Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[charlesmeryon]

Hello,

Your engine is very impressive !
Is that possible to have any demo of a small area with one vehicle (like airplane ) to test it ?
I hope it will not be like Infinity engine with impressive images and videos during many years without any chance to test it!

PS: sorry for my english.

[angrypig]

Your engine is very impressive !
Is that possible to have any demo of a small area with one vehicle (like airplane ) to test it ?
I hope it will not be like Infinity engine with impressive images and videos during many years without any chance to test it!

Hi,

Thanks.

We would like to release a public demo when we stabilize the functionality and resolve some of the more serious issues. A major problem are the AMD OpenGL 3.2 driver bugs. As the engine quite depends on the new features, these bugs render the AMD graphic cards unusable now. As many of users have AMD cards we have to have this fixed. We are working to resolve these issues with AMD but their reaction times are rather slow. Still, it's better than when they were ATI.

So what will probably be in the demo:
* Tatra 815
* Cessna 172
* A free camera controller (UFO mode with limits)
* terrain data will be downloaded by bittorrent (but we are still resolving how and where to seed the data)
* there will be probably only one land type with one or two tree types and there won't be grass blades (only as texture)

Regarding the performance, as there are few posts in other forums whether Outerra needs a super computer to run. The minimum configuration is a dual core CPU and a SM4 graphics card with at least 512MB memory and performance similar to Nvidia 8800GT. With this configuration you will get at least 30FPS. This is what we are getting in the current development version which is not optimized very well, but there is also still lot of stuff missing like the building etc.

[Abc94]

Yeah a demo would be nice.

I remember reading a YouTube Post by Cameni saying that they hope to have a demo by January/February 2010.  That post is about 7 months old now though.

Do you think there will still be a release in February?

Whoah, two posts a minute apart!

[helio2]

Well i hope u guys fix the ATI gfx card problem, because i got their Latest Card that barely came out like 3 months ago. And i want to SET IT LOOSE! It hungers for graphics intense games!

[cameni]

Well, we are currently at this with AMD cards:



versus


At least we aren't in a complete darkness like in the beginning
That's after a few bug reports filed, fortunately the guys from AMD seem to be quite helpful. We have some workarounds for all the bugs we've found so far, shame it took such a long time to isolate the causes.

We have found also a few bugs of our own; the most problematic issue is the "undefined behavior" that appears in many places in GL specification, and somehow on AMD it usually manifests as an unexpected one, whereas on Nvidia it does the presumed behavior anyway. Still, it's our bug, but the missing feedback from OpenGL that you are doing something wrong or not doing something at all really doesn't help to shorten the time spent on finding such bugs. It's real pain sometimes.

And from the screen shot you can see there will be some more of ones and/or the others.

[Abc94]

Haha that's a crazy screenshot, with the flying trees and the colored forests.

That's really good that you have a way to fix all the "known" bugs.   I have an AMD.

When you say AMD you mean AMD processors right?

[cameni]

No, AMD graphic cards. I should have rather said ATI
Their OpenGL driver is problematic.

Well, and there's apparently a lot of "unknown bugs" too

[Abc94]

Oh ok.

I have an ATI 4850.  I hope your able to fix it soon.  

[helio2]

Yeah AMD bought ATI about 2 years ago. So from the pictures it seems to be some kind of texture problem, but then again I'm not a program just 3d artist , lawlz

I actually used to get a similar problem that resembled that on my old graphics card when it overheated running some intense graphic applications. But its probably not related.

[djt]

Phenomenal looking game engine.

With it being OpenGL based it’s too bad you couldn’t team up with the X-Plane development team, I think it would be a great fit.

[cameni]

I think no existing simulator (or a simulator built on old principles) can easily use the engine, and it would be a huge amount of work to integrate it as well. But there are still some possibilities, like PSX where one can make a plugin for world rendering relatively easily, or the Jet Thunder, possibly.

BTW there will be also a DX10+ back-end later.

[djt]

I think no existing simulator (or a simulator built on old principles) can easily use the engine, and it would be a huge amount of work to integrate it as well. But there are still some possibilities, like PSX where one can make a plugin for world rendering relatively easily, or the Jet Thunder, possibly.

What would be an example of “old  principles”?

BTW there will be also a DX10+ back-end later.

Interesting, I would really like to hear more about this. I have been under the impression that it's either OpenGL or DirectX when you are developing a game engine not both?

[cameni]

What would be an example of “old  principles”?

Hmm, for example, we are doing everything to be fractal refineable, so there's always detail as you zoom in. Another property of the approach is that level of detail management and streaming is very effective. However, this also means the data has to be prepared, preprocessed - it cannot use just anything. So you cannot simply take existing meshes and textures and throw it at the engine.

I have been under the impression that it's either OpenGL or DirectX when you are developing a game engine not both?

It's certainly not, there are engines that more or less hide from you the underlying API. Nowadays, with everything going in one direction - GPU, shaders and everything, it's in fact easier to do than before, if one can get rid of the obsolete stuff entirely.
We've modeled our core engine to follow the DX10 API style, but have implemented OpenGL back-end under it. So now it should be easier to implement DirectX 10+ back-end when the whole thing was designed this way.

[djt]

Hmm, for example, we are doing everything to be fractal refineable, so there's always detail as you zoom in. Another property of the approach is that level of detail management and streaming is very effective. However, this also means the data has to be prepared, preprocessed - it cannot use just anything. So you cannot simply take existing meshes and textures and throw it at the engine.

Thank you for the explanation, again very interesting stuff here.

It's certainly not, there are engines that more or less hide from you the underlying API. Nowadays, with everything going in one direction - GPU, shaders and everything, it's in fact easier to do than before, if one can get rid of the obsolete stuff entirely.
We've modeled our core engine to follow the DX10 API style, but have implemented OpenGL back-end under it. So now it should be easier to implement DirectX 10+ back-end when the whole thing was designed this way.

Does this approach have a negative effect on the performance of the engine in the end; is it less efficient not being “dedicated” to OpenGL or DirectX?

When you say “DX10+” do you mean DirectX 10 and 11?

Thanks again for taking the time to respond.

[cameni]

There should be no big impact; it's not like we are creating a thick layer to hide all aspects of the APIs. The engine is mostly the shaders and the graphics API-independent code. We are also using code generators that hide the specifics without requiring heavy abstract code.

Yes, DX10+ is anything from DX10 and above. Mainly that means we aren't addressing DX9 because that has quite different and old architecture.