From chemistry-request@server.ccl.net Mon May 21 20:31:43 2001 Received: from mailhost.msi.com ([146.202.0.242]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4M0VhI06101 for ; Mon, 21 May 2001 20:31:43 -0400 Received: (from daemon@localhost) by mailhost.msi.com (980427.SGI.8.8.8/970903.SGI.AUTOCF) id RAA16340 for ; Mon, 21 May 2001 17:32:03 -0700 (PDT) Received: from unknown(146.202.5.50) by mailhost.msi.com via smap (V2.0) id xmab16316; Mon, 21 May 01 17:31:38 -0700 Message-Id: <5.1.0.14.0.20010521171516.00a838f0@pop3.norton.antivirus> X-Sender: jnauss/sparky2.msi.com@pop3.norton.antivirus X-Mailer: QUALCOMM Windows Eudora Version 5.1 Date: Mon, 21 May 2001 17:16:20 -0400 To: chemistry@ccl.net From: Jeff Nauss Subject: MSI Workshop on Macromolecular X-ray Structure Determination Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Molecular Simulations Inc, will be holding a 2-day "Macromolecular X-ray Structure Determination Workshop" at the San Diego Supercomputing Center, La Jolla, CA. The event will be held July 17-18. This workshop is designed for x-ray crystallographers who wish to be more effective with the MSI software for structure determination. The workshop will encompass MSI's principal software tools for protein crystallography; > from model building to structure refinement and validation. The QUANTA and CNX user interfaces will be used with an emphasis on efficient use of the software. During this two-day workshop, the following topics will be presented. After each lecture, an extended discussion/hands-on session will give attendees the opportunity to try out new strategies and methodologies and to discuss their own applications with scientists from MSI. De-novo electron density map interpretation: bones skeletonization and C-alpha atoms tracing Automated protein model building and refinement Automated ligand fitting in protein-ligand complexes using electron density maps Solvent placement Structure refinement using CNX interface of QUANTA Structure validation Phasing X-ray data using Molecular replacement in CNX Structure refinement using CNX No experience with molecular modeling is required; however, experience in x-ray crystallography is essential. Fees for the 2-day course are $1000 commercial, $500 government, and $400 academic. Please note that on June 1, MSI, Synopsys, Oxford Molecular, and GCG will become Accelrys. Further detailed information about this and other MSI training workshops, as well as on-line course registration, can be found at MSI's website (http://www.msi.com/about/events/training). Please do not hesitate to contact us should you have any questions. Thank you very much. Jeffrey L. Nauss MSI Customer Training Programs -- Jeffrey L. Nauss, PhD Phone: (858) 799-5555 Customer Training Programs Fax: (858) 458-0136 Molecular Simulations Inc. E-mail: jnauss@msi.com 9685 Scranton Road http://www.msi.com/about/events/training San Diego, CA 92121-3752 MSI, Synopsys, Oxford Molecular, and GCG will become Accelrys on June 1 From chemistry-request@server.ccl.net Tue May 22 08:03:27 2001 Received: from mailhost1.rdg.ac.uk ([134.225.16.4]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MC3QI30490 for ; Tue, 22 May 2001 08:03:26 -0400 Received: from indigog.chemistry ([134.225.168.11] helo=chemistry.reading.ac.uk) by mailhost1.rdg.ac.uk with esmtp (University of Reading Email Service) id {152AsU-0003dH-00} for chemistry@ccl.net; Tue, 22 May 2001 13:03:06 +0100 Received: from indigol.chemistry.reading.ac.uk (indigol.chemistry.rdg.ac.uk [134.225.168.15]) by chemistry.reading.ac.uk (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id OAA45964 for ; Sat, 31 Oct 1998 14:04:45 GMT Received: from reading.ac.uk (localhost [127.0.0.1]) by indigol.chemistry.reading.ac.uk (SGI-8.9.3/8.9.3) with ESMTP id LAA06818 for ; Tue, 22 May 2001 11:02:57 -0100 (BST) Sender: tris@chemistry.reading.ac.uk Message-ID: <3B0A5570.93F4C28D@reading.ac.uk> Date: Tue, 22 May 2001 11:02:56 -0100 From: Tris Youngs X-Mailer: Mozilla 4.7C-SGI [en] (X11; I; IRIX 6.5 IP22) X-Accept-Language: en MIME-Version: 1.0 To: chemistry@ccl.net Subject: Lanthanide complexes in Gaussian98. Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hi. I am a relative newcomer to computational chemistry, and part of my work focuses on the calculation of lanthanide-containing complexes. The complexes in question consist of Ln3+ (where Ln=La-Lu), a tridentate nitrogen-bearing ligand, a solvent molecule (water) and three counterionic NO3- groups. In G98 I have successfully run the calculations for La and Lu without problem, but the rest of the series (even Gd which should be next simplest as the f-shell is entirely spin-unpaired) has proven to be more difficult. I'm using ECP's from the Stuttgart group (Dolg et.al.) for the Lanthanides and the standard 6-31g(d) basis set for the rest of the atomic centres. My main problem is that of spin contamination using the UHF method - I had found one suggestions which was to use the ROHF method, but this was as unsuccessful. I have checked and double-checked the spin multiplicities (i.e. in Gaussian, Gd=8, Ce=2, Pr=3 e.t.c.). Any thoughts or comments would be appreciated. Regards Tris Youngs scr00tgy@reading.ac.uk University Of Reading From chemistry-request@server.ccl.net Tue May 22 08:45:05 2001 Received: from iris.sipp.ac.cn ([202.127.25.14]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MCj4I31536 for ; Tue, 22 May 2001 08:45:04 -0400 Received: from iris.sipp.ac.cn [202.127.25.100] by iris.sipp.ac.cn with ESMTP (SMTPD32-6.04) id AED77C5E00D2; Tue, 22 May 2001 20:43:03 +0800 Sender: swift@ccl.net Message-ID: <3B0B31FD.509E20F9@iris.sipp.ac.cn> Date: Tue, 22 May 2001 20:43:57 -0700 From: xmao Organization: iris.sipp.ac.cn X-Mailer: Mozilla 4.7C-SGI [en] (X11; I; IRIX 6.3 IP32) X-Accept-Language: en MIME-Version: 1.0 To: chemistry@ccl.net Subject: G2 model calculation Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dear everyone: I wonder if there are some freewares that could do G2 model calculation. Thanks. Regards, mao xiang From chemistry-request@server.ccl.net Tue May 22 11:43:03 2001 Received: from electre.pasteur.fr ([157.99.64.120]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MFh0I02959 for ; Tue, 22 May 2001 11:43:00 -0400 Received: (from nathsav@localhost) by electre.pasteur.fr (8.11.3/8.11.3) id f4MFgug292929; Tue, 22 May 2001 17:42:56 +0200 (CEST) Date: Tue, 22 May 2001 17:42:56 +0200 From: Nathalie Duclert To: chemistry@server.ccl.net Cc: Nathalie Duclert Subject: AMBER and GDP/GTP Message-ID: <20010522174256.A292312@electre.pasteur.fr> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i Hi, I'm a new Amber user. I would like to simulate complex protein-GDP and protein-GTP. Has anybody parameters for GDP or/and GTP ? I wonder if there is any existing file for NDP or NTP. Thanks very much for any suggestions, -- Nathalie Duclert-Savatier ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Unite de Bio-Informatique Structurale | nathsav@pasteur.fr Institut Pasteur | Tel : +33 1 45 68 87 37 28, rue du Docteur Roux | Fax : +33 1 45 68 87 19 75724 Paris cedex 15 | From chemistry-request@server.ccl.net Tue May 22 11:43:43 2001 Received: from outbound.mail.uoguelph.ca ([131.104.96.18]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MFhgI02979 for ; Tue, 22 May 2001 11:43:42 -0400 Received: from msnet.mathstat.uoguelph.ca (msnet.mathstat.uoguelph.ca [131.104.32.59]) by outbound.mail.uoguelph.ca (8.9.3 (PHNE_21697)/8.9.3) with