From: Ong Tze Lin (tzelin++at++sgisgp.singapore.sgi.com)
Date: 04/04/2001 09:39:01
Hi Lee,
IMHO lack of genlock will not affect your graphics
performance per se, because this is just a hardware
signal to tell your Display Generator cards: "please
start your video flyback and rescan now."
pfSync of course waits for the video retrace signal, but
genlock or no, that signal will happen at a fixed rate,
regardless. The only question is whether all three (or
more) pipes will have it happen simultaneously.
This translates into a visual quality issue- uneven
redraws between screens, very visible esp on large
projections- rather than a frames/sec performance hit.
If you are using stress, it just measures actual
over expected rendering time so I don't expect lack
of genlock to affect it unless- maybe, just maybe you're
on the hairiest of hairy edges of performance dropoffs.
This is because I think stress doesn't include the
time to sync, only the time to actually render
to the framebuffer... gotta get a confirmation on this
one. In any case sync's contribution is probably
small enough that tweaking the high and low stress levels
could handle it.
** I will defer to anyone who's had experiences/thoughts
contrary to this though... I can't think of any reason lack
of genlock will cause a performance problem.
Gangswap/swapready, on the other hand, will wait for
all pipes to be ready to swap simultaneously, so any
delay in any of the gangdrawn pipes will cause a
delay in all the pipes. So yes, this is potentially
very restrictive- leave it out for now.
I suggest you try and find bottlenecks in other stages
of the Performer and hardware pipes.
Good luck!
Cheers,
Tze Lin
--
Ong Tze Lin Tel/VNET: +65 771 0219
Principal Consultant Mobile : +65 9832 7125
SGI Asia Pacific Fax : +65 779 3650
89 Science Park Dr, #03-03/05, Singapore 118261
On Mon, 2 Apr 2001, Lee Marden wrote:
> Tze Lin,
>
> I'm having graphics stress related issues with my application on this
> 3 pipe (6 channel) system. My understanding is that if the
> pipes/channels are not synchronized correctly, then delays can be
> introduced in the rendering process. Since both perfly and my app issue
> this "No Genlock" warning, I thought that maybe my throughput issues
> might be related. I know my application is not setup, nor designed for,
> gangdraw. I've been told that I don't need gangdraw to optimally render
> on this hardware. However, I would like to get rid of this warning if
> possible. So... I'd like to know if the "No Genlock" message is truly
> genlock (i.e. video display) related, or if it's something more
> substantial, and the graphics pipes are out of alignment, and it's
> effecting my performance (cull/draw times, etc.).
>
> Thanks for your help,
> Lee
>
> Ong Tze Lin wrote:
> >
> > Hi Lee,
> >
> > Yup, Genlock locks the pipe to ensure they all
> > generate video frames at the same time. Yes, this can
> > also be used to ensure sync with the display device.
> >
> > Gangdraw OTOH makes sure they all swap buffers at
> > exactly the same time. This feature requires
> > you to cable up the Swap Ready ports on your
> > graphics hardware.
> >
> > So- for a fully synchronised hardware setup, you'd
> > connect up both genlock and swap ready.
> >
> > I'm not sure where exactly this check is performed,
> > although its bound to be at a very low level.
> >
> > Is there any particular reason you're looking for
> > this message? If you're looking just to disable it,
> > then use PFNFYLEVEL.. it will always appear when
> > multiple unsynchronized pipes are detected...
> >
> > Cheers,
> > Tze Lin
> >
> > --
> > Ong Tze Lin Tel/VNET: +65 771 0219
> > Principal Consultant Mobile : +65 9832 7125
> > SGI Asia Pacific Fax : +65 779 3650
> > 89 Science Park Dr, #03-03/05, Singapore 118261
> >
> > On Thu, 29 Mar 2001, Lee Marden wrote:
> >
> > > Thanks. When you mean hardware genlock between the pipes, are talking
> > > gangdraw between the pipes? My understanding is that gangdraw is
> > > responsible for synching between pipes, while genlock is for synching to
> > > an external video, such as the projectors. I believe this message is
> > > coming from Performer, but I don't know what sub-system, or library.
> > > Would you know the name of the part of Performer that's checking for
> > > this?
> > >
> > > Thanks,
> > > Lee
> > >
> > > Ong Tze Lin wrote:
> > > >
> > > > Hi Lee,
> > > >
> > > > Just means you don't have hardware genlock between
> > > > your pipes. Your channels are still sync'ed
> > > > internally, which may or may not be adequate for
> > > > your simulation needs.
