Kerbal Space Program 2 - Pre-Release Notes

[regex]

3 hours ago, SuperMiiBrother said:

I just realized…

If thermal systems are disabled, can I land on the sun?

AFAIK that was disabled for the ESA event, it may actually be enabled tomorrow.

OTOH, you should just land on the sun for because.

Edited February 24, 2023 by regex

[stephensmat]

Alright, first one to the Mun Arch brings snacks.

Actually, now that I think about it, if Multiplayer is possible, does that mean the Arch becomes something like a 'Hub'?  Assuming we pick which 'other' games we can link up to, it means a lot of us could conceivably meet somewhere in a 'common' game...

[RedderThanMisty]

12 hours ago, tomk said:

Dear @Intercept Games, beside features, will it also support multiple operating systems ?

I'm asking, because it feels to me that pretty graphics are creating a lock-in into DX ecosystem. ( Anyone typing now, suggesting VM or VINE, please don't ).

I play on Linux too, I'll post about how well proton works with ksp 2 once I have it downloaded, as I haven't booted into my windows drive in over two years and aren't going to do so anytime soon.

Who knows, perhaps proton single-handedly fixes the performance bugs just like it did on Elden Ring's launch. Wishful thinking, sure, but not without precedence.

1 hour ago, stephensmat said:

Alright, first one to the Mun Arch brings snacks.

Actually, now that I think about it, if Multiplayer is possible, does that mean the Arch becomes something like a 'Hub'?  Assuming we pick which 'other' games we can link up to, it means a lot of us could conceivably meet somewhere in a 'common' game...

I doubt it, it would likely lock out alot of new players who haven't mastered landing at a target yet, let alone landing on the mun. I don't believe that should be a requirement if they just want to goof around. However, using the mun arch as a main menu background for a multiplayer session browser could be a good idea.

[VITAS]

Hi,

ive added Kerbal Space Porgram 2 under https://spacedock.info/kerbal-space-program-2

to SpaceDock.

Because we dont know the Version Numbers yet i used a temporary Version number of 0.0.0.0.1. Once "real" Version numbers are added to SpaceDock you just have to click update on your mods like you would with any new game release.

Who will be the first to upload a KSP2 mod?

Have fun.

[tomk]

8 hours ago, R-T-B said:

WINE

quite, my bad.

8 hours ago, R-T-B said:

And if you haven't tried WINE recently, I'd say you'd be in for a shock, but due to lack of optimization it will probably still run badly.

Every time I approached wine in last 10 years it was requiring me to do library kungfu, and most of my rigs are dev rigs ... if I had a dedicated rig for games it would run as well windoze, but that would be sacrilege ... and frankly through the pandemic cs and ksp kept me entertained because it did not require any faffing about.

[Mekan1k]

On 2/20/2023 at 8:57 AM, AlphaMensae said:

This emphasises the "Early" in Early Access. This is not the final game, far from it. It's like a mix of pre-v0.13.3 KSP and the v1.1 Experimentals builds that Squad released, and those were a buggy mess.  

So when all the rockets exploded and we had to deal with the fact that exploding was the LEAST of our worries when the Kraken was waiting to eat our feets. Got it. 

Everyone- watch for tentacles.

[Waifu Art Thou Romeo]

14 hours ago, SuperMiiBrother said:

I just realized…

If thermal systems are disabled, can I land on the sun?

Ferb, I know what we're gonna do today.

[tstein]

On 2/21/2023 at 11:50 PM, K^2 said:

It's hard to say with certainty that it's a physics lag. You can see that, like most people who tried building bigger rockets, Everyday Astronaut had to add a lot of struts to the boosters to keep them stable, so that could be it, or we might see a return of a fuel pathing algorithm being really bad - like if the boosters are having hard time "convincing" themselves that they can't find a path to the main tank, or some other issue.

But yeah, regardless of whether or not it's specifically a problem with physics simulation, this is definitely a CPU lag, and likely because of something that's taking entirely too long to process on the main thread. Since we know from the min spec that the game is main thread bound, much like the original KSP, I don't think we'd see such a stark contrast in performance if the bottleneck was on the worker threads.

A lot of it is speculation, though. I should be able to take a closer look at which cores are doing what when I can run the game on my PC. I promise to share anything I can dig up without breaking the EULA.

What baffles me is how something as pathing can be a problem  on a system with  so few nodes? For god's sake, when I was at university in the late  80's  I remember doing  flow simulation  (not  a scale different from what a game like this need for the  tanks) on sewage system  of  real cities with HUNDREDS of nodes   with better performance and in a computer with 8 Mhz!

[zer0Kerbal]

All I care about is : how to mod it!

