From chemistry-request@ccl.net Mon Mar 8 04:03:46 2004 Received: from shiva.jussieu.fr (shiva.jussieu.fr [134.157.0.129]) by server.ccl.net (8.12.8/8.12.8) with ESMTP id i2893jjs008019 for ; Mon, 8 Mar 2004 04:03:46 -0500 Received: from ds10.itodys.jussieu.fr (ds10.itodys.jussieu.fr [134.157.24.13]) by shiva.jussieu.fr (8.12.10/jtpda-5.4) with ESMTP id i2895akE065576 for ; Mon, 8 Mar 2004 10:05:36 +0100 (CET) X-Ids: 168 Received: by ds10.itodys.jussieu.fr (8.12.1/1.1.2.11/09Jul02-1200PM) id i2895YcC450554; Mon, 8 Mar 2004 10:05:34 +0100 (MET) Date: Mon, 8 Mar 2004 10:05:34 +0100 (MET) From: Michel Petitjean Message-Id: <200403080905.i2895YcC450554-.at.-ds10.itodys.jussieu.fr> To: chemistry-.at.-ccl.net Subject: CCL:Re: software evaluation / validation X-Miltered: at shiva.jussieu.fr with ID 404C3760.000 by Joe's j-chkmail (http://j-chkmail.ensmp.fr)! X-Antivirus: scanned by sophie at shiva.jussieu.fr X-Spam-Status: No, hits=0.0 required=7.0 tests=none autolearn=no version=2.61 X-Spam-Checker-Version: SpamAssassin 2.61 (1.212.2.1-2003-12-09-exp) on servernd.ccl.net To: chemistry-.at.-ccl.net Subject: CCL:Re: software evaluation / validation Andras.Borosy-.at.-givaudan.com wrote: >... > How is it possible to reproduce scientific results obtained by commercial > _OR_ academic software which relies on _RANDOM_ methods ? >... Use a random generator returning the same sequence on all platforms. For successive runs on the same platform, a detailed description of the algorithm should give access to the number of calls to the RG. If you need different results on a given platform, just run several times the calculation within a single programme call, when possible. E.g. the ASV freeware contains an optional Monte-Carlo part: http://petitjeanmichel.free.fr/itoweb.petitjean.freeware.html#ASV (van der Waals surfaces and volumes; also calculated analytically) The Monte-Carlo results are the same on all platforms, provided that all input parameters are identical (includes the number of runs and the numbers of random observations) Performing several Monte-Carlo within a single call is possible. The same remarks apply to CSR (3D alignments): http://petitjeanmichel.free.fr/itoweb.petitjean.freeware.html#CSR Commercial softwares should be able to do so. If not, use adequate academic softwares, or built yourself the programmes. Michel Petitjean, Email: petitjean-.at.-itodys.jussieu.fr ITODYS (CNRS, UMR 7086) ptitjean-.at.-ccr.jussieu.fr http://petitjeanmichel.free.fr/itoweb.petitjean.html From chemistry-request@ccl.net Mon Mar 8 13:59:54 2004 Received: from sycorax.delfa.net (relay4.delfa.net [193.125.210.9]) by server.ccl.net (8.12.8/8.12.8) with ESMTP id i28Ixmjs007270 for ; Mon, 8 Mar 2004 13:59:49 -0500 X-Envelope-To: Received: from dux.ru (dux.ru [193.125.210.65]) by sycorax.delfa.net (8.12.11/8.12.11/Sycorax.Delfa) with ESMTP id i28J1aEw014171 for ; Mon, 8 Mar 2004 22:01:36 +0300 Received: from dux.ru (localhost [127.0.0.1]) by dux.ru (8.12.6/8.12.6) with ESMTP id i28J1aIU042677 for ; Mon, 8 Mar 2004 22:01:36 +0300 (MSK) (envelope-from dima*at*xenon.spb.ru) Received: (from uucp@localhost) by dux.ru (8.12.6/8.12.1/Submit) with UUCP id i28J1a8B042676 for chemistry*at*ccl.net; Mon, 8 Mar 2004 22:01:36 +0300 (MSK) (envelope-from dima*at*xenon.spb.ru) Received: from goblin ([192.168.5.1] helo=xenon.spb.ru ident=dima) by shadow with esmtp (Exim 3.35 #1 (Debian)) id 1B0Py4-0005Wb-00; Mon, 08 Mar 2004 21:59:12 +0300 Message-ID: <404CC27F.6030506*at*xenon.spb.ru> Date: Mon, 08 Mar 2004 21:59:11 +0300 From: Dmitry Rozmanov