Please help > Do you want game speed optimizations(FPS)? (Poll)

[Rainbowdesign]

Actually i am pretty sure the next update will contain some fixes. btw i checked it my savegame runs at 1-2 fps and deleting stuff gets it up to 6 fps. deleting the most of the colony gets it up to 10 fps with mods. I think i found the problem: Modded Auto sweepers with high range kill performance. Getting 12 fps after uninstall.

 

But then i remember the same problem from before the turbular update. I think i never managed to play a game above 300 cycles before its unplayable.

[Nitroturtle]

Babba requested my input here, so I'll give my thoughts.  First off, I'm a professional software developer, so I'm pretty familiar with the process.  The thing is, while performance is something that should be considered from the beginning, the real optimizations aren't going to happen until the content is mostly complete.  It really doesn't make much sense for them to be optimizing code that will potentially change in the near future.  Based on their stated roadmap, I do expect to see significant optimization done over the next year, though I'd guess most of that will probably be in the last quarter once they've fleshed out content a bit and fixed the bugs.

As for my personal experience with the game, I started playing with an i7 2600k Sandy Bridge processor.  The game ran very well early game, but obviously as my base grew and the cycles passed, the lag got worse and worse.  When they first added the space biome, I saw a significant performance hit once the surface was breached.  I believe this was due to the MASSIVE amount of regolith in the game initially, which has been mostly fixed in the recent patch.

About 2 months ago, I upgraded to an i7 8700k Coffee Lake CPU.  I'm currently running it at 4.6GHz and performance has been totally acceptible in my ~1500 cycle base, usually hovering around 30fps.  With that said, I'm always cognizant of performance when designing my base, and try my best to avoid things that I believe will have a large performance hit.  All my dupe pathing is fairly limited and deliberate, my jetsuits are confined to a limited area and I try to avoid excessive automation looping.  I do tend to use a lot of automation, but I make sure to design it such that it's only running when needed.  For example, door pumps that only operate when pressure reaches a certain amount, or auto-drop sweepers on clock timers to only run part of the day.

One thing I've noticed is that there seems to be a lot of people that have completely unrealistic performance expectations.  If you're running hardware that's more than a year or two old, the FPS problems you're having are on you.  You can't play this game on a 5 year old laptop with integrated graphics and 4GB of RAM and expect 60fps (or even 30, most likely).  No sane game development company is going to design games around outdated hardware, especially for a game in early access.  Sure, it's nice when games run well on older hardware, and I'm sure the game will improve in that regard once optimizations are done.  But it's still early in the game's life, and it doesn't make sense to limit development just so it will run on outdated hardware at this point in the process.

Another thing I'll mention is that I don't think people take into consideration that mods will have an impact on performance.  Sure, depending on the mod, that impact might be minimal.  However, it could potentially have a huge impact, again depending on the mod.  If you are having performance issues and running mods, the first step is to remove the mods and see if that helps.  

I think the best thing we all can do to help is just to keep playing and giving feedback.  I imagine the testing people have done on the performance impacts of various things is very valuable to the developers.  I know the developers don't have very strong communication with their playerbase (something I'd love to see improve) but they are listening, and I know they are taking our feedback into account.  So just keep playing, giving feedback, testing and having fun.  I've only been playing the game for about 8 months and I'm blown away by the amount of content that I've seen added over that time.  I can't wait to see how the game looks after this next year and I only hope that I don't get burned out on it before I get to see the final product.

[Mariilyn]

Performance optimizations would be welcome, but I'm mostly looking forward for the rocket bug fixes, balance changes and small quality of life improvements. In my point of view the FPS in a bigger base can be maintained to a tolerable amount (13-15 would be the minimum for me) by good habits like systematically limiting the amount of automation loops and pathing loops, sweeping the debris laying around, restricting critter's movement, etc.

It's just unfortunate because bigger bases cannot be brought into preview branches without manually disabling the debug tools and that newer content like jet suits won't be used extensively for now. The last two previews were the first ones I didn't hopped in right away because the new content was late game and I didn't feel like restarting a colony just to test the rocket stuff in survival.

