From owner-chemistry@ccl.net Thu Mar  6 07:49:00 2008
From: "Kalju Kahn kalju=chem.ucsb.edu" <owner-chemistry:-:server.ccl.net>
To: CCL
Subject: CCL: Coupled Cluster and MP4 in peptides
Message-Id: <-36432-080306071021-22801-gTFjY/jBE/pKXL9RFB0YNA:-:server.ccl.net>
X-Original-From: "Kalju Kahn" <kalju**chem.ucsb.edu>
Content-Transfer-Encoding: 8bit
Content-Type: text/plain;charset=iso-8859-1
Date: Thu, 6 Mar 2008 04:10:01 -0800 (PST)
MIME-Version: 1.0


Sent to CCL by: "Kalju Kahn" [kalju{:}chem.ucsb.edu]
Hi Pablo,

You may also want to consider that only few programs implement analytical
gradients for methods fancier than MP2.  The numerical optimization of
large flexible systems can be *very* time consuming.

As for single point jobs, get as much memory as you can ... the speed of
the transformation step in such large systems scales favorably with
available memory as fewer passes are needed.

> In my group, at the University of Zaragoza, we are planning to build a new
> cluster to perform high-level QC calculations. Meaning CCSD, CCSD(T),
> MP4(SDQ) and MP4, with basis sets (Dunning's, for example) of, say,
> double-, triple- and quadruple-zeta quality (if possible).
...
> We would like to calculate single-point energies, gradients and maybe
> second derivatives of the energy of these systems using the aforementioned
> levels of the quantum chemical theory.

Good luck,

Kalju


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Dr. Kalju Kahn
Department of Chemistry and Biochemistry
UC Santa Barbara, CA 93106