From owner-chemistry@ccl.net Wed Oct 3 04:21:01 2007 From: "Jens Thomas j.m.h.thomas()dl.ac.uk" To: CCL Subject: CCL:G: TDDFT electric transition dipole moment plotting Message-Id: <-35292-071002130306-9997-ytYN9rCGH/VvO0KHajJZhA]*[server.ccl.net> X-Original-From: Jens Thomas Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Tue, 02 Oct 2007 18:02:07 +0100 MIME-Version: 1.0 Sent to CCL by: Jens Thomas [j.m.h.thomas!A!dl.ac.uk] Dear Tom, The CCP1GUI has support for plotting vectors in a variety of ways (streamlines, hedgehog plots, glyphs etc), is free and runs on Windows, Mac and Linux. You can find out more about the CCP1GUI at: http://www.cse.scitech.ac.uk/ccg/software/ccp1gui/ The CCP1GUI supports a number of different file formats, but, for various reasons, we haven't been able to provide support for Gaussian log files yet. However, if you can send me some information about the format of Gaussian log files and some examples of your data, I should be able to get the CCP1GUI to do the plots you are interested in. Best wishes, Jens Tom FA de Greef t.f.a.d.greef!A!tue.nl wrote: > Sent to CCL by: "Tom FA de Greef" [t.f.a.d.greef#tue.nl] > Dear all, > Last week I asked if any programs are availabe to plot the ground to excited state transition electric dipole moment from gaussian TDDFT calculations (these vectors appear in the .log file). Due to the lack of response there are maybe no standard programs that can do the trick. However, maybe there are programs available that read the standard coordinates from the .log file and that give an option to plot a vector in terms of x,y,z coordinates. Does anyone know of such a program that can do this (windows or linux). > > best regards, > > tom> > > > > > -- =================================================================== Jens Thomas, email: j.m.h.thomas{=}dl.ac.uk STFC Daresbury Lab, tel: +44-1925-603849 Warrington, fax: +44-1925-603634 WA4 4AD, UK. http: http://www.cse.scitech.ac.uk =================================================================== From owner-chemistry@ccl.net Wed Oct 3 04:53:00 2007 From: "VITORGE Pierre 094605 Pierre.VITORGE__cea.fr" To: CCL Subject: CCL:G: Gaussian parallelism and parallelism in general Message-Id: <-35293-071003041137-20852-joc6O4FfksehdL8CWzcl7w * server.ccl.net> X-Original-From: "VITORGE Pierre 094605" Content-class: urn:content-classes:message Content-Type: multipart/mixed; boundary="----_=_NextPart_001_01C8058D.CCE30027" Date: Wed, 3 Oct 2007 09:17:49 +0200 MIME-Version: 1.0 Sent to CCL by: "VITORGE Pierre 094605" [Pierre.VITORGE-*-cea.fr] This is a multi-part message in MIME format. ------_=_NextPart_001_01C8058D.CCE30027 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable -----Original Message----- > From: Herbert Fruchtl herbert.fruchtl .. st-andrews.ac.uk = [mailto:owner-chemistry%a%ccl.net] Sent: Tue 02/10/2007 18:19 Subject: CCL:G: Gaussian parallelism and parallelism in general =20 ...(I know only two=20 programs with a good, robust general purpose optimizer). Herbert Which ones ? (!) Pierre Vitorge Laboratoire de Speciation des Radionucl=E9ides et Mol=E9cules (LSRM) CEA Saclay DEN/DPC/SECR UMR 8587 et Laboratoire Analyse et Mod=E9lisation pour la Biologie et = l'Environnement (LAMBE) Universit=E9 d'Evry UMR 8587 (CEA - Universit=E9 d'Evry-Val d'Essone - = CNRS) http://www.vitorge.name ------_=_NextPart_001_01C8058D.CCE30027-- From owner-chemistry@ccl.net Wed Oct 3 05:46:01 2007 From: "Andrew Turner andrew.turner###ed.ac.uk" To: CCL Subject: CCL: Gaussian parallelism and parallelism in general Message-Id: <-35294-071003042516-22930-QxTy9zUxWtebc9MxKc3/wA{}server.ccl.net> X-Original-From: Andrew Turner Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Wed, 03 Oct 2007 09:25:08 +0100 MIME-Version: 1.0 Sent to CCL by: Andrew Turner [andrew.turner_+_ed.ac.uk] > (I know only two > programs with a good, robust general purpose optimizer). > And they are? Or are you forbidden from telling us due to contractual obligations? As with most things it is horses-for-courses. There are many factors to weigh up when choosing which code to use. Often the best choice is the one that you have access to local expertise in. That way, you can easily get help when you need it. Your choice, of course, is also determined by what methods/techniques you want to use. All other things being equal, I will always choose a free/open source option over a commercial code. Andy -- ================================== Dr Andrew R. Turner Research Computing Officer e: andrew.turner*ed.ac.uk t: +44 (0)131 650 7748 w: http://www.eastchem.ac.uk/rcf icq: 370-899-715 p: School of Chemistry University of Edinburgh EH9 3JJ ================================== From owner-chemistry@ccl.net Wed Oct 3 06:38:00 2007 From: "Herbert Fruchtl herbert.fruchtl(~)st-andrews.ac.uk" To: CCL Subject: CCL:G: Gaussian parallelism and parallelism in general Message-Id: <-35295-071003063215-28767-7fC7NX8NTvcbdFZAU2qSAA*|*server.ccl.net> X-Original-From: Herbert Fruchtl Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Wed, 03 Oct 2007 11:30:50 +0100 MIME-Version: 1.0 Sent to CCL by: Herbert Fruchtl [herbert.fruchtl.:.st-andrews.ac.uk] VITORGE Pierre 094605 Pierre.VITORGE__cea.fr wrote: >> From: Herbert Fruchtl herbert.fruchtl .. st-andrews.ac.uk > ...(I know only two > programs with a good, robust general purpose optimizer). > > Which ones ? (!) Gaussian and NWChem (if one fails, I use the other, even though NWChem is painfully slow for normal-sized calculations on normal-sized computers). If optimization is not the main problem, there are better programs than these. Before I'm beaten up again: I have never used Turbomole, but heard good things about it. But I have heard good things about many programs, which then didn't always live up to it. In most programs, geometry optimization is treated as a stepchild. Somebody implemented BFGS long ago (if you're lucky), and nobody will touch this code ever again. I have seen a very expensive program fail on water dimer. As soon as you have some stiff and some soft modes (in other words: always :-), you're in trouble. Herbert > > Pierre Vitorge > > Laboratoire de Speciation des Radionucléides et Molécules (LSRM) > CEA Saclay DEN/DPC/SECR UMR 8587 > et > Laboratoire Analyse et Modélisation pour la Biologie et l'Environnement (LAMBE) > Université d'Evry UMR 8587 (CEA - Université d'Evry-Val d'Essone - CNRS) > http://www.vitorge.name -- Herbert Fruchtl EaStCHEM Fellow School of Chemistry University of St Andrews From owner-chemistry@ccl.net Wed Oct 3 07:12:01 2007 From: "Herbert Fruchtl herbert.fruchtl-.-st-andrews.ac.uk" To: CCL Subject: CCL: Gaussian parallelism and parallelism in general Message-Id: <-35296-071003063849-30731-LdcApNf2AbUHcXxwlXlLOA[]server.ccl.net> X-Original-From: Herbert Fruchtl Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Wed, 03 Oct 2007 11:34:45 +0100 MIME-Version: 1.0 Sent to CCL by: Herbert Fruchtl [herbert.fruchtl-$-st-andrews.ac.uk] > >> (I know only two programs with a good, robust general purpose optimizer). >> > And they are? Or are you forbidden from telling us due to contractual > obligations? > I have risen to the bait... It's not contractual obligations, but the inevitable "my program is better than yours" and "you didn't choose the right options" flame war that will inevitably follow now. Herbert From owner-chemistry@ccl.net Wed Oct 3 07:55:00 2007 From: "Utpal Sarkar utpalchemiitkgp*gmail.com" To: CCL Subject: CCL: parameters in an empirical Parameters in an emperical potential Message-Id: <-35297-071003070854-12486-uIdCwhjue3SaaKUi8ai+qw::server.ccl.net> X-Original-From: "Utpal Sarkar" Content-Type: multipart/alternative; boundary="----=_Part_3933_12551268.1191406044207" Date: Wed, 3 Oct 2007 12:07:24 +0200 MIME-Version: 1.0 Sent to CCL by: "Utpal Sarkar" [utpalchemiitkgp{=}gmail.com] ------=_Part_3933_12551268.1191406044207 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline Dear CCL Users, We are interested in optimizing the parameters in an empirical potential to give a better fit to DFT calculations (energies, bond lengths, etc.). Is anybody aware of any software, within DL_POLY or some other MD code, that enables us to do this? Any kind of help will be greatly appreciated. Thanking you in advance With best regards Utpal Sarkar ------=_Part_3933_12551268.1191406044207 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline Dear CCL Users,
                       We are interested in optimizing the parameters in an empirical 
potential to give a better fit to DFT calculations (energies, bond 
lengths, etc.).

Is anybody aware of any software, within DL_POLY or some other 
MD code, that enables us to do this?

Any kind of help will be greatly appreciated.

Thanking you in advance
With best regards
Utpal Sarkar
 ------=_Part_3933_12551268.1191406044207-- From owner-chemistry@ccl.net Wed Oct 3 09:53:01 2007 From: "Marcel Swart marcel.swart ~ icrea.es" To: CCL Subject: CCL:G: Geometry optimization techniques Message-Id: <-35298-071003090437-9801-LaJ3Ien0VqEj/dqtbKbz8g[]server.ccl.net> X-Original-From: Marcel Swart Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=ISO-8859-1; delsp=yes; format=flowed Date: Wed, 3 Oct 2007 13:29:11 +0200 Mime-Version: 1.0 (Apple Message framework v752.3) Sent to CCL by: Marcel Swart [marcel.swart++icrea.es] I would like to make a small point here: there are some programs that can do geometry optimization very well, and which are not necessarily tied to any QM program. Two programs that come to mind are ChemShell, and the Optimize code by Baker and co-workers. Furthermore, for the combination of strong and weak coordinates (or =20 as stated here: stiff and soft modes), you might want to have a look at the =20 adapted delocalized coordinates: M. Swart and F.M. Bickelhaupt "Optimization of strong and weak coordinates" Int. J. Quant. Chem. 2006, 106, 2536-2544 http://dx.doi.org/10.1002/qua.21049 This was already implemented in the QUILD program: M. Swart and F.M. Bickelhaupt "QUILD: QUantum-regions Interconnected by Local Descriptions" J. Comput. Chem. 2007, in press http://dx.doi.org/10.1002/jcc.20834 and is now also available with the latest version of ADF. On Oct 3, 2007, at 12:30 PM, Herbert Fruchtl herbert.fruchtl(~)st-=20 andrews.ac.uk wrote: > Sent to CCL by: Herbert Fruchtl [herbert.fruchtl.:.st-andrews.ac.uk] > VITORGE Pierre 094605 Pierre.VITORGE__cea.fr wrote: > >> From: Herbert Fruchtl herbert.fruchtl .. st-andrews.ac.uk > > ...(I know only two > > programs with a good, robust general purpose optimizer). > > > > Which ones ? (!) > > Gaussian and NWChem (if one fails, I use the other, even though =20 > NWChem is painfully slow for normal-sized calculations on normal-=20 > sized computers). If optimization is not the main problem, there =20 > are better programs than these. > > Before I'm beaten up again: I have never used Turbomole, but heard =20 > good things about it. But I have heard good things about many =20 > programs, which then didn't always live up to it. > > In most programs, geometry optimization is treated as a stepchild. =20 > Somebody implemented BFGS long ago (if you're lucky), and nobody =20 > will touch this code ever again. I have seen a very expensive =20 > program fail on water dimer. As soon as you have some stiff and =20 > some soft modes (in other words: always :-), you're in trouble. =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D dr. Marcel Swart ICREA researcher at Institut de Qu=EDmica Computacional Universitat de Girona Campus Montilivi 17071 Girona Catalunya (Spain) fax +34-972-418356 e-mail marcel.swart/a\icrea.es marcel.swart/a\udg.edu web http://www.icrea.es/pag.asp?id=3DMarcel.Swart http://iqc.udg.edu/~marcel =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D From owner-chemistry@ccl.net Wed Oct 3 13:10:00 2007 From: "Ol Ga eurisco1---pochta.ru" To: CCL Subject: CCL:G: DTol, SCF convergence and Mg n, where n ~ 50-200 Message-Id: <-35299-071003130648-6342-6On9GW8xymOeXjwl1Ej80Q%a%server.ccl.net> X-Original-From: "Ol Ga" Date: Wed, 3 Oct 2007 13:06:45 -0400 Sent to CCL by: "Ol Ga" [eurisco1]~[pochta.ru] Hello! :-) In the article Yan Zhao and Donald G. Truhlar "Benchmark Databases for Nonbonded Interactions and Their Use to Test Density Functional Theory" J. Chem. Theory Comput it is underlined an opportunity of application the density tolerance variable DTol in subroutine bc95ss.f on G03. Whether is there an opportunity of application for calculation DFT (PBEPBE, PBE1PBE) for systems Mg [n], where n ~ 50-200. If it is known, that