First of all, I'm not saying "let's switch to OpenRT now".  I am saying that 
if FlightGear's scenegraph ever requires a large restructure, it will be a 
good time to look at the feasibility of using OpenRT.

On October 20, 2004 11:50 pm, Boris Koenig wrote:
> nope, it wasn't required - after all it is supposed to be
> "software-raytracing" and not hardware, but I *assume* without
> a corresponding hardware board your CPU would probably be abused
> as a GPU - almost exclusively ...
No doubt.

>
> > The reason I brought the OpenRT topic up again is that (as far as I
> > understand) it can run on most people's desktop.
>
> yes, but I somehow doubt that there would remain much idle CPU time -
> you'd probably at least need a dual-processor board - but again, I'm
> *guessing*.
>
> Probably, they wouldn't start developing prototype hardware boards:
>
>  http://www.saarcor.de/
>
>> Now, if it can render a 350 million polygons model using a CPU that is 
slower 
>> than mine, then it should be able to handle FlightGear with ease.
>
> I am not so sure about that - in order to get representative data, one
> would need to have at least some simple tests available or maybe even a
> simple game that employs the software-based raytracing approach.
>
> On http://graphics.cs.uni-sb.de/RTGames/ there are various games
> mentioned, even screenshots are shown - and even very promising
> divx-videos - but there doesn't seem to be a simple demo for personal
> evaluation !?
>
> This might be the case because all these videos seem to have been
> created using clusters of a dozen or even more 1 GHZ machines ...
Even if FlightGear were to utilize OpenRT, I highly doubt it will ever have to 
render more than several hundred thousands polygons at a time.

Just look at Oasen at 
http://graphics.cs.uni-sb.de/~morfiel/oasen/features.html :
Some numbers to frighten children: 
500.000 polygons heightfields 
> 1000 objects, ~30.000 polygons each 
level complexity: 15 Mio polygons 
138 dynamical lightsources, all of them casting dynamic shadows from all 
objects to all objects 

The numbers already give a hint that this "game" is made for stressing the 
graphic library to the limit, not for you and I to play.


> Doesn't sound that feasible to me :-/
>
> But that page is also where you can find the quote I mentioned:
>
>  "We are very much interested in evaluating new ways for computer
>  games and therefore like to cooperate with the gaming industry.
>  Thus if  you are in such a position, please send us an email! "
>
> Maybe they've got a mailing list ? Possibly they would provide you with
> more details if you contact them...
>
> But I doubt that it is the ultimate solution - there are certainly
> drawbacks, and it's not necessarily suited for any purpose, either.
>
> Otherwise it would probably already be a lot more popular !?
Well, I guess the answer to that question will remain unknown until someone 
decides to look more deeply into it.

Ampere

_______________________________________________
Flightgear-devel mailing list
[EMAIL PROTECTED]
http://mail.flightgear.org/mailman/listinfo/flightgear-devel
2f585eeea02e2c79d7b1d8c4963bae2d

Reply via email to