Croquet Programming Document

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Croquet Programming Document

Derek O'Connell
Is there an updated version of Croquet Programming pdf? Also is there any performance data on
Croquet, eg, simple FPS vs NumberOfPolygons?


               
___________________________________________________________
Inbox full of spam? Get leading spam protection and 1GB storage with All New Yahoo! Mail. http://uk.docs.yahoo.com/nowyoucan.html

Reply | Threaded
Open this post in threaded view
|

Re: Croquet Programming Document

Joshua Gargus-2

On Sep 25, 2006, at 4:05 AM, dmoc wrote:

> Is there an updated version of Croquet Programming pdf?

Not that I am aware of, but I'm not quite in the know about it.  Are  
there specific parts of the guide that you found to be inadequate?  
Your input would be very helpful in improving the guide.

> Also is there any performance data on
> Croquet, eg, simple FPS vs NumberOfPolygons?

This is a meaningless performance metric.  If you have a scene with a  
handful of 100000-polygon meshes, then Croquet will perform as well  
as any other 3D engine.  If you have 10000 small meshes with the same  
total polygon count, then it will perform slower than many.  Note  
that rendering performance has received very little attention; the  
focus has been on the messaging layer of the system.  This is due to  
time constraints rather than to lack of interest.

I'm not sure what metric is better to use.  Croquet uses TCP, and  
therefore will not be able to achieve the extremely low latency that  
is desirable for a 1st-person shooter.  On the other hand, the  
latency is very comfortable for Second Life or World of Warcraft  
types of use cases.  Performance will also depend if you are  
streaming audio/video, playing video, computing complicated  
simulations, etc.  Unfortunately, the best advice I can give is to  
try it.

Josh


>
>
>
> ___________________________________________________________
> Inbox full of spam? Get leading spam protection and 1GB storage  
> with All New Yahoo! Mail. http://uk.docs.yahoo.com/nowyoucan.html
>


Reply | Threaded
Open this post in threaded view
|

Re: Croquet Programming Document

Derek O'Connell
In reply to this post by Derek O'Connell
Josh, thanks for replying...
 
> > Is there an updated version of Croquet Programming pdf?
>
> Not that I am aware of, but I'm not quite in the know about it.  Are  
> there specific parts of the guide that you found to be inadequate?  
> Your input would be very helpful in improving the guide.

I don't think the document is inadequate. I started reading it today and did not what to start
commenting on what may be an old version (noting the creation date is 14/4/06). This was why I
asked.

> > Also is there any performance data on
> > Croquet, eg, simple FPS vs NumberOfPolygons?
>
> This is a meaningless performance metric.  If you have a scene with a  
> handful of 100000-polygon meshes, then Croquet will perform as well  
> as any other 3D engine.  If you have 10000 small meshes with the same  
> total polygon count, then it will perform slower than many.  

IMHO and in my experience this is not meaningless but I of course I agree re polygon vs objects.
Maybe my question should be "as a 3D graphics engine how well does it scale?". Specifically, I
am
thinking as my first test of Croquet (graphics side to begin with) to convert my project which
has
~450 objects, of which ~300 are relatively simple but textured and the remainder rather more
complex and also textured. In my own code I get around ~170 fps but in previous versions using
other peoples "engines" (and with reduced objectives) I have had as low as 17 fps. So it is a
very
practical concern to me before embarking on what is a significant amount of work to canvas
opinion
on performance before starting, even if the response is subjective.

> Note that rendering performance has received very little attention; the  
> focus has been on the messaging layer of the system.  This is due to  
> time constraints rather than to lack of interest.

Understandable and I hope, once I am familiar enough with the Croquet rendering classes, to give
it some attention and maybe contribute to Croquet. I have a little experience of optimising
things
from the OpenGL side, assuming if in fact it is needed.

> I'm not sure what metric is better to use.  Croquet uses TCP, and  
> therefore will not be able to achieve the extremely low latency that  
> is desirable for a 1st-person shooter.  On the other hand, the  
> latency is very comfortable for Second Life or World of Warcraft  
> types of use cases.  

For my own purpose I categorise graphical environments between two extremes "Hands-off eye
candy"
and "flexible to point of being unuseable" (without being disrespectful to either). Examples of
the former would be the excellent products from the demo scene. For the latter I would say
VRML/X3D and the various players which allow for interaction but are ok only for relatively
simple
geometry (ie, they max-out fairly quickly). I suppose another dimension would be the level of
ability to truly interact with objects (as opposed to simply blowing things up - lol). My
interests lie somewhere inbetween and maybe a little off center in what use to be called
"visualization and direct manipulation" and what I now prefer to call "engaging(ly) interactive
data environments". To date I have not been concerned with multiple-user interaction but this is
a
nice-to and the Croquet environment (and indeed Squeak/Smalltalk) looks very promising in that
respect. Even if it turns out it cannot cope with my specific requirement I still intend to
learn,
develop and experiment with Croquet. I think what has already been achieved is fantastic.

> Performance will also depend if you are  
> streaming audio/video, playing video, computing complicated  
> simulations, etc.  Unfortunately, the best advice I can give is to  
> try it.
>
> Josh

Already am but thanks for your reply Josh!



               
___________________________________________________________
Try the all-new Yahoo! Mail. "The New Version is radically easier to use" – The Wall Street Journal
http://uk.docs.yahoo.com/nowyoucan.html

Reply | Threaded
Open this post in threaded view
|

Re: Croquet Programming Document

Milan Zimmermann-2
In reply to this post by Derek O'Connell
On 2006 September 25 04:05, dmoc wrote:
> Is there an updated version of Croquet Programming pdf? Also is there any
> performance data on Croquet, eg, simple FPS vs NumberOfPolygons?

Well, FWIW, this is a link I setup way back, to track people's experience
(FPS) on different hardware

http://minnow.cc.gatech.edu/croquet/17

Milan

>
>
>
> ___________________________________________________________
> Inbox full of spam? Get leading spam protection and 1GB storage with All
> New Yahoo! Mail. http://uk.docs.yahoo.com/nowyoucan.html

Reply | Threaded
Open this post in threaded view
|

Re: Croquet Programming Document

Derek O'Connell
Well, FWIW, this is a link I setup way back, to track people's experience
(FPS) on different hardware

http://minnow.cc.gatech.edu/croquet/17

Milan
Thanks Milan.