Unity 5 [Is now available]

[Beduino]

Imagine in this little demos you are using a plane, in game, you have planets, and orbits. Is a different calcule. Not only a one axis gravity...

Actually the demo is fully 3d simulation much like in KSP in term of physics.. There's a gravity vector in 3D i can change it to any direction.

[Red Iron Crown]

I would be curious to see if it multi-threads with only a single "ship" instead of five. I don't think it will but would love to be wrong.

[Climberfx]

---

Yes, any direction, one direction, and orbit is not an one direction. Is an center were around point at.

But like i said, i do not know for sure how KSP is programed, but i know Damion a long time ago, on the Twitch told that the center of the KSP universe, is your vessel, and all things move around it.

So, that could be the start of complexity. LOL.

They could start using multithread putting other cores or physic cores to do physics on the around vessels...

That would turn thins prettier, having parts physics working not only on the subject!

I have Unity 4.X and 5 installed...

After i finish my parts development, i will do some tests...

Edited April 6, 2015 by Climberfx

[Beduino]

Not to be that guy, but I've seen several attempts to show U5 handling the physics much better than we know KSP currently does, which is great. Could you please run the exact same simulation in U4 and post the results similarly, so we can compare the two side-by-side?

Sorry just purged unity 4, not going to download it again

[pincushionman]

Sorry, didn't mean to imply that you should for me. What I meant to point out is this simulation asks the question "does U5 run well?" And it shows that yes, it does. But that's really not the question we're discussing. The question is more properly "does U5 run better than U4?" And without an equivalent test in U4, we can't evaluate that claim.

Just, you know, as a note for the aspiring scientists out there: when designing your experiments, make sure you properly identify your hypothesis, and always include a control!

Edited April 6, 2015 by pincushionman

[Beduino]

Sorry, didn't mean to imply that you should for me. What I meant to point out is this simulation asks the question "does U5 run well?" And it shows that yes, it does. But that's really not the question we're discussing. The question is more properly "does U5 run better than U4?" And without an equivalent test in U4, we can't evaluate that claim.

Just, you know, as a note for the aspiring scientists out there: when designing your experiments, make sure you properly identify your hypothesis, and always include a control!

as per http://docs.unity3d.com/Manual/UpgradeGuide5-Physics.html

"Unity 5.0 physics could be expected to work up to 2x faster than in previous versions"

Im thinking more like proactive here.. let's use Unity 5 and optimize the hell out of KSP, because it shows much promise.

[Nemrav]

The only problem with u5 that I've encountered is the wheel colliders..... they're horrible (if you try to adjust default settings, they bounce you into oblivion) and plenty of people have the same sort of problems....

[boxman]

Those two clauses are in direct opposition to each other. Changing engines is the precise opposite of "optimization" and *especially* "bug fixing." The point of doing optimization and bug fixing is that you make minor changes to code. You don't do big things; you only change things if you absolutely have to. Switching engines creates new bugs. It's changing the code from something where the team has a pretty good sense how things work and what to do, and has incorporated all sorts of tweaks to fix issues, into an incompatible physics engine and a new game engine. It's not a simple task. If it were, this would be Unity 4.4 (I think). Likewise, the kind of optimization you do late in the release cycle is as unobtrusive as possible; big changes have heavy bug potential.

It's entirely possible that, had this come out six months ago, Squad would have changed over. Had it come out in December, Squad might have considered investing some dev time into seeing how it works, and trying to hack a port together to see how it runs (and see if it's a bugsplosion, a won't-compile-splosion, or runs decently well). But it came out in March, well into the final push for release. At that point, no matter what Squad had said about U5 a year ago, it'd be *insane* to go all-in on U5. I wouldn't be surprised if they tried compiling against U5, but if there were *any* serious issues, you'd be crazy to put off your 1.0 to change the game engine when you're nearing release (keep in mind, Squad also had and has a much better idea of the status of KSP's development than forumers; I suspect they knew they were nearing experimentals).

If they had made the engine more effecient, then who cares if it is called optimization or not?? They have not even done anything serious when it comes to optimization anyways, so going to a new engine wont hurt anything and can only IMPROVE it. Just because they are porting the game over, they dont have to release it until it has been tested and properly bug fixed. And just because it is a new version number of unity, does not mean it will be insanely difficult to port over. KSP was not so long ago running in unity 3 and the task of porting it over to version 4 was according to the devs for most part a very smooth process. Unity 5 has also been designed with ease of upgrading in mind, so that the process is as painless as possible.

And for your information it was squad themselves that stated that Unity 5 was such a high priority that they would immediatly start working on it as soon as they got access and would even halt development of other features temporarily so that they could spend every resource on porting it.

It was also squad that decided to suddenly out of nowhere push for 1.0, even though it was said it would remain in beta for quite some time. No one forced them to do this and it is in fact a bad idea to do so until the damn game is both polished and finished, which includes unity 5.

I really wonder why you people just assume that it is virtually impossible to port to unity 5, when everyone who has experience with it states the opposite.