> > > >
> > > > Cheers,
> > > > Tze Lin
> > > >
> > > > --
> > > > Ong Tze Lin Tel/VNET: +65 771 0219
> > > > Principal Consultant Mobile : +65 9832 7125
> > > > SGI Asia Pacific Fax : +65 779 3650
> > > > 89 Science Park Dr, #03-03/05, Singapore 118261
> > > >
> > > > On Wed, 28 Mar 2001, Lee Marden wrote:
> > > >
> > > > > pfUsers,
> > > > >
> > > > > I'm getting a warning when I run perfly, that indicates that there is
> > > > > a problem with genlock across my three pipes. I use the following to
> > > > > setup the three MCOs (2 channels each):
> > > > >
> > > > > ---
> > > > > set DISPLAY=:0.0
> > > > > /usr/gfx/setmon -x -S -p 0 2++at++1025x768_60
> > > > > set DISPLAY=:0.1
> > > > > /usr/gfx/setmon -g -x -S -p 1 2++at++1025x768_60
> > > > > set DISPLAY=:0.2
> > > > > /usr/gfx/setmon -g -x -S -p 2 2++at++1025x768_60
> > > > >
> > > > > /usr/gfx/stopgfx; /usr/gfx/startgfx &
> > > > > ---
> > > > >
> > > > > The scene looks ok, so I know that the MCOs are working, but I get the
> > > > > following warnings:
> > > > > 17724 PF Warning/Resource: !!!!!!!!!!!!!!!!!! NO GENLOCK
> > > > > !!!!!!!!!!!!!!!!!!!!!!
> > > > > 17724 PF Warning/Resource: Screen 1 is phase-shifted 68.657346
> > > > > degrees from master screen 0.
> > > > > 17724 PF Warning/Resource: !!!!!!!!!!!!!!!!!! NO GENLOCK
> > > > > !!!!!!!!!!!!!!!!!!!!!!
> > > > > 17724 PF Warning/Resource: !!!!!!!!!!!!!!!!!! NO GENLOCK
> > > > > !!!!!!!!!!!!!!!!!!!!!!
> > > > > 17724 PF Warning/Resource: Screen 2 is phase-shifted 78.312222
> > > > > degrees from master screen 0.
> > > > > 17724 PF Warning/Resource: !!!!!!!!!!!!!!!!!! NO GENLOCK
> > > > > !!!!!!!!!!!!!!!!!!!!!!
> > > > >
> > > > > I've looked through the perfly code and can't find this message. Is
> > > > > this a Performer message? Any ideas what might be causing this? Should
> > > > > I be concerned?
> > > > >
> > > > > Thanks,
> > > > > Lee
> > > > > --
> > > > > ------------------------------------------------------------------------
> > > > > Lee Marden Raytheon Company
> > > > > Principal Software Engineer Northeast C3I Systems
> > > > > Advanced Modeling & Simulation 50 Apple Hill Dr. (T3MJ21)
> > > > > Technology Section Tewksbury, MA 01876
> > > > > Phone: (978) 858-9487 Fax: (978) 858-4336
> > > > > Email:lee_marden++at++raytheon.com Intranet: http://amasts.msd.ray.com/~lms
> > > > > ------------------------------------------------------------------------
> > > > > -----------------------------------------------------------------------
> > > > > List Archives, FAQ, FTP: http://www.sgi.com/software/performer/
> > > > > Open Development Project: http://oss.sgi.com/projects/performer/
> > > > > Submissions: info-performer++at++sgi.com
> > > > > Admin. requests: info-performer-request++at++sgi.com
> > > > >
> > >
> > > --
> > > ------------------------------------------------------------------------
> > > Lee Marden Raytheon Company
> > > Principal Software Engineer Northeast C3I Systems
> > > Advanced Modeling & Simulation 50 Apple Hill Dr. (T3MJ21)
> > > Technology Section Tewksbury, MA 01876
> > > Phone: (978) 858-9487 Fax: (978) 858-4336
> > > Email:lee_marden++at++raytheon.com Intranet: http://amasts.msd.ray.com/~lms
> > > ------------------------------------------------------------------------
> > >
> >
> > -----------------------------------------------------------------------
> > List Archives, FAQ, FTP: http://www.sgi.com/software/performer/
> > Open Development Project: http://oss.sgi.com/projects/performer/
> > Submissions: info-performer++at++sgi.com
> > Admin. requests: info-performer-request++at++sgi.com
>
> --
> ------------------------------------------------------------------------
> Lee Marden Raytheon Company
> Principal Software Engineer Northeast C3I Systems
> Advanced Modeling & Simulation 50 Apple Hill Dr. (T3MJ21)
> Technology Section Tewksbury, MA 01876
> Phone: (978) 858-9487 Fax: (978) 858-4336
> Email:lee_marden++at++raytheon.com Intranet: http://amasts.msd.ray.com/~lms
> ------------------------------------------------------------------------
> -----------------------------------------------------------------------
> List Archives, FAQ, FTP: http://www.sgi.com/software/performer/
> Open Development Project: http://oss.sgi.com/projects/performer/
> Submissions: info-performer++at++sgi.com
> Admin. requests: info-performer-request++at++sgi.com
>
This archive was generated by hypermail 2b29 : Tue Apr 03 2001 - 19:02:36 PDT