Path: utzoo!utgpu!news-server.csri.toronto.edu!rutgers!bellcore!uunet!zaphod.mps.ohio-state.edu!caen!ox.com!yale!cs.yale.edu!news
From: dan@doctor.chem.yale.edu (Dan Severance)
Newsgroups: comp.sys.sgi
Subject: Re: Graphics slows down
Message-ID: <1991Jun10.233423.4299@cs.yale.edu>
Date: 10 Jun 91 23:34:23 GMT
Sender: news@cs.yale.edu (Usenet News)
Distribution: usa
Organization: Laboratory for Computational Chemistry, Yale University
Lines: 28
Nntp-Posting-Host: doctor.chem.yale.edu


>   Hi,
>..
>    It does an animation of molecular movement over time, but what I see
>is that it starts out running at N frames/second, and then after a couple
>minutes it starts to run about half the speed or slower.  If I tell the 
>program to rewind the file and do it again, it keeps running at this slower
>speed.  Then if I stop and restart the program it runs at the original faster
>speed once again, once more slowing down in a couple of minutes...
>    One thing I've noticed is that doing an osview shows that initially
>the graphics fifowait (what is this??) starts small and gradually gets
>larger; once it is large (14-15M) the program slows down.  Telling the
>program to stop animating causes the fifowait to disappear, but continuing
>it causes an immediate jump back up to 15M (not the gradual climb seen
>at the initial run of the program).

    Hi again.  Well, I had done something wrong, but I still don't understand
the behavior.  I was inadvertantly drawing 348 zero length vectors 
using v3f having two endpoints (bgnlin; 2 call v3fs; endlin).  Why would
this cause the slowing down over time?  
    I'm just curious now since I found the problem, but I'd still like
to know if this is the expected behavior or if it is a generally 
innocuous bug of some sort....
    Thanks,

      Dan Severance
      dan@omega.chem.yale.edu
      Laboratory of Computational Chemistry