[TheGCat]

1 hour ago, Nitroturtle said:

One thing I've noticed is that there seems to be a lot of people that have completely unrealistic performance expectations.  If you're running hardware that's more than a year or two old, the FPS problems you're having are on you.  You can't play this game on a 5 year old laptop with integrated graphics and 4GB of RAM and expect 60fps (or even 30, most likely).  No sane game development company is going to design games around outdated hardware, especially for a game in early access.  Sure, it's nice when games run well on older hardware, and I'm sure the game will improve in that regard once optimizations are done.  But it's still early in the game's life, and it doesn't make sense to limit development just so it will run on outdated hardware at this point in the process

In whole i sign your written. U forget something in your text. Some people bought the game, because the steam requirements say only "a dual core with 2GHZ at least". So, what's wrong when the people think, they can play the game with the notebook. The problem isn't the user, but rather the wrong requirements. And the requirements doesn't tell u, what generation of CPU they mean. A CPU ago 10 Years ? An actual ? They are not telling the truth in my opinion with this requirements. That's maybe an marketing thing, because nobody would buy the game if the requirements would say: "at least 4GHZ to play the game a little bit smooth :D"

I am also an PC Enthusiast and buy mostly the newest hardware. But like other said, it wasn't required in the last years, because EVERY other game runs perfect. The performance increase aren't so big as for 10 years ago. I checked by myself. If i would buy the i7 8700k Coffee Lake  like u did, i would have JUST 30-40% (For an 7 Years old CPU. Absolutely BAD!!!!!) better performance in whole (Mostly the multi core performance) The single core performance is nearly the same. So, for getting 10 FPS more just for ONI, i should buy a 500 $ CPU + a new Mainboard + RAM ! NOPE!

Ask yourself! Would u buy an complete new PC, when all other games run on ultra settings just for ONI ? It's a difficult question, i know

I sign your thoughts, but think about my written text. What would u do ?

Don't take it serious, but i had made a joke some time ago:

"Klei has a contract with the CPU manufacturers to sell the new expensive CPU's"

 

[babba]

14 hours ago, DaveSatx said:

I trust klei to work on whats important.  several things at once in fact.  its going to a matter of what gets fixed first not if. I'm sure they have people working on many aspects.. i have no idea how big the coding team is but given the size of the updates its not tiny.
reducing the save file size would have a bigger impact for me. it takes forever to save the later we go

 

Mhhh, saving should only take a few seconds.Whats your system specs ? Maybe some others can look in to it if you want to provide a savegame. Happy ONI

1 hour ago, Nitroturtle said:

Babba requested my input here, so I'll give my thoughts.  First off, I'm a professional software developer, so I'm pretty familiar with the process.  The thing is, while performance is something that should be considered from the beginning, the real optimizations aren't going to happen until the content is mostly complete.  It really doesn't make much sense for them to be optimizing code that will potentially change in the near future.  Based on their stated roadmap, I do expect to see significant optimization done over the next year, though I'd guess most of that will probably be in the last quarter once they've fleshed out content a bit and fixed the bugs.

As for my personal experience with the game, I started playing with an i7 2600k Sandy Bridge processor.  The game ran very well early game, but obviously as my base grew and the cycles passed, the lag got worse and worse.  When they first added the space biome, I saw a significant performance hit once the surface was breached.  I believe this was due to the MASSIVE amount of regolith in the game initially, which has been mostly fixed in the recent patch.

About 2 months ago, I upgraded to an i7 8700k Coffee Lake CPU.  I'm currently running it at 4.6GHz and performance has been totally acceptible in my ~1500 cycle base, usually hovering around 30fps.  With that said, I'm always cognizant of performance when designing my base, and try my best to avoid things that I believe will have a large performance hit.  All my dupe pathing is fairly limited and deliberate, my jetsuits are confined to a limited area and I try to avoid excessive automation looping.  I do tend to use a lot of automation, but I make sure to design it such that it's only running when needed.  For example, door pumps that only operate when pressure reaches a certain amount, or auto-drop sweepers on clock timers to only run part of the day.

One thing I've noticed is that there seems to be a lot of people that have completely unrealistic performance expectations.  If you're running hardware that's more than a year or two old, the FPS problems you're having are on you.  You can't play this game on a 5 year old laptop with integrated graphics and 4GB of RAM and expect 60fps (or even 30, most likely).  No sane game development company is going to design games around outdated hardware, especially for a game in early access.  Sure, it's nice when games run well on older hardware, and I'm sure the game will improve in that regard once optimizations are done.  But it's still early in the game's life, and it doesn't make sense to limit development just so it will run on outdated hardware at this point in the process.

Another thing I'll mention is that I don't think people take into consideration that mods will have an impact on performance.  Sure, depending on the mod, that impact might be minimal.  However, it could potentially have a huge impact, again depending on the mod.  If you are having performance issues and running mods, the first step is to remove the mods and see if that helps.  

I think the best thing we all can do to help is just to keep playing and giving feedback.  I imagine the testing people have done on the performance impacts of various things is very valuable to the developers.  I know the developers don't have very strong communication with their playerbase (something I'd love to see improve) but they are listening, and I know they are taking our feedback into account.  So just keep playing, giving feedback, testing and having fun.  I've only been playing the game for about 8 months and I'm blown away by the amount of content that I've seen added over that time.  I can't wait to see how the game looks after this next year and I only hope that I don't get burned out on it before I get to see the final product.