Of course there might be a few bugs here and there and some changes in code, but end result would lead to a better game in literally every way and will even make future development easier since unity 5 takes care of quite a few bugs, issues and performance problems.

[Eric S]

"Unity 5.0 physics could be expected to work up to 2x faster than in previous versions"

Could vs should. The numbers I've seen indicate a worst case speedup in the 10-20% range, and unless U5 lets Squad run multiple PhysX threads per craft, we'll probably see something close to that worst case for KSP when you're only dealing with a single craft. Docking two craft of similar complexity will be quite a bit nicer since we should be able to run each craft in a separate thread, coming close to doubling the FPS.

Personally, I think people are getting overly wound up over the issue. 1.0 has gone to experimentals, so there's almost no chance of it getting ported to U5. The devs have stated that they'll be evaluating a switch to U5 soon after 1.0 gets out the door, so a U5-based 1.1 is quite possible (or possibly 1.0.X, though I suspect those versions, if they exist, will be strictly for bug fixes). At this point we're basically debating which version of KSP should be called 1.0 and questioning whether Squad's evaluation of U5 will be adequate.

I'm not saying that U5 won't be an improvement over U4 for KSP, and I definitely look forward to seeing KSP moved over to U5. I am saying that I strongly doubt it will be the magic wand of FPS and 64-bit stability fixes that some people are assuming it will be.

[Climberfx]

Even if it are faster, they would put more thing to our delight, like visual post fx, bloom, glow, global illumination, etc...

In the and we gonna have practically same fps, but a lot better visual game!

[Eric S]

The source is KSPTV/Squadcast.

Sorry, not doing your work for you, you're gonna have to be a little more specific, otherwise I'll just assume that you've got your own interpretation of some offhand remark that I've already seen.

I really wonder why you people just assume that it is virtually impossible to port to unity 5, when everyone who has experience with it states the opposite.

I don't see any significant number of people saying it's virtually impossible to do the port. I see people saying it's not a trivial port, and seeing as KSP uses PhysX, and that's the one big area where U5 isn't backwards-compatible with U4, unlike the U3 to U4 transition, which all by itself took over a month (almost two, but the christmas holiday season fell in the middle of that), I'd say that's a reasonable expectation.

Edited April 6, 2015 by Eric S

[boxman]

Sorry, not doing your work for you, you're gonna have to be a little more specific, otherwise I'll just assume that you've got your own interpretation of some offhand remark that I've already seen.

I dont have anything to prove. You are free to believe whatever you want.

I am not going to dig through months of ksptv stream archives that are most likely long since gone thanks to twitch deleting old archived streams.

[basic.syntax]

Regardless of what was said in the past, this is, I think, the most recent statement about U5:

[twitch question: Unity 5?]

"Yes, we are running internal tests on Unity 5. It's looking promising, but also... it kinda looks like every time... how to put it...like, there's a lot that CAN happen with Unity 5, but the work we have shown [internally] so far, is that it can... just about IGNITE the computer it is running on."

[Think he is suggesting stuff like KSP getting bogged down, GPU/CPU overload.]

"We are absolutely committed to getting to Unity 5. It can't happen this release cycle ... the timeline simply doesn't pan out for the amount of work it has to be, the same goes for the 64-bit update [client] but yes, we absolutely want it in."

[Eric S]

I dont have anything to prove. You are free to believe whatever you want.

Do you remember exactly what was said? I am actually interested in that. Then again, I've seen so many times that gaming communities turn offhand remarks by people not in the position to commit the development staff into "promises," including this one (this community, not this specific claim), that I tend to be rather skeptical. Which isn't to say that there haven't been companies or development teams that delivered product far short of their early claims for that matter.

What I'm saying is that I've been following the discussion on a U5 port since before squad publicly commented on it, and while I do remember some very enthusiastic (and fairly unofficial) comments about it, I've never seen anything to indicate that they'd drop whatever they were doing when U5 became available.

[boxman]

Do you remember exactly what was said? I am actually interested in that. Then again, I've seen so many times that gaming communities turn offhand remarks by people not in the position to commit the development staff into "promises," including this one (this community, not this specific claim), that I tend to be rather skeptical. Which isn't to say that there haven't been companies or development teams that delivered product far short of their early claims for that matter.

What I'm saying is that I've been following the discussion on a U5 port since before squad publicly commented on it, and while I do remember some very enthusiastic (and fairly unofficial) comments about it, I've never seen anything to indicate that they'd drop whatever they were doing when U5 became available.

It would be impossible to remember exactly the words that was said since it was months ago before unity 5 had gone to beta.

But basically what was said was that they could not wait to get their hands on it and kinda joked about how they have been bugging unity 5 daily to see when they could get their hands on it. Then went on to say that they are dedicated to the port as they believed this would allow the game to finally utilize multicore/threads for physics due to the changes in unity 5. They then explained that even though it will be multicore and will increase performance with large part counts we should not expect double the performance from lets say dual core systems. Also talked about how they were excited about the improved physics engine as well and that this would most likely also improve performance as well as the improved graphical capabilities which could allow them to at a later date make the game prettier by using the GPU which was more or less underutilized right now.

Then they finally talked about how that it was such high priority that they would focus most of their resources on the port and even halt other development once unity 5 is released until that work has been done.

This was of course not a promise since they usually end such announcements with a disclaimer saying that things can change further down the line.

The order might be a bit out of order but this is roughly what was said during that squad cast and they seemed really excited and optimistic about it.

But it really shows that Squad themselves understood that Unity 5 is important for this game and that it would be better to port it sooner than later.

With the recent change and push towards 1.0 I fear this port will never happen...

[Eric S]

With the recent change and push towards 1.0 I fear this port will never happen...

I don't. The devs have stated that they're going to evaluate U5 after 1.0 gets out the door, and haven't contradicted that in any way. They've been repeating that 1.0 doesn't mean the end of new development and features. They may or may not prioritize U5 higher or lower than the features that they cut from 1.0, or they may stick them all into 1.1. They haven't said that they will definitely port to U5, but I suspect that that is because saying that before having the chance to evaluate the impact of U5. I strongly believe that the evaluation will come out in favor of U5 and that they'll do the port in 1.1 or 1.2.

[boxman]

I don't. The devs have stated that they're going to evaluate U5 after 1.0 gets out the door, and haven't contradicted that in any way. They've been repeating that 1.0 doesn't mean the end of new development and features. They may or may not prioritize U5 higher or lower than the features that they cut from 1.0, or they may stick them all into 1.1. They haven't said that they will definitely port to U5, but I suspect that that is because saying that before having the chance to evaluate the impact of U5. I strongly believe that the evaluation will come out in favor of U5 and that they'll do the port in 1.1 or 1.2.

I am skeptical, but I really hope you are right. They just no longer seem that optimistic about unity5.

And I think it is just a bad idea to release 1.0 first, since we all know that a new engine are bound to have some issues or bugs in the first KSP build.

But I guess it is too late now to delay the 1.0.

[ratchet freak]

If you want full on speculation about performance updates then consider unity 5.X using vulkan for rendering and using compute shaders for physics (like I suspect PhysX is doing behind the scenes).

[Climberfx]

Look guys:

lfMh

[NoMrBond]

The Worldmachine demo was pretty cool as well

And the Tenkoku Dynamic Sky

If you want full on speculation about performance updates then consider unity 5.X using vulkan for rendering and using compute shaders for physics (like I suspect PhysX is doing behind the scenes).

Unfortunately there is no timeline for Unity 5 Vulkan support (or any OpenGL update), DX12 support is due immediately following the Win-10 launch though which is coming up in a couple of next months.

The branch of PhysX that Unity is using is CPU only so no GPGPU accelleration is used at this time, it's still a major update from 2.8.3 (Unity 4) to PhysX 3.3 (Unity 5) though, with support for multi-threading and SIMD FPU acceleration.

Future APEX support with GPGPU acceleration via OpenCL has been mentioned, sadly not confirmed as yet though, and might only effect additional (APEX) modules (such as

, Clothing, and

) and not the main PhysX component. Edited April 12, 2015 by NoMrBond

[boxman]

The Worldmachine demo was pretty cool as well

Yet another good reason for why unity5 should be a high priority. I know it will probably not be possible to get it as good looking as these videos due to scale of planets, but I am sure they are able to more easily add some nice effects, like for example fog/mist, clouds, water and maybe even some dust being kicked up from rover wheels or rocket.

[Climberfx]

Great Unity 5 demos. I know Substance is not Unity 5, but is directly connect to the game development process. so i share here my first test on substance texture inside Maya. This is preparation to prepare the textures for my developing parts (Space Car), when i will use the Unity to get to KSP game.

Basic Diffuse Specular Normal texture created over a normal map generated from a high poly model for this low poly mesh.

[NoMrBond]

Reminds me of Pixologic's ZBrush or the Quixel Suite, I wish I could afford either.

[Climberfx]

Zbrush is more modeling tool, this one is procedural textures. It's model parameters drive de details in the texture. Helps a lot.

And have a indie option of subscription. Not expensive.

[Beduino]

Okey decided to learn some more of Unity and in a kerbal like way try the physics performance again.

+ Terrain

+ 1 Rocket (165 cylinders) using fixed joints (which are kind wobbly but at least the rocket explodes)

+ 5 White Stations 140 cylinders each plus configurable joints (set to work as fixed joint should, all motion axis locked and projection mode is set to very small values for position and rotation)

Rocket and one White Station is driven by adding forces.

+ 3 orange kraken like contraptions containing each 265 cylinders and a combination of fixed and spring joints with collisions between bodies enabled.

+ Floaty and static objects

All joints are breakable.

Total ~1700 physics objects.

30 FPS average with fluid physics most of time.