ESMTP id LAA16489 for ; Tue, 22 May 2001 11:43:40 -0400 (EDT) Received: from MSNET/SpoolDir by msnet.mathstat.uoguelph.ca (Mercury 1.48); 22 May 01 11:42:40 -0500 Received: from SpoolDir by MSNET (Mercury 1.48); 22 May 01 11:42:11 -0500 From: "Bill Smith" Organization: Math & Stats, University of Guelph To: Chemistry@ccl.net Date: Tue, 22 May 2001 11:42:03 -0400 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: Re: CCL:Group Contribution Methods Message-ID: <3B0A508C.29676.BA4770@localhost> Priority: normal In-reply-to: X-mailer: Pegasus Mail for Win32 (v3.12c) THERGAS, from (URL all on one line) http://www.ensic.u- nancy.fr/ENSIC/DCPR/cinetique/Chimieinfo/TEXTES/thergas.htm On 16 May 2001, at 15:23, Golab, Joseph T wrote: > Hello: > > In updating a software catalogue, are there any other programs (in > addition to NIST database 25 and CHETAH) that can calculate heats of > formation, entropy, heats of reaction, etc. based on Benson's (or > Benson-like) Group Contribution Methods. Thanks in advance. > > Joe > ________________________________________________________________ > Joseph T. Golab Phone: +1 (630) 961-7878 > BP Naperville Complex, C-7 Fax: +1 (630) 420-4382 > 150 W. Warrenville Rd e-mail: GolabJT@BP.com > Naperville, IL 60563 SOCON: 231 > > > > -= This is automatically added to each message by mailing script =- > CHEMISTRY@ccl.net -- To Everybody | CHEMISTRY-REQUEST@ccl.net -- To > Admins MAILSERV@ccl.net -- HELP CHEMISTRY or HELP SEARCH > CHEMISTRY-SEARCH@ccl.net -- archive search | Gopher: > gopher.ccl.net 70 Ftp: ftp.ccl.net | WWW: > http://www.ccl.net/chemistry/ | Jan: jkl@ccl.net > > > > > Best Regards, W. R. Smith, Professor Dept. of Mathematics and Statistics and School of Engineering Room 546 MacNaughton Building University of Guelph Stone Road and Gordon Streets Guelph, Ontario, CANADA N1G 2W1 Tel: 519-824-4120, ext. 3038; FAX: 519-837-0221; http://www.mathstat.uoguelph.ca/faculty/smith/ From chemistry-request@server.ccl.net Tue May 22 12:46:48 2001 Received: from ivory.trentu.ca ([192.75.12.103]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MGkmI04196 for ; Tue, 22 May 2001 12:46:48 -0400 Received: from trentu.ca ([204.225.13.50]) by trentu.ca (PMDF V5.2-32 #29543) with ESMTP id <01K3V5D5RI260019Z5@trentu.ca> for chemistry@ccl.net; Tue, 22 May 2001 12:45:26 EDT Date: Tue, 22 May 2001 12:51:22 -0400 From: elewars Subject: MOLECULAR MECHANICS CALC OF IR SPECTRA To: chemistry@ccl.net Message-id: <3B0A9909.77631193@trentu.ca> MIME-version: 1.0 X-Mailer: Mozilla 4.7 [en] (WinNT; I) Content-type: text/plain; charset=us-ascii Content-transfer-encoding: 7bit X-Accept-Language: en 2001 May 22 Hello, Molecular mechanics programs can now calculate the positions (cm**-1) of normal-mode vibrations. However, MM by itself cannot calculate the _intensities_ of these vibrations, since this depends on the change in dipole moment, which in turn depends on electron distribution, a concept outside MM theory. It seems, however, that in principle MM could calculate intensities using assigned bond moments or atom charges, which would provide a rapid way of calculating IR spectra. Has this been implemented in any program? Thanks E. Lewars ======== From chemistry-request@server.ccl.net Tue May 22 13:10:01 2001 Received: from sims.csuohio.edu ([137.148.5.58]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MHA1I04463 for ; Tue, 22 May 2001 13:10:01 -0400 Received: from dhcp.csuohio.edu (scifac209-53.dhcp.csuohio.edu [137.148.209.53]) by sims.csuohio.edu (Sun Internet Mail Server sims.4.0.1999.09.28.17.31.p2) with ESMTP id <0GDQ00M88YIO6W@sims.csuohio.edu> for CHEMISTRY@ccl.net; Tue, 22 May 2001 13:13:37 -0400 (EDT) Date: Tue, 22 May 2001 11:27:21 -0400 From: David W Ball Subject: CBS-Q