SCF does not converge at practically any choice of Functional in such case. Whether is it possible to manage and this utility, but to make the manipulations with the same results. Thank you very much in advance. Sincerely Yours, Ol GA From owner-chemistry@ccl.net Wed Oct 3 17:06:01 2007 From: "Soren Eustis soren%jhu.edu" To: CCL Subject: CCL: Output from Cubeman and probability density Message-Id: <-35300-071003170339-23639-5F98/mJlgqFMGxf349dedw#server.ccl.net> X-Original-From: "Soren Eustis" Date: Wed, 3 Oct 2007 17:03:35 -0400 Sent to CCL by: "Soren Eustis" [soren . jhu.edu] I am trying to calculate the probability density for a given diffuse singly occupied molecular orbital. I have squared the orbital using cubeman, however I am not entirely sure what the output is telling me. What I am interested in is creating a corresponding contour plot where it encompasses 95% or 90%, etc. of the electron density. Thus e/volume of orbital = 0.95, 0.9, etc. I ought to be able to take the probability density plot as encompassing ~100% of the orbital volume (given that the cube was generated with a massive number of steps). Any thoughts? Soren N. Eustis, M.A. Graduate Research Assistant Department of Chemistry Johns Hopkins University soren a jhu.edu From owner-chemistry@ccl.net Wed Oct 3 17:40:00 2007 From: "David Blackwell dtblackwell() gmail.com" To: CCL Subject: CCL: Gaussian 03 MP2 performance Message-Id: <-35301-071003161014-19629-xMFnDSYB5PWkSHn8UzDnNw+*+server.ccl.net> X-Original-From: David Blackwell Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Wed, 03 Oct 2007 20:09:23 +0100 MIME-Version: 1.0 Sent to CCL by: David Blackwell [dtblackwell[-]gmail.com] Hi all, Thanks to everyone who replied to my previous post. A few people recommended that I try RI-MP2 in ORCA; this has given very satisfactory preliminary results. Cheers, David Blackwell From owner-chemistry@ccl.net Wed Oct 3 23:29:00 2007 From: "Yangsoo Kim vsmember-x-gmail.com" To: CCL Subject: CCL:G: Question: Best combination of Linux OS & compiler for Gaussian03 Message-Id: <-35302-071003112909-10288-zAlRR4w4twQCrbh6X+RJVA(-)server.ccl.net> X-Original-From: Yangsoo Kim Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="us-ascii" Date: Wed, 3 Oct 2007 23:25:18 +0900 MIME-Version: 1.0 Sent to CCL by: Yangsoo Kim [vsmember]![gmail.com] Dear all, It looks like the recommendation from Gaussian company is not fresh enough for the Linux OS & compiler selection when a newer hardware system is used. According to their website (http://www.gaussian.com/g03_plat.htm), SuSE Linux 9.0, 9.1, 9.2, 9.3; SuSE Linux Enter. Server 9 SP3; RedHat Enterprise Linux 4 Update 2 should be used with the compiler version of Portland F77 6.1-1, while newer computer hardware (CPU, M/B...) is coming nearly every 6 month. It is, of course, understandable that Gaussian cannot update their test for all the new computer specs. On the other hand, however, it is critical for the end user to set up the fast and reliable system for running Gaussian. For example, INTEL Core2Quad Q6600 on [GIGABYTE]GA-G33-DS3R system with SATA2 HD that recently we purchased refused to install the old version of SuSe that Gaussian recommend due to the lack of drivers. My focus is how we could choose the most efficient (fast + stable) Linux OS + compiler for the newer hardware? More specifically, what OS & compiler version should be used for the INTEL Core2Quad Q6600 on [GIGABYTE]GA-G33-DS3R system? Should we use Portland F77 instead of Intel compiler all the time, even if it is known that Intel hardware performance enhances a lot when Intel compiler is used? I believe Gaussian cannot test all the possible combinations of hardware/OS/compiler, so it might be useful to setup a user report system with a bench test problem on various different combination of hardware/OS/compiler, if not already exist. Any comment is greatly appreciated, and I'll summarize the answers on request. Yangsoo Kim, Ph.D.