Jim Galis (jgalis++at++paradigmsim.com)
Thu, 11 Sep 1997 09:43:31 -0500
If you have looked at this (above) and that isnt it, there is a sure fire
method to find it: WindView. this will analyze (in realtime) your processes
on each CPU, and you can instrument your code to find the problem areas.
It is very easy to use, and non-intrusive. Even without the instrumentation,
WindView exposes all tasks running. You can get this software for eval,
and try it out. You must be running rtmond for it to work.
>Thanks to all who've replied, and sorry I haven't summarized sooner!
>
>I still have the ~2 second "glitch", and here's what I've done so far:
>
>- Stopped displaying my xclock (thanks Angus)
>
>- Added the NOINTR directive in my irix.sm (thanks Dr. Colin) as shown
> in the FAQ. I now have protected processors 9, 10, and 11 from
> broadcast interrupts and lock them in my app.
>
>- I run a "desktop" with three "desks", some with nice picture backgrounds.
> Haven't had the chance to disable this, but the desk I run the app on
> has no background defined. Also, the framebuffer is 2560x1024 across
> 2 monitors so I can debug if necessary. (printf's off when testing
> framerate, however). Haven't compare it to a 1280x1024 framebuffer
> yet.
>
>Most curious is this sockd, which seems to be running with a rather
>high priority and takes up quite of bit of CPU time, too. I got a
>little info on sockd, which is a firewall proxy daemon? The NFS
>mounts are few and checked to be inactive during testing. Maybe
>bdflush?
>
>I do run an intersect process, but the stats show "0" time spent
>(and I've disabled the intersect and same prob stays)
>
>Once again, iR, 12CPU, 2-64Mb RM6, 2Gb RAM, IRIX 6.2, Performer 2.1
>
>OK, Round 2 begin!
===========================================
Jim Galis
Paradigm Entertainment, Inc.
jgalis++at++paradigmsim.com
---- There are no accidents ----
===========================================
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
Submissions: info-performer++at++sgi.com
Admin. requests: info-performer-request++at++sgi.com
This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:55:54 PDT