Hi Jacob,
Thanks for your answer. I have just joined the Discussion List and I think
that my post has been lost in others subjects. Maybe cause it deals with
OpenGL Display List as well. But it doesn't deal with the bench????
I will repeat it:

I would like to display a lot of times the same object at different
positions.
For instance :
display 1 time --------+
                       |
                       + Transform3D 1 +
                       |               |
                       |               + Cube 1
                       |
                       + Transform3D 2 +
                                       |
                                       + Cube 2
display 2 time --------+
                       |
                       + Transform3D 1 +
                       |               |
                       |               + Cube 1
                       |
                       + Transform3D 2 +
                                       |
....                                   + Cube 2
display n time --------+
                       |
                       + Transform3D 1 +
                       |               |
                       |               + Cube 1
                       |
                       + Transform3D 2 +
                                       |
                                       + Cube 2


I have tried to do a sharedgroup, sharing geometry but it doesn't speed up
the display.
On opengl, I was using display lists and calling n times the DL.
Is there a way to produce the same effect?
Thanks for your help!

-----Message d'origine-----
De : Jacob Marner [mailto:[EMAIL PROTECTED]]
Envoyé : mercredi 6 mars 2002 15:26
À : [EMAIL PROTECTED]
Objet : Re: [JAVA3D] Java 3D performance.


Hi Oliver,

I am not sure I understand this post. But if I get it right
you say that because the benchmark repeats the same
geometry raw OpenGL compare better.

I think you should know that in C++ and GL4Java
the boxes are not the same geometry. They are represented
as seperate display lists. In OpenGL it is common
practice (when tuning for performance anyway) to propagate
transforms into static geometry data manually. java3D 1.3
should do this automatically assuming that no capability
bits are set in the scene graph below the transform group
and the geometry is not done by reference.

So there is one display list for each box in the OpenGL
application - just as had the geometry been completely
different.

Did I understand your post correctly? If not, please
clarify.

Jacob

----- Original Message -----
From: "Olivier Bourion" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Wednesday, March 06, 2002 15:14
Subject: Re: [JAVA3D] Java 3D performance.


Hi,
The trouble thing is that in my case, the performances are the same if I
either use a SharedGroup, I share the Geometry or that I repeat the
"geometry" with no sharing at all.
I can garantee you that In that case, OpenGL Display lists are efficients.
The worst is that on pure performance, OpenGL and J3D arevquite similar (I
mean displaying over a thousand of random lines). It is jsut in the case of
a repetition of geometry that J3D perf falls down versus OpenGL.

Thanks for your comments!!!

Oulr

-----Message d'origine-----
De : Joachim Diepstraten
[mailto:[EMAIL PROTECTED]]
Envoyé : mercredi 6 mars 2002 15:03
À : [EMAIL PROTECTED]
Objet : Re: [JAVA3D] Java 3D performance.


Hi

> scen graph. At run-time the node should have been converted to display
> lists (I hope) and the original representation will have no importance.

Just a comment:

Well In my experience display lists in OpenGL are not always good
performance wise (at least on NVIDIA Cards)
it really depends on how many display lists and how big the display lists
are. Sometimes it's better not to use display lists at all.

EOF,
 J.D.

--
Explore SRT with the help of Java3D
(http://wwwvis.informatik.uni-stuttgart.de/relativity/minkowski)
(http://www.antiflash.net/java3d/relativity (mirror)

===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

=========================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

==========================================================================To 
unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

Reply via email to