KSP2 System Requirements

[K^2]

17 hours ago, Vl3d said:

Then why are the CPU requirements so low? Even with multithreading we should expect higher loads for physics calculations. Is it possible that CPU requirements will increase as EA advanced?

Why? KSP2 does continuous collisions, but other than that, the physics load is about the same. Think about KSP requirements to run a "reasonably sized" rocket. We know that Unity's PhysX did receive a few optimizations in the past few years, so we expect it to run a little better out of the box, and if Intercept switched to Havok it's an even better improvement on top of that. In addition, a lot of the KSP's performance problems with larger ships had nothing to do with physics. If these things were fixed, we expect CPU requirements to be fairly manageable. In its current state, I expect KSP2 to run better than KSP for the same size rocket on the same CPU.

I still have concerns about the CPU, but they have mostly to do with all the things that aren't the core physics simulation. Things like running 1M time warp with your engines firing for an interstellar transfer, while there are dozens of other ships and satellites buzzing about, and several colonies doing their thing. None of this is in yet. The core physics was always supposed to be just a little heavier than the originals, but with some optimizations to compensate for it.

Hopefully the above is something that Intercept can make use of the threading for. In that case, the AMD min spec can remain pretty much the same, and Intel's go up only a little bit for the full release. Maybe there are even optimizations to still have, and the CPU requirements can stay as is. But again, none of that has to do with physics. It has to do with absolutely everything else the game has to do, which is still a lot.

15 hours ago, mcwaffles2003 said:

Were it to be done, what extra potential do you think is there in such a kind of implementation over running on a CPU?

And I wanted the context of the above for the reply here. For a ship with a hundred parts or so, I don't think it matters. There are things you can do with aerodynamics occlusion, for example, that would be way more efficient on the GPU, but there hasn't been an indication of it being considered. You're really only getting benefits of physics on the GPU when you do a lot of it. Which KSP and KSP2 do a lot more than most games, sure, but for comparison, on the last AAA game I've worked on, we had a cap of about 2,000 collision particles for the FX system. Each one could make multiple contact points with the ground or other objects in the scene, and participated in the collision solver in much the same way as a part would in KSP, resulting in thousands of constraints that had to be solved for. And while this was a high-end CPU feature, it was something we were doing in a game that already had a lot going on with AI, networking, animation systems, and scripting system going on on the CPU. So I don't think we had any significant advantage over KSP, other than running a better physics engine.

If you want a 10,000 part ship flying through the atmosphere at a 10x time warp, yeah, the CPU is going to start having problems with it. That's not a common scenario, though. It'd be interesting to see how far the game can be pushed with a simulation like this, but I don't think it's strictly necessary.

But if we do want to speculate, one interesting avenue is deformable terrain. You can do it with CPU, of course. So many games have, but there are limitations. It's a lot easier to do this if you have a unified system handling terrain collisions and the rendering. You can then have something as small as a kerbal's footprint making an impact on a tiny rover, or as large as a new crater forming from a comet impact. All procedural and without significant impact on performance via the magic of virtual textures and your collision system doing tests against that same virtual texture. There are some clever optimizations you have to do to account for scaling, but so long as you are dealing with a height map (read, the crater is significantly smaller than the body it's in) there are clever ways to handle that within the constraints of virtual texturing.

So that would be my callouts for what you could do with GPU physics that would be very difficult with CPU - absurdly large part counts and high fidelity procedural craters.

[dave1904]

6 hours ago, Bej Kerman said:

Talk about the framerate if you like, I personally love the fact that the launch loading times don't leave room to make several cups of tea.

And besides, there's not an apparent physics delta, so at least 20fps doesn't mean waiting 3 times longer to get to orbit

There is a delta i believe. This guy said it took a while to get into space.  https://youtu.be/TFC4ul2uCCg

 

[Vl3d]

4 hours ago, alexalex said:

Myself I will eagerly await EVE volumetrics, finish the mod update I'm working on, and play original KSP while  KSP2 matures.  Eventually KSP2 will surpass KSP1 but it won't be in the first few months after EA release.

Good plan!

[Incarnation of Chaos]

5 hours ago, alexalex said:

 

Have you ever coded anything?  Its pretty clear this game is in the early stages (probably shouldn't be released just yet) with missing core fundamentals such as heat transfer etc, and as such it looks to be very unoptimized.  Personally, I think they are releasing it now because those are the orders from on high they have received rather than thinking its ready themselves which explains the current state of things.  It also doesn't take very much at all for a game, any game, to slow a comp to a crawl.  Just one bad piece of code will do it, and clearly given the letter from Nate he seems to know that there are some big bad pieces of code in there. 