[K^2]

3 hours ago, tstein said:

What baffles me is how something as pathing can be a problem  on a system with  so few nodes?

A graph with relatively few nodes can have a huge number of possible paths if you have a good number of edges. Note that a lot of these ships have a huge number of struts. If struts are considered for fuel flow, for example, the number of possible paths from engine to every tank can be enormous. A good algorithm doesn't have to touch every single path. We're not solving a traveling salesman here. But it's very easy to make a mistake in your algorithm that effectively results in every single possible combination being tried,  and it can be hard to detect that problem in testing. If your test craft had 10 parts and no struts, it will still be instant, and the problem might only appear when you start building giant ships - which, yeah, will show up in testing, but there could have been other performance issues masking it until recently.

If this is a problem, it's kind of understandable and should be easy enough to resolve now. It would just be ironic, seeing how that was a known problem with KSP, and somebody really ought to have checked to make absolutely sure they're not stepping on the same rakes.

And, of course, it could be something entirely different. It is possible that PhysX itself has some sort of traversal issues with joints that this is manifesting. That might be trickier to identify and solve, but there are certainly ways of solving this one way or another. There is no legitimate reason for that kind of a poor performance. Just a bunch of plausible errors in code that can and should be fixed.

[sunsetteOS]

14 hours ago, stephensmat said:

Alright, first one to the Mun Arch brings snacks.

Actually, now that I think about it, if Multiplayer is possible, does that mean the Arch becomes something like a 'Hub'?  Assuming we pick which 'other' games we can link up to, it means a lot of us could conceivably meet somewhere in a 'common' game...

If you count one of the Mun arches glitching into space and crashing into my ship after a quicksave/quickload in Munar SOI then I think I'm the winner of the snacks.  But I'm safely guessing you mean in a less Krakenesque way LMAO

[tomk]

15 hours ago, RedderThanMisty said:

I play on Linux too, I'll post about how well proton works with ksp 2 once I have it downloaded, as I haven't booted into my windows drive in over two years and aren't going to do so anytime soon.

Who knows, perhaps proton single-handedly fixes the performance bugs just like it did on Elden Ring's launch. Wishful thinking, sure, but not without precedence.

I would say don't bother - I've bitten the bullet and I can report that result is "no cigar"

"

Failed to initialize graphics.

Make sure you have Direct 11 installed, have up to date drivers for your graphics card and have not disabled 3D acceleration in display settinas.

InitializeEngineGraphics failed

"

[Madrocketman]

On 2/17/2023 at 11:32 AM, Intercept Games said:

Clicking and dragging the middle mouse button allows you to track the VAB camera along the vehicle’s long axis. To focus the camera on a particular part, middle-click that part. Use the scroll wheel to zoom

This isn't working for me. I use a laptop, but I can't go up or down. Anyone else figure out how this is supposed to work for a laptop?

Edited February 25, 2023 by Madrocketman
can't

[Linkageless]

9 hours ago, tomk said:

I would say don't bother - I've bitten the bullet and I can report that result is "no cigar"

Thanks for the report.   I would dearly love to play KSP2 on my os of preference, Linux, and am still hopeful it will be a standard option in due course.

I did, however, come to the conclusion long ago that I needed to build my first personal windoze machine in about 25 years in order to make life easy for myself when KSP2 came out.  I've built an awesome PC with a cheap but legit windows licence and I'm not regretting it.

[Negative Periapsis]

7 hours ago, Madrocketman said:

This isn't working for me. I use a laptop, but I can go up or down. Anyone else figure out how this is supposed to work for a laptop?

Nope, doesn't work for me either. For a game that has supposedly had an expert user experience team redesign said user experience, they seemingly haven't done a lot of user experience testing.

This just adds to the disappointment I feel after the first few hours of playing. The excitement has evaporated as the game just isn't pleasant to use. The map view with the fiddly little click areas to get AP/PE information which disappears when you want to adjust your maneuver node is just another example. Of course the map view is affected by the MMB problem as well, as there is no other way to move the camera freely there. Add the mushy look and the pixely text.

Wake me up when the next update comes around. I'm off getting my satellites set up in orbit around Eve in KSP 1 until then.

Edited February 25, 2023 by Negative Periapsis
After restart I got better FPS in the VAB.

[Mutex]

On 2/24/2023 at 8:23 AM, RedderThanMisty said:

I play on Linux too, I'll post about how well proton works with ksp 2 once I have it downloaded, as I haven't booted into my windows drive in over two years and aren't going to do so anytime soon.

For me I just had to change the Proton version to Experimental, and it works fine (I mean, no more issues than people on Windows anyway).

