How long before a performance increase?

[Motokid600]

I guess we can debate all we want, but I guess the end all be all is the engine. So.. how big of an overhaul is it to swap out what's under the hood?

[ScottyDoesKnow]

I have a pretty good computer, about a year and half old, self built. My space station is probably under 200 parts and I get 5-7 fps near it. I understand the "if you give them more, they'll just want even more" side of it, but I think being able to do something like 1000 parts before it lags this bad is more reasonable. Unfortunately if you use a high level framework like Unity to make a complex game, you're gonna have these problems. Unity is good for making relatively simple (as in the most processor intensive part is the graphics) games quickly.

http://en.wikipedia.org/wiki/Minecraft

Point delivered. Its one of those games where you can make 32bit computers.

This is a ridiculous example to use. I haven't played in a while but view range was a limiting factor for sure, it can only load so many chunks into memory. And it's limited for the exact same reasons as KSP, a high level language was used to make it (java) and it's not well optimized. And try building that 32bit computer and see how fast it actually runs, if I remember correctly you'd get a lot of lag with large redstone creations. On top of that, multiplayer compounds the problems.

[Nao]

Do you have win32 or something?

Lol the game itself is "32", i also have seen no performance change between for example XP32bit and Win764bit on the same machine.

Last version had quite a bit of performance boost with large ships for me.

And you have to realize that squad never realized that people would build ships with thousands of parts or that it was even possible to send such massive rockets/objects into space.

And what do you think will happen with every performance increase? People will just build bigger and bigger and always reach that limit where performance goes down. If they make 1500 part ships run smoothly then people like me and most likely you will just start building 3000 part ships.

It has already improved alot since .17 and i bet it will improve a bit with future versions as well.

If you play this game like it was intended to then there should be no problems getting smooth framerate. This game runs acceptable even on my old intel core 2 duo when launching medium sized rockets.

Agreed that people will always go up to the limit. The problem is that there is really big difference between different users limits. And im not talking about somebody trying to run this game on a notebook.

I have C2Duo too, even OC'd by 33% but for me lag starts sometimes way to early. Possibly its a driver or OS issue, but i tried many different things and it still not smooth where it should be. (I'm on W764 with all the CPU intensive stuff disabled, GTS450)

I honestly believe that half of the FPS problems are caused because of bad optimization. I've launched 800+ part ships into orbit with ~10 fps but at the same time the game starts lagging with ships below 150 parts that have a lot of lights, ladders or solar panels. There also problems with debris near the launch pad, that sometimes create heavy lag even on low part counts.

Scrapping the engine now would be an overkill for a small studio like SQUAD but i think there is still a lot to optimize at least for some people. Also the game is still in Alpha so this kind of problems are understandable from design standpoint, but it would be nice to get some more updates .

[Clamps]

What a nice attitude you have, charming indeed.

I didn't mean to come off that way, it just seemed silly that you would totally disregard the option of hiring. I know it's expensive, but ksp has sold a lot of copies. I wish they would publish the actual figures so we know how much they made.

[Sir Nahme]

I didn't mean to come off that way, it just seemed silly that you would totally disregard the option of hiring. I know it's expensive, but ksp has sold a lot of copies. I wish they would publish the actual figures so we know how much they made.

no project survives an engine change

[Themohawkninja]

no project survives an engine change

I've heard that Duke Nukem did.

OT: Probably not until the game is nearly done.

[Clamps]

no project survives an engine change

Someday we will have...

Kerbal space program 2!

[notfruit]

Why do people on the internet get so agitated about little things?

Sometimes I wish I didn't live on the internet.

[numerobis]

You could change the physics engine without ditching Unity entirely. But if I were the boss, I'd do just about anything else to make the game better.

There have been fairly significant performance improvement in two recent releases, so it's not like Squad is just ignoring the issue.

[Uberick]

I'm not a programmer or anything, but I have to think that changing out the whole engine at this point would be an exercise in futility; the developers would probably be better off heavily modifying the existing engine to better optimize it.

I'd assume that there's at least enough room for optimization to probably double current frame rates, assuming that they haven't done much yet(which I don't really know). If they could get the optimization to the point where most people could assemble a decently sized 400-500 part station without lagging too bad, I think that'd be a great start.

[Sir Nahme]

I've heard that Duke Nukem did.

OT: Probably not until the game is nearly done.

Have you PLAYED DNF? I reiterate, no project survives an engine change.

Edited June 22, 2013 by Sir Nahme

[Overlag]

my game struggles with 150 parts, dunno how you are getting 1000 parts! lol

[asmi]

no project survives an engine change

Wrong. I can name few that did right off top of my head.

[Sir Nahme]

Then do so

[asmi]

Then do so

Entropia Universe

Eve Online

[m4v]

Changing the engine would be a pain, all KSP mods out there would need to be rewritten from scratch.

