Hi Michele,
On Wed, Jan 18, 2006 at 11:29:38AM +0100, Michele Puccini wrote:
> >> RFE: It would be nice to have a .show(boolean VSynch) in java2d. The
> >> main
> >> problem is the CPU load: the caller waits until it gets the vsynch. The
> >> implementation could setup an event and be signalled when the vsynch
> >> arrives.
> >
> > Well, the users don't complain about this for the full-screen
> > BufferStrategy (aka FlipBufferStrategy).
>
> FlipBufferStrategy is already VSynched. BlitBufferStrategy is not. That's
> the case I would like to address.
I realize that. What I'm saying is that since a Vsynched windowed
buffer strategy would be implemented in the same way the
FlipBufferStrategy is, there won't be a problem with the cpu usage.
> > I don't think asynchronous vsync event is a good idea, though, as it
> > typically turns out to be too much hassle and potential
> > multithreading issues.
>
> Yes. I try to explain better (oofff). Using a brute-force approach {while
> !vblank();} leads to a high CPU load. This could be avoided by "sleeping"
> the waiting thread for a moment (some ms for each loop). The dshow
> implementation can signal an event. This means that we could be simply
> wait() until signalled.
I see. We're talking about the same thing. Of course I wouldn't
recommend spinning in a loop waiting for v-sync, but I thought
you're suggesting to expose the way we're waiting for the vsync on
the api level.
Thanks,
Dmtiri
>
> There's a nice article here -> http://www.compuphase.com/vretrace.htm
>
>
> > Hmm. I don't see why it wouldn't work.
> >
> > Do you use BufferStrategy with the same canvas? Could your
> > rendering be overwritten by the contents of the back-buffer?
> >
> > Could you try running your app with ddraw disabled
> > (-Dsun.java2d.noddraw=true)?
> >
> > Also, do you add your canvas before or after entering the
> > full-screen mode? We do some funny stuff under the covers when
> > entering the full-screen mode (especially if you're using Window
> > instead of Frame to enter the FS mode).
>
>
> At the moment I'm not going to spent too much time here as it would be a
> very implementation-specific topic.
> Even more the fullscreen exclusive mode is not suited for the kind of app
> we're going to produce (the "auto-iconify" is too much dangerous in the
> case you count on a 24h/24h full screen output).
>
> Anyway I'll take some time to investigate further
> with -Dsun.java2d.noddraw=true and let you know.
>
> Cheers!
>
> Mik
> ============================================================================
> >ClassX Development Italy Via Francesca, 368/I I-56030 S.M. a Monte (PI) <
> >Tel.(+39)-0587-705153 Fax.(+39)-0587-705153 WEB: http://www.classx.it <
> ============================================================================
>
===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA2D-INTEREST". For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".
