Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!husc6!cmcl2!brl-adm!adm!bzs@bu-cs.bu.edu
From: bzs@bu-cs.bu.edu (Barry Shein)
Newsgroups: comp.unix.wizards
Subject: BSD4.2 paging heuristics and vadvise(VA_ANOM)
Message-ID: <8673@brl-adm.ARPA>
Date: Tue, 4-Aug-87 19:49:53 EDT
Article-I.D.: brl-adm.8673
Posted: Tue Aug  4 19:49:53 1987
Date-Received: Thu, 6-Aug-87 06:43:57 EDT
Sender: news@brl-adm.ARPA
Lines: 23


Something you can do for free (maybe you know this) is to do a 'ps
agx' and kill off any daemons et al that you don't really need. Things
like rwhod run every 30 seconds are so and on a memory limited machine
will just swap in and out and try to hang onto a working set. Also
those pretty digital clocks, they have to get into memory to update
the hands, at least slow them down if there's an option (I know on X's
clock you can regulate update time from the startup command line,
consider every 5 minutes or get a watch and don't run it at all.)
Consider whether you need a sendmail daemon running (I don't have mail
delivered to my workstation, only the server, not particularly for
this reason but it doesn't hurt.)

Also watch out for lots of windows and possibly rlogins which might be
getting woken up in the background. The best thing to do is start up
your lisp on the barest possible system, look at vmstat and how it
feels and proceed from there. It might be something small which is the
culprit like rwhod (or, of course, you might just be out of memory,
consider a 3/60 w/ 24MB + 3MIPs, that would help :-)

Cheers.

	-Barry Shein, Boston University