manually X-Sender: d.ball@popmail.csuohio.edu To: CHEMISTRY@ccl.net Message-id: <5.0.2.1.0.20010522112025.00a417a0@popmail.csuohio.edu> MIME-version: 1.0 X-Mailer: QUALCOMM Windows Eudora Version 5.0.2 Content-type: text/plain; format=flowed; charset=us-ascii Content-transfer-encoding: 7BIT I'm trying to determine the CBS-Q energy of a molecule manually, using data generated by G98. I've figured out all the terms in the final summary section except for the one labeled 'DE(CBS)'. The original references aren't specific to the G98 way of doing it, of course, and the GAUSSIAN web pages don't have that specific into. Can anyone tell me what raw data are used to determine this correction? Many thanks. dwb From chemistry-request@server.ccl.net Tue May 22 11:58:25 2001 Received: from ccl.net ([192.148.249.4]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MFwPI03368 for ; Tue, 22 May 2001 11:58:25 -0400 Received: from gap.cco.caltech.edu (gap.cco.caltech.edu [131.215.139.43]) by ccl.net (8.9.3/8.9.3/OSC 2.0) with ESMTP id LAA02419 for ; Tue, 22 May 2001 11:58:22 -0400 (EDT) Received: (from news@localhost) by gap.cco.caltech.edu (8.9.3/8.9.3) id IAA26382; Tue, 22 May 2001 08:39:02 -0700 (PDT) To: mlist-chemistry@nntp-server.caltech.edu Path: news From: "Richard P. Muller" Newsgroups: mlist.chemistry Subject: Summary: Integrals over Gaussian Functions Date: Tue, 22 May 2001 08:38:24 -0700 Organization: California Institute of Technology, Pasadena Lines: 37 Message-ID: <3B0A87F0.1036D7C7@wag.caltech.edu> NNTP-Posting-Host: dhcp-26-202.wag.caltech.edu Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 4.76 [en] (Win98; U) X-Accept-Language: en I've gotten many good references for integrals over Gaussians. The one that everyone seems to like is: "Molecular Integrals over gaussian basis functions" Peter M.W. Gill, Adv. Quant. Chem. 25 (1994) 141. which is a very complete article on the subject. One person mentioned: "Gaussian Basis Sets and Molecular Integrals" by Trygve Helgaker and Peter Taylor in Modern Electronic Structure Theory Part II, edited by David Yarkony and published by World Scientific, 1995. This is a massive article. I haven't made it all the way through (yet), but it is extremely well detailed. Amazing how much I didn't know about Gaussian basis functions. Here are a couple more references that ccl-ers may find interesting. Isaiah Shavitt's "The Gaussian Function in Calculations of Statistical Mechanics and Quantum Mechanics," _Methods in Comp. Phys_, v. 2., B. Adler, S. Fernbach, M. Rotenberg, eds. (Academic Press, 1963) is very well written. I haven't (yet) gotten Boys RMP article, but this came out shortly after that article, and with that article is one of the foundations of the field. My own personal favorite reference for one-electron integrals is "General Quantum Mechanical Operators. An Open-Ended Approach for One-Electron Integrals with Gaussian Bases." J. Augspurger and C. Dykstra, J. Comp. Chem. 11, 105-111 (1990). Thanks for everyone's help. Rick -- Richard P. Muller, Ph.D. rpm@wag.caltech.edu http://www.wag.caltech.edu/home/rpm From chemistry-request@server.ccl.net Tue May 22 12:06:06 2001 Received: from mail.takas.lt ([212.59.0.10]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MG65I03506 for ; Tue, 22 May 2001 12:06:05 -0400 Received: from julius (dialup391.vln.takas.lt [212.59.3.159]) by mail.takas.lt (8.9.1/8.9.0) with SMTP id SAA2561155 for ; Tue, 22 May 2001 18:05:56 +0200 (GMT+0200) Message-ID: <000201c0e2d8$d3b2d620$9f033bd4@ktl.mii.lt.lorka.ktl.mii.lt> From: "Olegas Eicher-Lorka" To: Subject: G98W on dual processors: SUMMARY Date: Tue, 22 May 2001 16:07:28 +0200 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_002B_01C0E2D9.4C5CBAA0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2615.200 