So personally I don't think its the end of the world for KSP2 just yet, rather that its unfortunately following the trend of a lot of modern games and releasing before they should be due to shareholder and corporate pressures.  Myself I will eagerly await EVE volumetrics, finish the mod update I'm working on, and play original KSP while  KSP2 matures.  Eventually KSP2 will surpass KSP1 but it won't be in the first few months after EA release.

 

Have actually coded quite a bit myself, it's one of the reasons I'm not confident they'll improve performance significantly.

The performance we're seeing almost looks like a debugger is hooked up (can't be the case because it's a retail exe, but that's the example that comes to mind)

We were promised no early access and a full game on release, and we got early access and only sandbox mode that's a 20 FPS slideshow that needs more hardware than games that look 10x better to do it.

I do understand that milliseconds and cycles matter, even a loop that's running a few ms slow can add up to massive performance issues depending on how often it's being called. 

The fact that the physics is still causing lag is also a pretty grave concern for me, because either they don't have their physics LOD enabled or weren't able to get it working. That's a low-level, tightly integrated system that likely cannot just be "patched in" due to the dependencies of parts/colliders/meshes etc. So if that's not working, one of the biggest things that KSP2 was going to do to optimize just went straight out the window.

And if they have all this performance still on the table like everyone seems to be thinking, why ship? Game was already delayed three years, I seriously can't fathom 2K coming down and being like "yeah bruh you gotta send it now" after 3 years of delays lol

None of it makes sense, and none of it bodes well for the future of the game. But all I see is people who want to believe it'll all work out without any hard evidence, praise and ignoring the fact that this is another broken promise on top of the pile.

That's the crux of the issue, the hardware requirements are a symptom of a much larger issue than just "muh fps". Could they prove me wrong in time? Maybe, but they're certainly on a losing streak as of late.

 

[Kerbin Launch Coalition]

1 hour ago, Incarnation of Chaos said:

Have actually coded quite a bit myself, it's one of the reasons I'm not confident they'll improve performance significantly.

The performance we're seeing almost looks like a debugger is hooked up (can't be the case because it's a retail exe, but that's the example that comes to mind)

We were promised no early access and a full game on release, and we got early access and only sandbox mode that's a 20 FPS slideshow that needs more hardware than games that look 10x better to do it.

I do understand that milliseconds and cycles matter, even a loop that's running a few ms slow can add up to massive performance issues depending on how often it's being called. 

The fact that the physics is still causing lag is also a pretty grave concern for me, because either they don't have their physics LOD enabled or weren't able to get it working. That's a low-level, tightly integrated system that likely cannot just be "patched in" due to the dependencies of parts/colliders/meshes etc. So if that's not working, one of the biggest things that KSP2 was going to do to optimize just went straight out the window.

And if they have all this performance still on the table like everyone seems to be thinking, why ship? Game was already delayed three years, I seriously can't fathom 2K coming down and being like "yeah bruh you gotta send it now" after 3 years of delays lol

None of it makes sense, and none of it bodes well for the future of the game. But all I see is people who want to believe it'll all work out without any hard evidence, praise and ignoring the fact that this is another broken promise on top of the pile.

That's the crux of the issue, the hardware requirements are a symptom of a much larger issue than just "muh fps". Could they prove me wrong in time? Maybe, but they're certainly on a losing streak as of late.

 

I found the videos we saw from the likes of Matt Lowne interesting to be honest because it's not how I feared it would be. It does appear *very* slow at times but it's a 'smooth' slow at least, there doesn't appear to be jank, stutters, that sorta thing. I think most Kerbal Space Program players can deal with that for the time being. The big question mark is, can the developers improve this significantly? It's also worth noting the game seems to speed up a lot after the transition to 'space', which suggests that the aero model might be the root cause of that, I dunno. Kerbal Space Program (the original) had the same tendency, where any slow-down and stuttering tended to be reserved for in-atmosphere flight. That said, I would have expected after three years of delay that the performance would have been better at this point, but I do *really* like a lot of the GUI and interface changes, those are really nice.

[JoeSchmuckatelli]

The point above about similar part count crafts working the same or better is likely well found. 

I'm interested to see what people get up to in the first weeks of the game.  I think there are likely to be a lot of (if not most) EA players who fall into two main camps:

1. Highly experienced players who want to recreate their Grand Tour crafts or make the most awesome or absurd GT crafts with very high part counts and then go looking for all the 'new' of the Kerbolar System, and 

2.  Experienced players who want to recreate past missions or build complex crafts for visiting bodies they found complicated or difficult to achieve in KSP. 