This is on Arch + Plasma (Wayland), using Steam Flatpak, and AMD 6800XT.

[tstein]

18 hours ago, K^2 said:

A graph with relatively few nodes can have a huge number of possible paths if you have a good number of edges. Note that a lot of these ships have a huge number of struts. If struts are considered for fuel flow, for example, the number of possible paths from engine to every tank can be enormous. A good algorithm doesn't have to touch every single path. We're not solving a traveling salesman here. But it's very easy to make a mistake in your algorithm that effectively results in every single possible combination being tried,  and it can be hard to detect that problem in testing. If your test craft had 10 parts and no struts, it will still be instant, and the problem might only appear when you start building giant ships - which, yeah, will show up in testing, but there could have been other performance issues masking it until recently.

 

 

True but that complexity can be solved at design time and  pre stored. All the non tank  nodes can  be mapped on if they are reachable  from each tank  without passing for a decoupler and   everything that can be reached  by tank A can be  treated as a single same  enter point.  It is  simple transitivity.  If  A can reach  C trough B then C and B are the same node for this  purpose and all vertex leading to C or B can be merged. It is simply the finite automata simplification algorithm. That reduces  the number of vertex    drastically. That is how we did  in the 80's We did not care for  EVERY single sewage entrance  but for every connection that was reachable by one.  Surely it is not easy to test as a human,  but you do  make a generated set of scenarios and compare your simulation with an FMM (plenty of implementations ready to use) if your results  are different.. you have a bug.

 

I really hope it is not a bug in the fuel transfer because  that is a very bad prospect for QA of the game. If it is a PhysX issue at least it is more understandable.

[tomk]

7 hours ago, Mutex said:

For me I just had to change the Proton version to Experimental, and it works fine (I mean, no more issues than people on Windows anyway).

This is on Arch + Plasma (Wayland), using Steam Flatpak, and AMD 6800XT.

Huh ? first time I hear of "steam flatpak" ... I'll give it a shot, maybe in some future ... (also I did change proton to experimental and had no luck on my steam) 

Edited February 25, 2023 by tomk

[Mutex]

4 minutes ago, tomk said:

Huh ? first time I hear of "steam flatpak" ... I'll give it a shot, maybe in some future ... (also I did change proton to experimental and had no luck on my steam) 

Flatpak is a way of packaging and distributing Linux software so that it can run on (basically) any distro. It's the easiest way to install Steam on Arch.

[Waifu Art Thou Romeo]

On 2/24/2023 at 11:35 AM, Waifu Art Thou Romeo said:

Ferb, I know what we're gonna do today.

UPDATE: ✨no✨

 

The sun doesn't have a surface, but you can do some wicked slingshot maneuvers since you can go all the way to the core. Shame we can't do interstellar yet. An 80km orbit alone is like 100k m/s

 

UPDATE 2: Tim C. Kerman is currently traveling away from the kerbolar system at 2,000 km/s

Edited February 26, 2023 by Waifu Art Thou Romeo

[K^2]

13 hours ago, tstein said:

True but that complexity can be solved at design time and  pre stored.

You could, but even without caching, so long as you do this right, it's not an expensive algorithm. You can run it every frame and not make an impact. It's only when you screw it up that it gets expensive. I was merely pointing out that there are ways it can go very wrong, still give you correct results, but get very computationally expensive with larger rockets due to bad algorithm choice. It's one possible explanation for performance issues among many.

One thing that hints to this possibility being slightly more likely, by the way, is that radial separators seem to have a cross-feed bug and continue providing fuel to the entire rocket regardless of stages. This can all be related. Or it can be entirely coincidental.

[J.Random]

On 2/17/2023 at 9:32 PM, Intercept Games said:

Clicking and dragging the middle mouse button allows you to track the VAB camera along the vehicle’s long axis.

This is not necessarily stupid, but why couldn't you just allow players to use scroll wheel? There is enough modifier keys on a keyboard for all three axis.

[tomk]

3 hours ago, J.Random said:

This is not necessarily stupid, but why couldn't you just allow players to use scroll wheel? There is enough modifier keys on a keyboard for all three axis.

Not everybody has a scroll wheel. 

[J.Random]

1 hour ago, tomk said:

Not everybody has a scroll wheel. 

Ditto MMB, so what?

Edit: also, how do these people which don't have mouse wheel zoom in and out? Same keys, with the same modifier keys, could work for camera movement.

Edited February 26, 2023 by J.Random

[000PainKiller000]

46 minutes ago, tomk said:

Not everybody has a scroll wheel. 

We learn something new and exciting every day. I'm getting used to the new controls.