From: "Fabrizio Nunnari" <[EMAIL PROTECTED]>
Sent: Wednesday, March 07, 2001 11:45 AM
> On Wed, 7 Mar 2001, Fred Klingener wrote:
>
> > Thanks, Fabrizio. If others want to volunteer to run this for other
OSs,
> > email me the results at [EMAIL PROTECTED], and I'll summarize on my
> > website. I've included a cleaned up version of the source.
>
> Uhmm... in my assembler experience I learned that a procedure call can be
> more time consuming than the the code of the procedure itself ;-)
>
> I guess the following version is more precise:
Uhmm... I _did_ say that it was ugly :-). I had reservations about exposing
my deep lack of understanding of key Java programming concepts, but I _did_
want to find out how other OSs worked.
> 42 49 *************************
> 43 49 *************************
> 44 49 *************************
> 45 49 *************************
> 46 49 *************************
> 47 49 *************************
> 48 49 *************************
> 49 49 *************************
>
> what a nice shape. We wait 29, 39 or 49... nothing in between !?!?!
> Uhmm....
Yeah. Cool, huh? I'm interested in high fidelity physics simulations, and
I use a simulation clock based on a Behavior with WakeupOnElapsedTime(dt
/*milliseconds*/) and t +=dt at the top of the processStimulus(). The
WakeupOnElapsedTime(dt) works exactly like wait(dt), so unless I pick a dt
just below one of the steps shown on the plots, the sim clock can slip a lot
with respect to the system clock. I have to measure the OSdt before I set
up the clock. Nasty little gotchas hiding everywhere.
Cheers,
Fred
===========================================================================
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".