Thank you for your extensive and sophisticated input. I was Coproducer on Crysis, the reason why people suffered from running the game on most PC`s was that the company owners had little interest in spending more money on engine optimization after presenting them in detail the tech specs of the target audience and what our current engine situation is... "I do not care" was the phrase which always stayed ringing in my head. Anyways, the saying "Can it run Crysis?" got famous because of that. As long EA could brag with the best graphics, the milestone monies for our team of +100 peeps came in.

ONI: The game will run good if the devs reduce the map size again to 20% (1/5). Do we want that ? No. Those who do not suffer any problems have a rather little base, light automation or have already so much experience with the game that they know about all the things which add to the lag and work around them. ONI is a great game, players being able to grind a default map size to 0.1 FPS or a few frame per second with $500 chips IMHO its a problem. The game does not have the bangbang success which Cities Skylines has, with CS its easy for buyers to understand that they need a strong cpu&gpu if they want to play a map with a few million city peeps. "Oh big city, oohhh looks awesome look how big...oh yes I want to build that too.Here is my money, cpu/gpu bling catshing catshing.

So hopefully we threw so much dust and tank tracks into the air that we will get some greater FPS optimization efforts. I need a shot now

P.S. I do not use mods on this game, it would be hard to tell what caused a crash and it would just add to the lag fest. I change 95% to 99% of the map into non-simulated Neutronium in the first seconds. For sophisticated pro gamers, the ONI specs would be 6700k 7700k 8700k or higher.

Its hard for casual gamers to see what this game is capable of and what cpu grind ONI can become after a few hundred cycles.

[Nitroturtle]

13 minutes ago, DustFireSky said:

In whole i sign your written. U forget something in your text. Some people bought the game, because the steam requirements say only "a dual core with 2GHZ at least". So, what's wrong when the people think, they can play the game with the notebook. The problem isn't the user, but rather the wrong requirements. And the requirements doesn't tell u, what generation of CPU they mean. A CPU ago 10 Years ? An actual ? They are not telling the truth in my opinion with this requirements. That's maybe an marketing thing, because nobody would buy the game if the requirements would say: "at least 4GHZ to play the game a little bit smooth :D"

I am also an PC Enthusiast and buy mostly the newest hardware. But like other said, it wasn't required in the last years, because EVERY other game runs perfect. The performance increase aren't so big as for 10 years ago. I checked by myself. If i would buy the i7 8700k Coffee Lake  like u did, i would have JUST 30-40% (For an 7 Years old CPU. Absolutely BAD!!!!!) better performance in whole (Mostly the multi core performance) The single core performance is nearly the same. So, for getting 10 FPS more just for ONI, i should buy a 500 $ CPU + a new Mainboard + RAM ! NOPE!

Ask yourself! Would u buy an complete new PC, when all other games run on ultra settings just for ONI ? It's a difficult question, i know

I sign your thoughts, but think about my written text. What would u do ?

 

It's important to understand, that "Minimum Requirements" don't mean "Minimum Requirements with acceptable framerate".  I'm sure the game will run on the hardware they specify, however I doubt it would be a pleasant user experience.  It's been like that with games for ages.  If you are at the low end of the requirements, you can either expect the game to play like a slideshow, or plan to upgrade.  As you said, it's largely a marketing thing, and they aren't going to put specs higher than they need to or they risk losing customers.  Some people are less bothered by low framerate, or just have never had better.  I have two young nephews who play games like ONI as a slideshow on way outdated hardware and they don't know any different.  I however can't even stand to watch them play like that.

I agree that hardware performance increase over the past several years has been minimal (SSDs have probably had the biggest impact in the last 5 years).  I used to upgrade roughly every 2 years, where this last upgrade I waited 6.  I'm happy with spending ~$400 to upgrade after 6 years, versus paying $300-400 every 2 years.

As to your question about buying a new PC to play ONI.. well, the only game I play currently is ONI, and I just upgraded, so that kinda answers the question.  Though I'm fortunate to be in a position with enough disposible income to build a new PC.  But ONI doesn't require a top of the line PC to run reasonably well, especially early game.  It also doesn't require a $400-600 video card like a lot of modern titles.  As I said, my previous 6 year old CPU was completely fine for ONI, though performance did start to suffer late game, it was still very playable.

I think one thing that contributes to the complaints is that performance is directly related to game progress and base design.  I'm sure many people start playing with inadequate hardware that runs perfectly early game, then once they start reaching late game the performance goes downhill quickly.  It's unfortunate, but it's the nature of this type of simulation game.  Optimizing the game should help in the end, but as I mentioned, I don't see that happening until shortly before ending early access status.