I suspect that both of these groups will be stressing the system and having fun and finding lots of bugs. 

There is also a third camp - one that is split into two groups:

A. Players like me that enjoy the game but never really got past the moons (maybe one or two other bodies), and

B. New players / returning players who were interested but never quite got the hang of KSP. 

This group is going to be looking hard at the onboarding (tutorials) and ease of access stuff as well as the 'fun' of the game in ways I don't think the preceding two groups will. 

We are likely to be building rovers and planes and trying to land on the moons or other biomes of Kerbin and generally noobing around for the fun of it. 

My hope is that the new /inexperienced voices don't get drowned out. 

If too many people are building absurd part-count crafts and then griping about performance it is going to have a negative impact on the interest of the new / returning players who are unlikely to be building high part count crafts during the first few months. 

 

 

Edited February 21, 2023 by JoeSchmuckatelli

[AtomicTech]

On 2/17/2023 at 1:37 PM, SolarAdmiral said:

Admittedly, it's been a long time since I paid attention to a game's system requirements.

Anyone know what resolutions they're usually talking about? Like would the "recommended" RTX 3080 be at 4k? I'm hoping this isn't recommending a RTX 3080 just for 1080p.

Aye, I'm hopin' that I can get a good experience out of my new laptop RTX 3070.

[Leopard]

15 hours ago, Incarnation of Chaos said:

How many times have I heard that same thing? Performance will improve with time.

You know how many times I've seen that actually occur? Maybe once, twice in my entire gaming experience.

You know what else they said? We'd get a complete game and not this early access hackjob. I've been willing to give intercept immense slack, but this is just downright shameful.

Forgive me for not caring about what they say about their plans at this point, because they have proven time and time again they won't hold themselves to it

what "performance will improve with time" usually means is essentially "hardware will improve at a rate slight ahead of us adding bloat to the code"

 

though they are being very clear and have been for some time that early access is a seriously cut down version of the game, makes sense as for now they can focus on getting the UI right and working before adding more "stuff"

[Bej Kerman]

26 minutes ago, Grenartia said:

Also, a lot of these "just play KSP1 with mods, its better anyways" people bug me because KSP2 having those features means those features are officially supported. How long has Airplanes Plus (as much as I love it), for instance, had landing gear that flips your craft around before you can even launch? At least since game version 1.8. I strongly suspect the only reason the orbital decay bug hasn't been tackled yet is because I'm the only one whose OCD it triggers enough to complain about, and because 90% of the development effort is going towards KSP2. But modders, as much as they are pillars of the community, aren't necessarily required to fix bugs (especially if something comes up IRL). For them, its a hobby they get tipped for, not a job with a living wage. 

EXACTLY!! I can't begin to understand how having these things integrated into the game is somehow on the same level as needing to stack 30 mods to cover basic things like this!

As Matt Lowne stressed, this is the baseline! This isn't modded, this is the pad modders launch from!

Edited February 21, 2023 by Bej Kerman
Dirty Typo

[Vanamonde]

A number of posts about KSP1, and the business of game development, have been removed, because they are not about KSP2 system requirements. 

[ArmchairGravy]

If it hasn't been said yet, allow me: If you don't meet minimum requirements your bug report will not matter.

[MechBFP]

43 minutes ago, ArmchairGravy said:

If it hasn't been said yet, allow me: If you don't meet minimum requirements your bug report will not matter.

That’s not true at all. Many things have nothing to do with minimal specs. Although such feedback may be less prioritized at times. 

[Grenartia]

1 hour ago, MechBFP said:

That’s not true at all. Many things have nothing to do with minimal specs. Although such feedback may be less prioritized at times. 

Honestly, it should probably be prioritized highly, in order to meet the stated goal of making the game more compatible with less powerful systems.

[Arugela]

If we are lucky the high requirements are physics on GPU for realistic aero and other things. Not that I can necessarily play it. It will be interesting if I can get it to work on a phenom1100t cpu with a 10603gb on lowest settings, on linux via proton. 8)

Edited February 22, 2023 by Arugela

[Starcatcher2009 - Modder]

I have a PC, integrated CPU and GPU, and I don't care if my computer becomes a BBQ griller. I will have this game

[ondrafialka]

On 2/21/2023 at 12:15 AM, Incarnation of Chaos said:

Forgive me for not caring about what they say about their plans at this point, because they have proven time and time again they won't hold themselves to it

This is how I feel now. I still believe they want to be honest and many of their statements are honest promises, but they are mixed with things they probably have to say in order to keep Take Two (or whoever gives them money) happy and not closing the project. For example, I do not believe the official reasons for going to EA. I think it was purely in order to generate income and avoid being closed. Now I am not sure if I can beileve what they say about performance improvements. I don't want to blaim the developpers, just stating my feelings.

