From chemistry-request&$at$&www.ccl.net  Wed Feb 24 11:29:59 1999
Received: from lovelace.usfca.edu (root %-% at %-% lovelace.usfca.edu [138.202.192.10])
        by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id LAA20376
        Wed, 24 Feb 1999 11:29:58 -0500 (EST)
Received: from delta.usfca.edu (deltaenet.usfca.edu [138.202.3.254])
	by lovelace.usfca.edu (8.9.1a/8.9.1) with SMTP id IAA04171
	for <chemistry- at -www.ccl.net>; Thu, 25 Feb 1999 08:30:55 -0800
Date: Thu, 25 Feb 1999 08:30:55 -0800
Received: from [138.202.97.100] by delta.usfca.edu with ESMTP
          for chemistry |-at-| www.ccl.net; Wed, 24 Feb 1999 8:29:58 -0800
X-Sender: karney /at\pop.admin.usfca.edu
Message-Id: <l03110700b2f96bcf2afd ^at^ [138.202.97.103]>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
To: CCL <chemistry-: at :-www.ccl.net>
From: William Karney <karney -A_T- usfca.edu>
Subject: memory usage in G94


Hello,
   I'm using Gaussian94 on an SGI Octane.  We recently upgraded to 768MB of
RAM, with the intention of running large jobs (e.g. CASSCF) incore.  But
for some reason, I'm unable to use any more than about 512MB for a single
job.  That is, I can specify up to mem=63000000 in the input file, and it
will run (though, it seems, more slowly than normal), but if I request more
than 63 megawords of memory, the job just hangs up.
   Does anyone have any suggestions as to how to overcome this problem, and
persuade G94 to use more than 63 MW of memory for a job?  Any advice would
be greatly appreciated.

Thanks,
William