[babba]

Great ongoing discussion, I would like to point out that I have a little pc builder thread for little monies - For those which want to upgrade and are undecided what is good value for little gold taler coins.

Its possible to run i7 6700k 7700k 8700k cpus within 30 minutes from 60 fps down to 30 fps, just by extensively excavating the map and making those tiles part of the simulation.

Within 1 hour I took those cpu`s down to 10fps. Add heavy automation = 0.1 to 1 FPS. Not even talking of using jetpackers or "stupid build methods". These are the root problems in my opinion.

[Nitroturtle]

Actually, in fairness, I didn't really upgrade only to play ONI.  I'm looking to pick up a VR setup in the next 6 months, and I wanted hardware that would support it.  The fact that it helped a little with ONI was just a nice benefit, but my previous setup would have probably been fine if I hadn't wanted to upgrade anyways.  As I said, I usually upgrade about every 2 years, so I figured with 6 years I had gotten my money's worth out of the previous setup.  Besides, then I can pass down the old hardware to my nephews so they can actually play games without them being a slideslow. 

[TheGCat]

7 minutes ago, Nitroturtle said:

.... my nephews so they can actually play games without them being a slideslow. 

As a kid, the only important thing was to play the game. Slideshow or not, kids haven't high expectations It was the fascination for the thing. :))) A Pc game is like magic for a kid. Sometimes i miss that... But now i am grumpy cat. To be grumpy is the only important thing for me

[Rainbowdesign]

I really disagree there Modern Pcs even the laptop ones are real powerhorses. It was already 10 years ago possible to program good games and also demanding ones. Todays processors have 4-10 times the processing power. I think most should run well unless there are bigger wastes of processing power somewhere. Surely oni needs to process quite many things in realtime, but factorio is more excessive but from performance i think its running better.

Also i feel it is a good idea to try to constantly work on performance with every update, especially fix bigger things that really drain. Sure is it so that parts of the code change later but there are too many parts in the game that are finished and could be optimized.

 

Btw i find it horrible to make apps with only the newest and strongest rigs in mind it is like making games only for the richest and most enthusiast players while ignoring the casual players who have a computer in first place to write mails but want to sometimes play a round.

4 minutes ago, Nitroturtle said:

 Besides, then I can pass down the old hardware to my nephews so they can actually play games without them being a slideslow. 

Guess they will understand why i feels its horrible.

[Tekky]

8 hours ago, Rainbowdesign said:

Surely oni needs to process quite many things in realtime, but factorio is more excessive but from performance i think its running better.

I fully agree. Despite Factorio having to simulate a lot more, it runs much better than ONI, because it is well optimized.

It is not my intention to criticize ONI for being badly optimized. Since ONI is early access, there will still be a lot of opportunity for further optimizations. Also, Factorio raises the bar pretty high, so it would not be fair to criticize games for not meeting Factorio's high standards.

However, I'm afraid that ONI may never be able to reach Factorio's level of optimization, due to the limitations of C# and the Unity Engine. Factorio was programmed in C++ for a good reason.

Luckily, Unity does give you the option of using C++ code in an external DLL.

[DaveSatx]

  i have to amend what I posted earlier.. i'm nto seeing any fps issues but i am seeing a lot of the dupes doing the 'let me stand here for 2-3 seconds between actions' lag.. lol i was being literal when i answered before.

it's alot like playing stellaris after reaching end game and days are passing like months used to. Patience is a virtue apparently..

[strawberrygirl]

I came here to add my vote for the the CPU issue. I haven't been playing the game much (if at all) since the performance issues started to seriously affect my game play. I love the game, just can't play it right now.

[byterussian]

Quote

Its possible to run i7 6700k 7700k 8700k cpus within 30 minutes from 60 fps down to 30 fps, just by extensively excavating the map and making those tiles part of the simulation.

I stopped to play because of this. It's annoying to spend money for top class pc components(i7 8700k in my case)  to play games at 30 fps(or less).  

[Calvin Valerian]

Yes please please please... I had once tried to load someone else's save game with 1000+ cycle and my PC can't handle the load and FPS performance very poor like nearly unplayable.

My rig is low end Intel Core i3 2100 with only 2 cores.

With AMD HD 5700 Series 1GB VRAM.

[PhailRaptor]

The game clearly needs a full on optimization patch to clear out a lot of issues that are plaguing everything from AI pathfinding to automation lag to phase change shennanigans.

But....

Wouldn't the necessity of such a patch depend on a lot of factors?  And can't some of that be done in pieces to minimize the impact such a wide-ranging change would likely require?  I'm finding the poll options limiting in this regard.  If there was a middle ground about knocking down the low hanging fruit while planning for more later, I'd pick that one.