[magnemoe]

11 hours ago, MechBFP said:

That’s not true at all. Many things have nothing to do with minimal specs. Although such feedback may be less prioritized at times. 

Part of the problem is that marginal systems get lots more bugs, now too slow cpu or to little memory probably give more than to week graphics card but much more weird stuff happens on weak systems. 

[Chibbob]

On 2/21/2023 at 3:36 AM, Rocket Farmer said:

Per the steam GPU survey it’s somewhere around 50%.

Around 68% of steam users don’t have the required specs, so I was 10% off.

[Rocket Farmer]

2 hours ago, Chibbob said:

Around 68% of steam users don’t have the required specs, so I was 10% off.

You forgot about the AMD systems as by a quick count I see 35% of users are on Nvidia systems that meet or exceed that limit.  Add in AMD and it goes up around another 10%+.  
 

Also that’s all steam users of which there are a bunch with potato’s playing Tetris.  The people playing games like Kerbals would, on average, have better systems.  A not insignificant percentage of the systems in there wouldn’t even play KSP 1 so they aren’t KSP players to begin with.

 

So you said 20%.  It’s somewhere around 45% by the survey and in reality probably tops well over 50%.  So no,  not “close” at all.

 

[sjwt]

On 2/19/2023 at 7:06 PM, snkiz said:

That isn't what this chart says. The axis are cut off but the vertical is framerate, and the chart is the Tom's Hardware 2023 roundup. This is the 1080p ULTRA settings average. Context makes all the difference.

There is no point in listing cards that don't work ..

The 1060 3GB could not run all the tests, the 6GB did, both are below minimum line. And the whole 9XX family is below its 100% not needed 

To me, those lines are drawn strofht across at the lowest listed card for the minimum and high specifications, the other listed cards below are plotted at there realtive scores based on having enough features that they should be able to run, but only a few days to go.. 

[snkiz]

14 minutes ago, sjwt said:

There is no point in listing cards that don't work ..

The 1060 3GB could not run all the tests, the 6GB did, both are below minimum line. And the whole 9XX family is below its 100% not needed 

who said anything about that? You misunderstood my post. 

 

14 minutes ago, sjwt said:

To me, those lines are drawn strofht across at the lowest listed card for the minimum and high specifications, the other listed cards below are plotted at there realtive scores based on having enough features that they should be able to run, but only a few days to go

It doesn't matter what you think those lines are. It matters what they actually represent. The chart has nothing to do with KSP. It's a chart showing the trends in system requirements for recent games. With full context, what it shows is that KSP2's current system specs are just out of line with the rest of the industry. And way out wack with the state of the GPU market. To me that's a settled issue, my concerns have been addressed in follow up posts. But that chart, it doesn't say what you think it does. I believe the person who anointed it only intended to show the range of gpus ksp2 would run on. The broken telephone that is the internet is has given it a life of own though.

Edited February 23, 2023 by snkiz

[MiniMatt]

Not sure I'm adding anything to repeat what has likely already been said a hundred times, but adding my disappointment that my computer which makes Cyberpunk & Forza 5 look amazing, can track the movements of millions of moving parts in Factorio mega factories, fly around the world in MS Flight Sim, and has provided thousands of hours of entertainment in KSP1 won't even hit the minimum specs for KSP2.

Not angry about it, it just is what it is. It's turned a day one purchase (even at that ~€50 price point) into either a months-from-now purchase when minimum reqs are tuned down, or a years-from-now-on-deep-sale purchase when I next upgrade.

Wish the devs and the game the very best of luck, bit sad I'll be wishing from the sidelines this time.

[ICantDockShips]

Probably still gonna buy the game but I’ll have to wait and get myself a new pc to play. Luckily I was planning on it anyways.

[histalonia]

Just an FYI.  The KSP2 standalone .zip download is just under 16GB.  Plan on about an hour to download.  My download is halfway done with another 25 minutes left.  (I'm getting 5.5 MB/sec and my internet download speed typically tests in at about 25 MB/sec.)

[tilliepops]

everyone the system requirement have been changed look on steam.