[Sir Nahme]

Entropia Universe

Eve Online

EVE never changed engines. It had various updates to the graphics engine but that is only the graphics, the core engine is an inhouse drive that CCP maintains.

Hell if you can consider upgrading from Trinity 1 to Trinity 2 as an engine change then EVERY time KSP has been recompiled in a new version of Unity it has "changed engines".

Entropia was rereleased with a new engine and is effectively closer to a sequel than than a "project that switched engines"

[Stealth2668]

Changing the engine would be a pain, all KSP mods out there would need to be rewritten from scratch.

I'd rather have a smooth game than a heavily modded slideshow.

[asmi]

EVE never changed engines. It had various updates to the graphics engine but that is only the graphics, the core engine is an inhouse drive that CCP maintains.

Hell if you can consider upgrading from Trinity 1 to Trinity 2 as an engine change then EVERY time KSP has been recompiled in a new version of Unity it has "changed engines".

No, there were very significant changes both in server side and client side.

Entropia was rereleased with a new engine and is effectively closer to a sequel than than a "project that switched engines"

Well THAT is what changing engine does! Entropia as project continued, it's not a sequel.

More importantly, you're proven wrong

[Sir Nahme]

Upgrades to engine =/= engine change

In that case Crysis 2 was just crysis 1, same with far cry, unreal tournament, hell any game sequel in existance! Why, when they released Baldurs Gate with the new engine why did I have to pay again?

FYI KSP isn't an mmo

[Themohawkninja]

Upgrades to engine =/= engine change

In that case Crysis 2 was just crysis 1, same with far cry, unreal tournament, hell any game sequel in existance! Why, when they released Baldurs Gate with the new engine why did I have to pay again?

FYI KSP isn't an mmo

So, a new version of the engine doesn't count as a new engine? So, GeoMod 2 and 2.5 (or is it 3?) for all intents and purposes the same as GeoMod 1?

Also, how is a different genre relevant when talking about changing engines?

[Sir Nahme]

So, a new version of the engine doesn't count as a new engine? So, GeoMod 2 and 2.5 (or is it 3?) for all intents and purposes the same as GeoMod 1?

An update to an existing engine applied to a game title does not equate to a change of engine

Also, how is a different genre relevant when talking about changing engines?

Very simply and effectively. There is a massive difference in development cycle between the two stated genre's

[Themohawkninja]

An update to an existing engine applied to a game title does not equate to a change of engine

I suppose... but some engines have some really big "updates" with graphics, physics, and a decade's worth of other upgrades, and simply calling that an update just seems... like an understatement.

Very simply and effectively. There is a massive difference in development cycle between the two stated genre's

I understand that, but what does a different development cycle have to do with changing the engine that the developers use? I can understand in the case of multiplayer vs. non-multiplayer games where server support comes into play, however even though you do use MMO's vs. KSP, which would fall under that category, I do know that Unity supports multiplayer (it has it's own mini-engine designed specifically for multiplayer).

[Fel]

no project survives an engine change

Didn't SimCity 3000 start in 3D and then changed over to 2D Sprites?

In a general sense, yeah... changing engines means a large rewrite; they even had "a great deal of programming" to go from Unity 3 to 4, I recall.

Changing the engine would be a pain, all KSP mods out there would need to be rewritten from scratch.

Yes, no.... most won't. We rarely use Unity to actually do anything with Mods, except when KSP's API fails to provide us with the mechanisms we want.

Edited June 23, 2013 by Fel

[Sir Nahme]

I suppose... but some engines have some really big "updates" with graphics, physics, and a decade's worth of other upgrades, and simply calling that an update just seems... like an understatement.

The primary difference is that most of what you've done in the "previous" engine is still valid and does not require reworking when it is updated to the "upgraded" engine. Where as an engine change (to a new engine) pretty much invalidates all previous work. Entropia is a good example of that, some systems they had when they changed (nearly 4 years ago) still aren't back up and running and many of the ones that have been reintroduced took quite some time.

The second is lifespan. A pay to play (or entropia's pay to win model) can handle long drawn out development cycles. The money made over it's life span exceeds the single purchase games margin, and after 6 years of active game play they can AFFORD to change engines to some degree. Most simply call it a sequel, entropia simply "retconned" the changes to keep the same name, but it is in no way any less a sequel.

I understand that, but what does a different development cycle have to do with changing the engine that the developers use? I can understand in the case of multiplayer vs. non-multiplayer games where server support comes into play, however even though you do use MMO's vs. KSP, which would fall under that category, I do know that Unity supports multiplayer (it has it's own mini-engine designed specifically for multiplayer).

Multiplayer has nothing to do with it. There are major differences in process, length, order, and complexity between a simulation's development cycle and an MMO's.

edit: I checked yeah sim city 3000 was, I can accept that as an acceptable example. However it does have the advantage of EA backing, they have the money to get away with that kind of shenanigan drawing out development time.

Edited June 23, 2013 by Sir Nahme