Kurien, that's the very point of Java3D (and possibly of any other
true 3D application)! The performance highly (if not thouroughly) depends on the
complicity of the scene. Moreover, performance would be different on
different video cards with different amount of onboard RAM. For
instance, hellouniverse example goes quite decently even on the cards
without any hardware acceleration, while anything that use textures,
etc, wouldn't go at all (or refresh rate would be around 1/sec). So,
do simplify your scene and do test it on many computers. Moreover i
can't see any sence in detecting some 'absolute' performance not
related to a real scene (and in fact u do need to build some scene
before testing the engine itself).
CU,
vladimir
-=V=-
>-------<=============>-------<
Join in Java community now!
http://JavaCafe.VirtualAve.net/
>-------<=============>-------<
In your previous letter u wrote:
--------------------------------
KJ> Thanks Vladimir,
KJ> I have one more doubt.:
KJ> Assume if 1 cube is there in a scene and I attach the Keyboard navigation
KJ> of the scene. The movement inside the Scene very smooth
KJ> Now if 100 cubes are kept in the scene, the movement inside the scene is
KJ> quite breaky(not smooth), so as to assume it takes x time duration for the
KJ> View to render(or to reorient or whatever jargon) after each KeyPress. So
KJ> this value of x will vary in both cases of 1 cube with that of 100cubes, and
KJ> I could assume x(1) is very less compared to x(100).
KJ> Is there any way to measure this x ?
KJ> Thanks once again.
KJ> Regards
KJ> Kurien
KJ> -----Original Message-----
KJ> From: Vladimir Olenin <[EMAIL PROTECTED]>
KJ> To: [EMAIL PROTECTED] <[EMAIL PROTECTED]>
KJ> Date: Tuesday, October 12, 1999 2:33 PM
KJ> Subject: Re: [JAVA3D] Performance Issues
KJ> Dear Kurien!
KJ> I guess the best way to calculate performance is to measure frame
KJ> refresh
KJ> rate. Anyway it's the only parameter that end user cares about. U can do
KJ> it
KJ> either by implementing postRender/preRender functions of Canvas3D class or
KJ> by
KJ> handling FrameElapsedEvent. As an output u can use standard System.out
KJ> stream
KJ> (the easiest way) or to send data to some customized more pleasantly
KJ> looking
KJ> component (like ProgressBar, for instance).
KJ> -=V=-
>>-------<=============>-------<
KJ> Join in Java community now!
KJ> http://JavaCafe.VirtualAve.net/
>>-------<=============>-------<
KJ> In your previous letter u wrote:
KJ> --------------------------------
KJ>> Dear Gurus,
KJ>> I want to know how one can find the performance of the Java3D
KJ> application developed. My Objective are as follows
KJ>> 1.Compare the Performance after using a Single geometry and appearance
KJ> instance and create multiple Shape3D instances with that of multiple
KJ> instances of Shape3D objects(i.e clone the shape3d
KJ>> object)
KJ>> 2. Peformance due to usage of compile.
KJ>> 3. Performance of using LOD
KJ>> I have seen a performance monitor in the Java2D demo applet which come
KJ> along with JDK1.2. How is this performance calculated ?
KJ>> What is the criteria under which I should calculate Performance for a
KJ> Java3D application.
KJ>> Thanks a Lot
KJ>> Regards
KJ>> Kurien
KJ> ===========================================================================
KJ> To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
KJ> of the message "signoff JAVA3D-INTEREST". For general help, send email to
KJ> [EMAIL PROTECTED] and include in the body of the message "help".
KJ> ===========================================================================
KJ> To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
KJ> of the message "signoff JAVA3D-INTEREST". For general help, send email to
KJ> [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".
