From: Glenn Meyer (grmeyer++at++mail.arc.nasa.gov)
Date: 02/26/2004 11:28:43
I haven't addressed this problem on our Onyx4, but have you checked out the
most recent SGI OpenGL Performer manual, Chapter 17? Here's an excerpt:
> Multiple windows, particularly those on separate
> graphics pipelines, that are intended to produce
> results that can be seen as a single image, such
> as projected side by side on a large screen or to
> video outputs used for a stereo display, must have
> their video vertical retraces synchronized with
> genlock(7) and their double buffering
> synchronized. This is necessary for both image
> quality and performance reasons as the last window
> to finish operation can hold up all of the
> rendering processes. Window double-buffering
> synchronization can be done through pfChannel
> share groups with the PFCHAN_SWAPBUFFERS_HW token,
> as discussed in Chapter 2, "Setting Up the Display
> Environment", or explicitly by attaching the
> windows to form window swap groups with
> pfAttachPWinSwapGroup(groupPWin, attachee) as
> discussed in Chapter 16, "Windows". pfPipeWindows
> can have pfWindows in their swap groups if a
> pfPipeWindow is the main group window. The sample
> program in file
> /usr/share/Performer/src/pguide/libpf/C/multipipe.c
> for IRIX and Linux and in file
> %PFROOT%\Src\pguide\libpf\C\multipipe.c for
> Microsoft Windows demonstrates multipipe
> synchronization.
Glenn
Jeroen Akershoek wrote:
> Hello,
>
> On our Onyx4 we're experiencing 'hickups' during Performer applications. This happens when geometry is 'moved' (rotating the scene for example) from 1 pipe to the other. I suppose this is because it needs to transfer the data from 1 pipe/module to the other. Does Performer provide a way to pre-load geometry to all pipes or perhaps some other solution to get rid of the delay?
>
> Regards,
> Jeroen
-- -- XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX Glenn Meyer Principal Engineer office: (650)604-1491 or -3281 QSS Group, Inc. fax: (650)604-3729 NASA Ames Research Center (MS 262-4) grmeyer++at++mail.arc.nasa.gov Moffett Field, CA 94035-1000 XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
This archive was generated by hypermail 2b29 : Thu Feb 26 2004 - 11:31:09 PST