User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.5) Gecko/20031107 Debian/1.5-3 X-Accept-Language: ru, en MIME-Version: 1.0 To: Shaji Chempath , CCL Subject: Re: CCL:Geom. Optimization in Solvent (Gaussian, Jaguar) References: <40480263.6040705*at*berkeley.edu> In-Reply-To: <40480263.6040705*at*berkeley.edu> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Status: No, hits=0.0 required=7.0 tests=none autolearn=no version=2.61 X-Spam-Checker-Version: SpamAssassin 2.61 (1.212.2.1-2003-12-09-exp) on servernd.ccl.net Hi Shaji, Have the same problem with GAMESS. Cannot find stationary point with PCM water solvation. It stucks at some point with the oscilating energy. Exactly as in your case. Though still have not solved the problem I think that the larger number of point/tesserae is the right way. Have you tried to enlarge the number? With what result? Best wishes. ---Dmitry. Shaji Chempath wrote: > > Hello CCLers, > > Few weeks beore I had asked a question about geometry optimization in > the presence of a solvent. I would like to thank all the people who > replied to that. ( see summary of replies below). > However I still have lots of problems with optimizing in the presence of > solvent. I have used both Gaussian and Jaguar for doing my polarizable > continuum optimizations and still no luck. > Here is what I understand about what is going on, please let me know if > you have any opinions. It appears as though the gradients calculated in > the presence of solvent are far less accurate. This may be because the > 'Poisson Boltzmann' equation is solved numerically for the solvent > region and the reaction field in the solvent is mapped on to limited > number of "Tesserae" (points on the solute-solvent boundary). This > somehow makes the potential energy surface very rough, compared to the > smooth surface in gas phase. Look at the graphs (on my webpage) to see > what I mean > > http://zeolites.cqe.nwu.edu/shaji/solvation.html > > You can see that the gas phase optimization proceeds smoothly, but > liquid phase optimizations run into bumpy potnetial energy surface. > (even after staring from a gas phase optimized geometry). This is a > general phenomena (the roughening of the potential energy surface) I > have observed for different systems I have tried . > > So based on this I hav ecome to the conclusion that the optimization > with solvent is very tough and works only for very simple molecules. Am > I wrong? > > Summary of help I got from others regarding PCM (Gaussian) optimizations > ------------------------------------------------------------- > > a) initially do a gas phase geometry and then start liquid phase > optimization from there > > b) use "OPT(calcfc)" to calculate initial hessian > > c) try iefpcm instead of pcm or scipcm > > d) try "loose" optimization > > e) try changing TSNUM (number of tesserae) or TSARE(ares of tesserae) in > scrf > > f) use nosymm option > > Anyway none of them seem to have helped in particular for my molecules > Thanks again > > Shaji > > > > > > -= This is automatically added to each message by the mailing script =- > To send e-mail to subscribers of CCL put the string CCL: on your > Subject: line > and send your message to: CHEMISTRY*at*ccl.net > > Send your subscription/unsubscription requests to: > CHEMISTRY-REQUEST*at*ccl.net HOME Page: http://www.ccl.net | Jobs Page: > http://www.ccl.net/jobs > If your mail is bouncing from CCL.NET domain send it to the maintainer: > Jan Labanowski, jlabanow*at*nd.edu (read about it on CCL Home Page) > -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ > > > > > >