From owner-chemistry@ccl.net Mon Jan 14 04:52:00 2008 From: "Tom Keal keal[a]mpi-muelheim.mpg.de" To: CCL Subject: CCL: Asking for "Usage of C sharp in Computational Chemistry" Message-Id: <-36025-080114042158-30478-Azw4cfh3xHO8Wqw2AS5TKw++server.ccl.net> X-Original-From: Tom Keal Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1 Date: Mon, 14 Jan 2008 09:51:20 +0100 MIME-Version: 1.0 Sent to CCL by: Tom Keal [keal,+,mpi-muelheim.mpg.de] Hi Lukasz and Mahmoud, I absolutely agree with Lukasz's recommendations, but it's no longer the case that C# is restricted to Windows - the Mono project (http://www.mono-project.com) provides a .NET environment for Linux which includes support for C#. I'm not aware of any computational chemistry programs that use it though... In general the language you use will be dictated by whatever project you're working on. But Python is definitely a good language to start with because it will be immediately useful for scripting and will also teach you all the principles of a modern programming language. From there you can easily pick up Fortran or C later if you need to. Best wishes, Tom Keal Lukasz Cwiklik cwiklik::gmail.com wrote: > Dear Mahmoud, > Your question and my answer can start a "flame war" at CCL list, > however, I will try to write what I think as a computational chemist > who is doing quite a lot of software development in everyday work. Of > course, my opinion can be biased by my experience/environment but > there are no independent opinions in this topic anyway. > 1. As for now, C# is almost completely useless for a computational > chemist with one exception: if writes applications for MS Windows. If > you do not develop for Windows, C# is not a good choice. > 2. A knowledge of C or C++ would be necessary if you want to write a > new code. Basic knowledge of Fortran is needed if one need to work > with existing Fortran codes. In my biased opinion, in these days > starting a new project in Fortran is not an optimal choice. > 3. A knowledge of a good, modern and flexible scripting language is > also a must and I propose Python. I think, Python would be as useful > for a young computational chemist as C, C++ or Fortran. > 4. But one must remember - the choice always depends on particualar > conditions. If your advisor has a lot of experience in scientific > software development and works in this field currently, maybe his/her > reasons for choosing C# have a good basis. > > I hope other CCLers will add their suggestions here. > Best, > Lukasz From owner-chemistry@ccl.net Mon Jan 14 08:37:00 2008 From: "Rene Thomsen rt|molegro.com" To: CCL Subject: CCL: Molegro releases Molegro Molecular Viewer Message-Id: <-36027-080114075512-8861-eDAy5arUVRUdMo0s1jdFiw^^^server.ccl.net> X-Original-From: "Rene Thomsen" Date: Mon, 14 Jan 2008 07:55:09 -0500 Sent to CCL by: "Rene Thomsen" [rt .. molegro.com] Aarhus, Denmark, January 14th, 2008 - Molegro is pleased to announce the release of Molegro Molecular Viewer, a free cross-platform application for visualization of molecules and Molegro Virtual Docker results. Molegro Molecular Viewer offers a high-quality visualization tool combined with a user interface experience focusing on usability and productivity. Highlights of Molegro Molecular Viewer: * Share and view results from Molegro Virtual Docker docking runs. * Imports and exports PDB, SDF, Mol2, and MVDML files. * Automatic preparation of molecules. * Molecular surface and backbone visualization. * Labels, sequence viewer, and biomolecule generator. * Cropping of molecules and clipping planes. * Structural protein alignment. * Cross-platform: Windows, Linux, and Mac OS X is supported. To download Molegro Molecular Viewer, please visit our company website at: http://www.molegro.com. For more information contact: Rene Thomsen, CEO Molegro Hoegh-Guldbergs Gade 10, Bldg. 1090 DK-8000 Aarhus Denmark E-mail: rt]![molegro.com Phone: (+45) 89 42 31 65 About Molegro Molegro is a Danish company founded in 2005. Our company concentrates on developing high-performance drug discovery solutions leading to a faster drug-development process. Our goal is to provide scientifically superior products focusing on both state-of-the-art algorithms and an intuitive graphical user interface experience. From owner-chemistry@ccl.net Mon Jan 14 09:18:01 2008 From: "Michel Petitjean ptitjean:+:itodys.jussieu.fr" To: CCL Subject: CCL: Re(2): Asking for Usage of C sharp in Computational Chemistry Message-Id: <-36028-080114035100-21418-WS5oSwDEm5oVnQ30WBzhUw() server.ccl.net> X-Original-From: Michel Petitjean Date: Mon, 14 Jan 2008 09:50:45 +0100 (MET) Sent to CCL by: Michel Petitjean [ptitjean-x-itodys.jussieu.fr] To: chemistry . ccl.net Subject: CCL: Re(2): Asking for Usage of C sharp in Computational Chemistry /* Apologies for multiple posting of my previous reply, due to some machine troubles */ Dear Mahmoud Arafat, Since I never used Windows (yes indeed !!!), I just discovered C-sharp from your posting and one of its replies. However I heard a lot about the multiple difficulties encountered by Windows users. Unix systems are found on any kind of machine, from the smallest low-cost microcomputers including those selled with Windows (after having installed linux) and macosx (which encounters an increasing success), to the most powerful supercomputers (parallel and/or vector machines). Windows is restricted to microcomputers, and, IMHO, is inadequate for professional uses. I do not recommend to develop for this platform, except if it is to develop rapidly low quality products of small life expectation. So, I do not recommend to develop softwares with languages depending on Windows. Unix, C and FORTRAN will undoubtly survive for a long. I do not know about the others. Michel Petitjean, DSV/iBiTec-S/SB2SM (CNRS URA 2096) CEA Saclay, bat. 528 91191 Gif-sur-Yvette Cedex Phone: +33(0)1 6908 9681 / Fax: +33(0)1 6908 4007 E-mail: petitjean . itodys.jussieu.fr, michel.petitjean . cea.fr http://petitjeanmichel.free.fr/itoweb.petitjean.freeware.html Formerly: ITODYS (CNRS, UMR 7086), 1 rue Guy de la Brosse, 75005 Paris, France. Sent to CCL by: "Mahmoud Arafat Ibrahim" [M_Arafat82 * Yahoo.Com] > Dear Dr. / Prof. > In the fact, I am a new researcher in Computational Chemistry Subject and I am going to specialized in Drug Design. I have started studying C sharp as a programming language as my first step in this new field (it was an advise from one professor). After an intermediate step in C sharp, I am now confusing about its usage in my new field. Indeed, I found that all programs have been used in Computational Chemistry are based on FORTRAN or C language. So, I am now hesitating about my step and asking for advice. > > May you direct me to the best step? > -Stop studying Programming languages as whole where they are usefulness in Computational Chemistry. > -Study FORTRAN or C language where they are more abundant in Computational Chemistry. > or > -Continue studying C sharp where it is the most modern language. > I hope to read from you as soon as possible. > Sincerely; > M. Arafat, > Chemistry Department, > Faculty of Science, > Minia university, > Egypt. From owner-chemistry@ccl.net Mon Jan 14 09:53:01 2008 From: "Michel PETITJEAN michel.petitjean]|[cea.fr" To: CCL Subject: CCL: Asking for Usage of C sharp in Computational Chemistry Message-Id: <-36029-080113164355-28413-f4LV2Dbs3gGyu38QZflnmg|-|server.ccl.net> X-Original-From: "Michel PETITJEAN" Content-class: urn:content-classes:message Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="iso-8859-1" Date: Sun, 13 Jan 2008 21:48:40 +0100 MIME-Version: 1.0 Sent to CCL by: "Michel PETITJEAN" [michel.petitjean]-[cea.fr] To: chemistry ~~ ccl.net Subject: CCL: Re: Asking for Usage of C sharp in Computational Chemistry Dear Mahmoud Arafat, If you are willing to develop sophisticated programs in the future, it is much more useful to know a bit of C AND a bit of FORTRAN rather than knowing deeply C. So, you will be able to select the most adequate for a given function to programme. Note that C and FORTRAN are easily mixable (calling or called) on most unix systems, provided that you take some care in the arguments type you use. Furtermore, using advanced features of languages is not a good idea if you would like to generate portable programmes. Most people agree than FORTRAN is better for numerical programming (although it is fine for many other purposes), and than C is better for system-oriented applications. Anyway, please be able to programme by yourself rather than being dependant on somebody else that you will have to pay much and implore=20 on the knees to do the job. Michel Petitjean, DSV/iBiTec-S/SB2SM (CNRS URA 2096) CEA Saclay, bat. 528 91191 Gif-sur-Yvette Cedex Phone: +33(0)1 6908 9681 / Fax: +33(0)1 6908 4007 E-mail: petitjean ~~ itodys.jussieu.fr, michel.petitjean ~~ cea.fr http://petitjeanmichel.free.fr/itoweb.petitjean.freeware.html Formerly: ITODYS (CNRS, UMR 7086), 1 rue Guy de la Brosse, 75005 Paris, France. Sent to CCL by: "Mahmoud Arafat Ibrahim" [M_Arafat82 * Yahoo.Com] > Dear Dr. / Prof. > In the fact, I am a new researcher in Computational Chemistry Subject = and I am going to specialized in Drug Design. I have started studying C = sharp as a programming language as my first step in this new field (it = was an advise from one professor). After an intermediate step in C = sharp, I am now confusing about its usage in my new field. Indeed, I = found that all programs have been used in Computational Chemistry are = based on FORTRAN or C language. So, I am now hesitating about my step = and asking for advice. > > May you direct me to the best step? > -Stop studying Programming languages as whole where they are = usefulness in Computational Chemistry. > -Study FORTRAN or C language where they are more abundant in = Computational Chemistry. > or > -Continue studying C sharp where it is the most modern language. > I hope to read from you as soon as possible. > Sincerely; > M. Arafat, > Chemistry Department, > Faculty of Science, > Minia university, > Egypt. From owner-chemistry@ccl.net Mon Jan 14 11:46:00 2008 From: "Jim Harrison jim456harrison!=!yahoo.com" To: CCL Subject: CCL:G: optimization problem, g03 Message-Id: <-36030-080113032312-5990-7Zd+lArbnhN8SySZuWinfQ . server.ccl.net> X-Original-From: Jim Harrison Content-Transfer-Encoding: 8bit Content-Type: multipart/alternative; boundary="0-644376299-1200208954=:62408" Date: Sat, 12 Jan 2008 23:22:34 -0800 (PST) MIME-Version: 1.0 Sent to CCL by: Jim Harrison [jim456harrison- -yahoo.com] --0-644376299-1200208954=:62408 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit > from my own experience, I know that when convergence failure occurs, this implies that the input geometry is very far from the optimized state as u may already know. what I normally do is to take the final structure of the output with the convergence failure and then use it as the input. but if this final structure is again too far from the optimized one, then it will take too much time. also use opt only instead of opt=tight . if u want then u can take the final structure of the opt operation as input for another run with opt=tight. hope that will help Martha. bye. "Kowalczyk, Marta MKowalczyk++gc.cuny.edu" wrote: optimization problem, g03 Hello, I do optimizations(in G03) for squaramide diphen molecues with nitro substitution in ortho, meta, para position. I use TD DFT method with b3lyp functional (and unrestricted formalism for ions). Base 6-31g(d,p) has been used. But with 631g+(d,p) base calculations stop via the l502.exe. ----------------------------------------- Convergence failure -- run terminated. Error termination via Lnk1e in /opt/g03/l502.exe ----------------------------------------------------- This problems is present only for molecule with para-substitution(both neutral and ionic ground state). Keywords such as Fopt=tight, Fopt=verytight or change from lower to higher multiplicity doesn't help. Using option symmetry(PG=C2v) doesn't help neither. What should I do? Best regards, Marta Kowalczyk Phd student, CUNY New York --------------------------------- Never miss a thing. Make Yahoo your homepage. --0-644376299-1200208954=:62408 Content-Type: text/html; charset=iso-8859-1 Content-Transfer-Encoding: 8bit > from my own experience, I know that when convergence failure occurs, this implies that the input geometry is very far from the optimized state as u may already know. what I normally do is to take the final structure of the output with the convergence failure  and then use it as the input. but if this final structure is again too far from the optimized one, then it will take too much time. also use opt  only instead of opt=tight . if u want then u can take the final structure of the opt operation as input for another run with opt=tight. hope that will help Martha. bye.

"Kowalczyk, Marta MKowalczyk++gc.cuny.edu" <owner-chemistry~~ccl.net> wrote:
optimization problem, g03
Hello,
I do optimizations(in G03) for squaramide diphen molecues with nitro substitution in ortho, meta, para position. I use TD DFT method with b3lyp functional (and  unrestricted formalism for ions). Base 6-31g(d,p) has been used. But with 631g+(d,p) base calculations stop via the l502.exe.
-----------------------------------------
Convergence failure -- run terminated.
 Error termination via Lnk1e in /opt/g03/l502.exe
-----------------------------------------------------
This problems is present only for molecule with para-substitution(both neutral and ionic ground state).
Keywords  such as Fopt=tight, Fopt=verytight or change from lower to higher multiplicity doesn't help. Using option symmetry(PG=C2v) doesn't help neither.
What should I do?

Best regards,
Marta Kowalczyk
Phd student, CUNY
New York

Never miss a thing. Make Yahoo your homepage. --0-644376299-1200208954=:62408-- From owner-chemistry@ccl.net Mon Jan 14 12:21:00 2008 From: "Noel M O Boyle baoilleach]^[gmail.com" To: CCL Subject: CCL: Calculating Continuous Symmetry for wave-functions Message-Id: <-36031-080114115516-11432-MlJUHAopw/DOkg6TREUnHQ---server.ccl.net> X-Original-From: "Noel M O Boyle" Date: Mon, 14 Jan 2008 11:55:12 -0500 Sent to CCL by: "Noel M O Boyle" [baoilleach===gmail.com] You may find it more useful to start using an open source quantum mechanics program. That way, you will be able to use its functions for calculating overlap of MOs in your own program. I recommend you look at PyQuante, a QM program implemented in Python. It is especially designed with algorithm development in mind (I think). Noel O'Boyle > "Chaim Dryzun chdnew . yahoo.com" wrote: > > Sent to CCL by: Chaim Dryzun [chdnew::yahoo.com] > --0-2041898603-1199980433=:870 > Content-Type: text/plain; charset=us-ascii > > Hello, > > My name is Chaim Dryzun and I am a PhD student of professor David Avnir at the Hebrew University of Jerusalem, Israel. My main interst is measuring symmetry and chirality. My work includes developing new methods for calculating Continuous Symmetry Measures (CSM) and Continuous Chirality Measures (CCM). > Recently I've started to develop new method for calculating symmetry for functions. I would like to calculate symmetry for wave function of molecules and relate it to some observables (like charge transfer, absorbsion coofiecient and so on). In order for me to do that, I need an analytical function to work on. I can run the molecule at the quantum cacculation program (like Gausiaan program) with some basic set (I stared with simple ones like STO-6G and 6-31G). The programs gives me the Molecular orbitals - with the occupations of each MO and the cooficient of the AO that build the MO. I need construct all the MOs and than on each MO I will operate with a symmetry operator and than I will have to overlap the resulting function with all the MOs - this will give me a value Si (i is an index of the MO i've been working on) and the sum of all Si's=S will be the symmetry measure for the wave-function. How can help me build a program that do all this ? > > thank you in advance, > Chaim Dryzun > > > ____________________________________________________________________________________ > Looking for last minute shopping deals? > Find them fast with Yahoo! Search. http://tools.search.yahoo.com/newsearch/category.php?category=shopping > --0-2041898603-1199980433=:870 > Content-Type: text/html; charset=us-ascii > >
>
Hello,
>
 
>
My name is Chaim Dryzun and I am a PhD student of professor David Avnir at the Hebrew University of Jerusalem, Israel. My main interst is measuring symmetry and chirality. My work includes developing new methods for calculating Continuous Symmetry Measures (CSM) and Continuous Chirality Measures (CCM).
>
Recently I've started to develop new method for calculating symmetry for functions. I would like to calculate symmetry for wave function of molecules and relate it to some observables (like charge transfer, absorbsion coofiecient and so on). In order for me to do that, I need an analytical function to work on. I can run the molecule at the quantum cacculation program (like Gausiaan program) with some basic set (I stared with simple ones like STO-6G and 6-31G). The programs gives me the Molecular orbitals - with the occupations of each MO and the cooficient of the AO that build the MO. I need construct all the MOs and than on each MO I will operate with a symmetry operator and than I will have to overlap the resulting function with all the MOs - this will give me a value Si (i is an index of the MO i've been working on) and the sum of all Si's=S will be the symmetry measure for the wave-function. How can help me build a > program that do all this ?
>
 
>
thank you in advance,
>
Chaim Dryzun

>
Looking for last minute shopping deals? > Find them fast with Yahoo! Search. > --0-2041898603-1199980433=:870-- > > From owner-chemistry@ccl.net Mon Jan 14 14:30:01 2008 From: "Karen.Green-#-sanofi-aventis.com" To: CCL Subject: CCL: some history of computational programming RE: Asking for "Usage of C sharp in Computational Chemistry" Message-Id: <-36032-080114142745-31698-g4o7he2DueXu++LGDBCL2w(-)server.ccl.net> X-Original-From: Content-class: urn:content-classes:message Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="us-ascii" Date: Mon, 14 Jan 2008 11:55:13 -0700 MIME-Version: 1.0 Sent to CCL by: [Karen.Green\a/sanofi-aventis.com] Dear M. Arafat, Here is my take on it. (Others please correct any errors you might find.) In my opinion, to answer this question requires you know a little about the history of computational programming. A LITTLE HISTORY OF COMPUTATIONAL PROGRAMMING The oldest computational programs were written to be stand-alone programs in the computer languages available at the time that had natively-defined math functions. This usually meant FORTRAN. Later in the '80's you saw the development of C and so some later programs were beginning to be written in this systems language -- primarily for portability on different platforms. =20 Stand-alone programs written in these earlier languages became difficult to maintain and further develop so a new paradigm was sought. Object-oriented programming was introduced to try to better organize the code, and so object-oriented versions of common languages were developed, for example, C++ (the object-oriented version of C). In the mid-90's the internet was made available to a broader, public audience. After that, there was more focus on internet applications and direct machine-machine connectivity. The limitations of ftp were beginning to become a problem. Microsoft developed the .NET framework to address the need for direct machine connectivity over the internet -- specifically, to allow easier (more native) creation of web services. C# is the .NET version of C++. Also, there is a difference between compiled (FORTRAN,C,C++) vs. scripted languages (Perl). Usually compiled languages are used for large scale projects. WHICH LANGUAGE TO USE/LEARN IS DETERMINED BY WHAT YOU NEED TO DO So, in my opinion, what language you learn depends on what you specifically need to do, in other words, what programs you need to work with or develop to interact with. You first need to define the requirements of your programs and that will help you better understand which language(s) you should learn. =20 Generally, after a while, once you know enough computer languages, it is not too difficult to learn a new one -- similar concepts appear throughout. Hope the information helps, Karen M. Green -----Original Message----- > From: owner-chemistry.:.ccl.net [mailto:owner-chemistry.:.ccl.net]=20 Sent: Sunday, January 13, 2008 1:57 AM To: Green, Karen SMA/US Subject: CCL: Asking for "Usage of C sharp in Computational Chemistry" Sent to CCL by: "Mahmoud Arafat Ibrahim" [M_Arafat82 * Yahoo.Com] Dear Dr. / Prof. In the fact, I am a new researcher in Computational Chemistry Subject and I am going to specialized in Drug Design. I have started studying C sharp as a programming language as my first step in this new field (it was an advise from one professor). After an intermediate step in C sharp, I am now confusing about its usage in my new field. Indeed, I found that all programs have been used in Computational Chemistry are based on FORTRAN or C language. So, I am now hesitating about my step and asking for advice. May you direct me to the best step? -Stop studying Programming languages as whole where they are usefulness in Computational Chemistry. -Study FORTRAN or C language where they are more abundant in Computational Chemistry. or -Continue studying C sharp where it is the most modern language. I hope to read from you as soon as possible. Sincerely; M. Arafat, Chemistry Department, Faculty of Science, Minia university, Egypt. -=3D This is automatically added to each message by the mailing script = =3D-http://www.ccl.net/cgi-bin/ccl/send_ccl_messageSubscribe/Unsubscribe:=20http://www.ccl.net/spammers.txt From owner-chemistry@ccl.net Mon Jan 14 16:30:01 2008 From: "Michael K. Gilson gilson!^!umbi.umd.edu" To: CCL Subject: CCL: Easy links to binding data in BindingDB Message-Id: <-36033-080114162721-4370-bEKZMB89/3og0oBRdadsuQ_-_server.ccl.net> X-Original-From: "Michael K. Gilson" Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 14 Jan 2008 16:27:08 -0500 MIME-Version: 1.0 Sent to CCL by: "Michael K. Gilson" [gilson(-)umbi.umd.edu] Dear Colleagues, If you are interested in protein-ligand affinity data, you might be interested in a new set of links into BindingDB that allow programmatic access to some standard searches. They allow one to make links directly to a BindingDB query from another web-site, such as another database. For example, if you want a link to binding data for a compound with a certain SMILES string, there is a URL template into which you can insert the SMILES string and thus navigate to the desired data. There are links by: Protein (target) name BLAST by protein sequence SMILES exact compound match SMILES substructure SMILES similarity (Tanimoto with Jmol fingerprints) PDB ID (returns binding data for proteins 85% similar to the protein in the PDB ID, and with a matching ligand) PubMed ID (returns binding data from the selected article) The URL templates are here: http://www.bindingdb.org/bind/SearchTemplates.jsp Comments and suggestions are always welcome! Best regards, Mike -- Michael K. Gilson, M.D., Ph.D. CARB Fellow and Professor Center for Advanced Research in Biotechnology University of Maryland Biotechnology Institute 9600 Gudelsky Drive Rockville, MD 20850 Voice: 240-314-6217 Fax: 240-314-6255 gilsonumbi.umd.edu Lab Page: gilsonlab.umbi.umd.edu BindingDB: www.bindingdb.org From owner-chemistry@ccl.net Mon Jan 14 18:03:01 2008 From: "Soren Eustis soren|jhu.edu" To: CCL Subject: CCL:G: Stable, NoSymm and Wavefunction Stability Message-Id: <-36034-080114174956-4782-XBT8h2NpYnYvnOzPtFlqxA=server.ccl.net> X-Original-From: "Soren Eustis" Content-language: en-us Content-Type: multipart/alternative; boundary="----=_NextPart_000_0061_01C856CD.6BBF6950" Date: Mon, 14 Jan 2008 16:49:25 -0500 MIME-Version: 1.0 Sent to CCL by: "Soren Eustis" [soren]_[jhu.edu] This is a multipart message in MIME format. ------=_NextPart_000_0061_01C856CD.6BBF6950 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit I was looking at the Gaussian site and their instructions for calculating energies for antiferromagnetic complexes. In doing so, I came across the Stable keyword as a method of checking wavefunction stability. They found a few situations where the wavefunction was not stable! I always use unrestricted wavefunctions, so I have no constraints there. However, I DO NOT request nosymm for my calculations - thus G03 symmetrizes the density matrix by default, hence adding a constraint. So my question is this: Should I always test my UHF or UDFT wavefunction symmetry? Or is this just a factor in rare cases? Perhaps removing the symmetry constraint is a safer (yet costly) option in general. What are people's thoughts on the stable keyword and/or using SCF=nosymm? Thanks in advance. Soren Soren N. Eustis, M.A. Graduate Research Assistant Department of Chemistry Johns Hopkins University ------=_NextPart_000_0061_01C856CD.6BBF6950 Content-Type: text/html; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable

I was looking at the Gaussian site and their = instructions for calculating energies for antiferromagnetic complexes.  In doing = so, I came across the Stable keyword as a method of checking wavefunction = stability. They found a few situations where the wavefunction was not stable! I always = use unrestricted wavefunctions, so I have no constraints there.  = However, I DO NOT request nosymm for my calculations – thus G03 symmetrizes the = density matrix by default, hence adding a constraint.

 

So my question is this:  Should I always test = my UHF or UDFT wavefunction symmetry?  Or is this just a factor in rare = cases? Perhaps removing the symmetry constraint is a safer (yet costly) option in general.  What are people’s thoughts on the stable keyword = and/or using SCF=3Dnosymm? 

 

Thanks in advance.

 

Soren

 

Soren N. Eustis, M.A.

Graduate Research Assistant

Department of Chemistry

Johns Hopkins University

------=_NextPart_000_0061_01C856CD.6BBF6950-- From owner-chemistry@ccl.net Mon Jan 14 20:21:00 2008 From: "Kalju Kahn kalju^^chem.ucsb.edu" To: CCL Subject: CCL: extrapolation schemes Message-Id: <-36035-080114193511-10345-Z+govhv+cFrGMa6Et5GKfg-#-server.ccl.net> X-Original-From: "Kalju Kahn" Content-Transfer-Encoding: 8bit Content-Type: text/plain;charset=iso-8859-1 Date: Mon, 14 Jan 2008 15:56:36 -0800 (PST) MIME-Version: 1.0 Sent to CCL by: "Kalju Kahn" [kalju**chem.ucsb.edu] Dear Jozsef, There are lots of papers on this topic, here is just a selection of works that have influenced my choice of extrapolation schemes: For the HF energy extrapolation comparisons see: 1) Halkier, A.; Helgaker, T.; Jorgensen, P.; Klopper, W.; Olsen, "Basis-set convergence of the energy in molecular Hartree-Fock calculations" J. Chem Phys Lett 1999, 302, 437. 2) Jensen, F. "Estimating the Hartree-Fock limit from finite basis set calculations" Theor Chem Acc 2005, 113, 267 Both studies conclude that the exponential form is better than the inverse power form, and caution that the extrapolation of HF energies > from Dunning-type basis sets has a limited practical accuracy. For correlation energy (E2) in the MP2 approach, see: 3) Klopper, W. "Highly accurate coupled-cluster singlet and triplet pair energies from explicitly correlated calculations in comparison with extrapolation techniques." Mol Phys 2001, 99, 481. 4) Valeev, E. F.; Allen, W. D.; Hernandez, R.; Sherrill, C. D.; Schaefer, H. F. I. "On the accuracy limits of orbital expansion methods: Explicit effects of k-functions on atomic and molecular energies" J Chem Phys 2003, 118, 8594 These works suggest that a separate extrapolation of singlet and triplet pair contributions using inverse power law formulas provides the best results. Also, double-zeta data should be omitted, so minimally one needs cc-pVTZ and cc-pVQZ energies (although I have come across a few cases where the two-point extrapolation from cc-pVTZ and cc-pVQZ energies gives E(2) that is inferior to cc-pVQZ results. For correlation energy E(3) in MP3, see: 5) Kahn, K.; Granovsky, A. A.; Noga, J. "Convergence of Third Order Correlation Energy in Atoms and Molecules" J Comput Chem 2007, 28, 547 Again, singlet and triplet pair energies behave differently with singlet pair contribution frequently showing a delayed onset of convergence. Thus, cc-pVDZ and cc-pVTZ data should not be used for the extrapolation of E(3). Extrapolation formulas now require three data points (cc-pVQZ, cc-pV5Z, and cc-pV6Z) making such extrapolations quite expensive. For correlation energy in coupled cluster methods, see: 6) Halkier, A.; Helgaker, T.; Jørgensen, P.; Klopper, W.; Koch, H.; Olsen, J.; Wilson, A. K. "Basis-set convergence in correlated calculations on Ne, N2, and H2O" Chem Phys Lett 1998, 286, 243 3) Klopper, W. "Highly accurate coupled-cluster singlet and triplet pair energies from explicitly correlated calculations in comparison with extrapolation techniques." Mol Phys 2001, 99, 481. Double-zeta data must be excluded, and rather large basis are needed for accurate results. Finally it looks like that one can truncate the basis sets on hydrogen atoms without drastically loosing accuracy in extrapolations. Please see: 7) Mintz, B.; Lennox, K. P.; Wilson, A. K. "Truncation of the correlation consistent basis sets: An effective approach to the reduction of computational cost?" J Chem Phys 2004, 121, 5629 8) Mintz, B.; Wilson, A. K. "Truncation of the correlation consistent basis sets: Extension to third-row (Ga-Kr) molecules" J Chem Phys 2005, 122, 134106/1 9) Kahn, K.; Kahn, I. "Improved efficiency of focal point conformational analysis with truncated correlation consistent basis sets" J Comp Chem 2008 (ASAP Online: http://www3.interscience.wiley.com/cgi-bin/abstract/116835736/ Hope this helps, Kalju > > Sent to CCL by: Jozsef Csontos [jozsefcsontos%x%creighton.edu] > Dear All, > > could you point me to papers which compare different extrapolation > procedures to the basis set limit. > > Thank you, > Jozsef > > -- > Jozsef Csontos, Ph.D. > > (jozsefcsontos_at_creighton_dot_edu) > Department of Biomedical Sciences > Creighton University, > Omaha, NE> > >