorio is quite playable in the 30-60 UPS range !

Temporarily and/or on a smaller base I guess yes. Long term and on a large base...meh...just build solar fields on sub 50 UPS is annoyingly slow (it's annoyingly slow also on 60UPS but then it gets unbearably bad ).

BlueTemplar wrote: Sat Apr 16, 2022 12:03 pm Isn't the UPS-minimizing approach the *definition* of megabases ?? What *else* is there to do (strategy-wise) once you finish the game ?

I have decently optimized blueprints, 12 beacons, as low number of entities as possible etc. Just at certain point (IMHO) I feel you are not optimizing UPS to play but optimizing to optimize. Sure, if you like optimizing it's still fun, otherwise...

BlueTemplar wrote: Sat Apr 16, 2022 12:03 pmHow is this *not* regular gameplay ? IMHO it's the opposite that isn't "regular gameplay" : with rare exceptions like iron and copper.

I do it, but for the sake of i.e. production ratios. Starting to do it (or rather HAVING to do it) in order to save 0.001ms on update because the inserter then does not have to reach for other side of the belt...is it still fun?

And I truly believe large number of players is still finding a lot of fun in Factorio. I still turn on the Krastorio game occasionally. I raised my concerns as a noob casual player.

Hello. My game on Steam is still at version 1.57, although the update is active in the settings. Why is Steam auto? not updating? What should I do?

oyunbagimlisi wrote: Wed May 11, 2022 1:58 am Hello. My game on Steam is still at version 1.57, although the update is active in the settings. Why is Steam auto? not updating? What should I do?

The more recent versions are experimental (beta). By default, you’re on the stable branch, so won’t get these latest ones until they push them to stable.

FuryoftheStars wrote: Wed May 11, 2022 2:00 am

oyunbagimlisi wrote: Wed May 11, 2022 1:58 am Hello. My game on Steam is still at version 1.57, although the update is active in the settings. Why is Steam auto? not updating? What should I do?

The more recent versions are experimental (beta). By default, you’re on the stable branch, so won’t get these latest ones until they push them to stable.

thank you.