iR Glitch Revisited...

New Message Reply Date view Thread view Subject view Author view

Jan Barglowski (jan++at++archimedes.vislab.navy.mil)
Thu, 30 Oct 1997 08:50:01 -0800 (PST)


Hello!

I'm still trying to fix a drawing "glitch" that happens about every 2
seconds. Here's what I've done so try and narrow it down:
- Everything in the FAQ: no desktop, turned off unecessary chkconfigs
  (including nfs), NOINTR in irix.sm, no extra graphics processes
  (xclock, xload, gr_osview, etc...), no remote logins.
- I try to lock down the cpus and prioritize the processes to realtime
  (+35), but to no avail. Perfly crashes when I use the -L -N options
  as root (but either by itself doesn't crash) I can run perfly with
  -L, then raise the priority via npri, but the glitch still happens
  (and the machine locks when closing perfly) This is probably another
  problem in itself :-(

An output of top while running perfly (no prioritize procs) looks like:

---
IRIX64 cyclops 6.2 03131016 IP25 Load[0.42,0.22,0.08] 08:16:04  102 procs
    user   pid  pgrp   %cpu proc  pri  size   rss    time  command        
    root  1001   996  20.10    *   79 17741   798    0:04  perfly_ogl
    root  1003   996  10.32    *   61 17609   659    0:02  perfly_ogl
    root   999   996   9.49    *   66 17609   656    0:02  perfly_ogl
    root   996   996   9.10    *   73 17594   165    0:02  perfly_ogl
    root  1000   996   8.60    *   60 17065   129    0:01  perfly_ogl
    root   998   996   5.84    *   67 17065   127    0:01  perfly_ogl
    root  1002   996   4.99    *   60 17065   128    0:00  perfly_ogl
    root  1004   996   2.92    *   60 17065   108    0:00  perfly_ogl
     jan   984   984   2.85    1   61   119    80    0:00  top
    root    22     0   2.56    *  +39     0     0    0:20  sockd
    root  1005   996   0.62    *   60 17065   109    0:00  perfly_ogl
    root   770   770   0.32    *   60   865   540    0:04  Xsgi
     jan   924   924   0.13    *   60   264   157    0:00  xterm
    root   398     0   0.05    *   39    98    54    0:00  sysctlrd
    root   997   996   0.02    *   39 16987    57    0:00  perfly_ogl
    root   260     0   0.00    *   61     0     0    0:00  nfsd
---

My guess is some other system process is taking a hit on the resources every once in a while. Suspicious processes are sockd and rtnetd, of which I can find little info and have had no luck downgrading priority.

pfStats show the procs easily maintaining 60Hz, but the glitch "reorganizes" the staggered proc display every couple of seconds. The "Misses" row shows on average 2 misses under "total", but nothing under any of the processes. Running APP_CULL_DRAW, Free Run, video at 60Hz.

More stuff: iR, 16 cpus, IRIX 6.2, Performer 2.1

Someone mentioned a program to monitor processes that might be interfering, but I lost my email (long story) and the Performer archives don't have the September messages yet.

Thanks for any suggestions!

jan

-- 
Jan Anthony Barglowski, C723100D
927-1057 jan++at++chinalake.navy.mil
=======================================================================
List Archives, FAQ, FTP:  http://www.sgi.com/Technology/Performer/
            Submissions:  info-performer++at++sgi.com
        Admin. requests:  info-performer-request++at++sgi.com

New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:56:09 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.