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2615.200 This is a multi-part message in MIME format. ------=_NextPart_000_002B_01C0E2D9.4C5CBAA0 Content-Type: text/plain; charset="windows-1257" Content-Transfer-Encoding: quoted-printable Few days ago, I asked the following question: Dear CClers, We've been using G98W for about a year on a lowly PC. We're just about = to get a dual processor system (2*PIII 933MHz) and were looking forward = to fast calcuations only to be told that G98W only runs on a single = processor. Any add-ons, tips or advice? Thanks Dr. Olegas Eicher-Lorka Institute of Chemistry Akademijos 7 e-mail: lorka@takas.lt tel.: +(370-)2-729372 We're waiting to find out how easy it is to "format" our system so that = one processor is used for only G98W and the other does everything else. = As soon as we find out we'll tell you. All in all, we didn't quite receive what we wanted but... Thanks to all for your answers. Here are a summary of the answers, I = received: Shure: Install Linux and Jaguar on your PC, if you are looking forward for really fast calculations. We have ported parallel Jaguar to Linux and it will be available in the next weeks - I hope. If you are interested ask me for more informations or check out http://www.schrodinger.com/Products/jaguar.html Thanks, Gerd =20 Hi, If you want to run G98 in parallel on Intel, I think you will have to switch to Linux, and compile an SMP binary. This assumes that you buy the source code and the Portland Group Fortran compiler which is required to compile Gaussian. Gaussian may sell an Intel SMP binary, but I'm not sure. As I recall, you cannot run two copies of G98W at the same time on Windows, which would at least allow you to utilize both processors for two concurrent jobs. Don't take my word on that, though, as I haven't run the Windows version. - John bushnell@chem.ucsb.edu G98W cannot work parallely on Windows platform because G98W was not = programmed for parallel computations. But G98 can work parallely on dual processors = Linux systems. That is to say, you need install Linux on your new system and = get a linux version of G98. Yubo Dear Olegas, I asked Gaussian (about a year ago) about running G98W on multiple processors on a PC. Douglas Fox responded and told me that they are = working with Portland to get a version to run parallel on Linux, and that a = parallel version for Windows would come afterwards. I got the impression that it would be quite a while until a parallel Windows version would be = available. Kind regards, Gerald J. Tanoury, Ph.D. Principal Research Chemist Sepracor Inc. 111 Locke Dr. Marlborough, MA 01752 Ph: (508) 357-7442 Fax: (508) 357-7808 E-mail: gtanoury@sepracor.com If you're looking for a speedup from parallelization, you're not going = to get much from 2 processors. To get much benefit you need a massively parallel system. Why not just run 2 jobs at once, getting essentially = the same benefit as running one in parallel. You also have the flexibility = of being able to run one job and have a processor free for other tasks. Jaime Martell, Ph.D. Assistant Professor Department of Chemistry SUNY Potsdam Potsdam, NY 13676 Phone: 315-267-2268 Fax: 315-267-3170 e-mail: marteljm@potsdam.edu WWW: http://www2.potsdam.edu/marteljm/ If you want dual processor and parallel support for QC calculations on a Windows PC, try pcGAMESS. It can be located at: http://classic.chem.msu.su/gran/gamess/index.html Jim Kress So maybe this explains why G98W is half the price of G98? ------=_NextPart_000_002B_01C0E2D9.4C5CBAA0 Content-Type: text/html; charset="windows-1257" Content-Transfer-Encoding: quoted-printable
Few days ago, I asked the following=20 question:

Dear CClers,
We've been using G98W for about a = year on a=20 lowly PC. We're just about to get a dual processor system (2*PIII = 933MHz) and=20 were looking forward to fast calcuations only to be told that G98W only = runs on=20 a single processor.
 
Any add-ons, tips or = advice?
 
Thanks
 
Dr. Olegas Eicher-Lorka
Institute = of=20 Chemistry
Akademijos 7
e-mail: lorka@takas.lt
tel.:=20 +(370-)2-729372
 
We're waiting to find out how easy it is to "format" our system so = that one=20 processor is used for only G98W and the other does everything else. As = soon as=20 we find out we'll tell you.
 
All in all, we didn't quite receive what we wanted but...
 
Thanks to all for your answers. Here are a summary of = the=20 answers, I received:
Shure: Install Linux and Jaguar on = your PC, if=20 you are looking forward
for really fast calculations. We have ported = parallel=20 Jaguar to Linux
and it will be available in the next weeks - I hope. = If you=20 are
interested ask me for more informations or check out

http://www.schro= dinger.com/Products/jaguar.html

Thanks,

Gerd
 
Hi,

  If you want to run = G98 in=20 parallel on Intel, I think you
will have to switch to Linux, and = compile an=20 SMP binary.
This assumes that you buy the source code and the=20 Portland
Group Fortran compiler which is required to compile=20 Gaussian.
Gaussian may sell an Intel SMP binary, but I'm not = sure.
As I=20 recall, you cannot run two copies of G98W at the same
time on = Windows, which=20 would at least allow you to utilize
both processors for two = concurrent=20 jobs.  Don't take my
word on that, though, as I haven't run the = Windows=20 version.

       -=20 John           &nb= sp;  =20 bushnell@chem.ucsb.edu

= G98W=20 cannot work parallely on Windows platform because G98W was not = programmed
for=20 parallel computations. But G98 can work parallely on dual processors=20 Linux
systems. That is to say, you need install Linux on your new = system and=20 get a
linux version of G98.

Yubo

Dear Olegas,
I = asked=20 Gaussian (about a year ago) about running G98W on multiple
processors = on a=20 PC.  Douglas Fox responded and told me that they are = working
with=20 Portland to get a version to run parallel on Linux, and that a=20 parallel
version for Windows would come afterwards.  I got the=20 impression that it
would be quite a while until a parallel Windows = version=20 would be available.

Kind regards,

Gerald J. Tanoury,=20 Ph.D.
Principal Research Chemist
Sepracor Inc.
111 Locke=20 Dr.
Marlborough, MA  01752
Ph: (508) 357-7442
Fax: (508)=20 357-7808
E-mail: gtanoury@sepracor.com

If=20 you're looking for a speedup from parallelization, you're not going = to
get=20 much from 2 processors.  To get much benefit you need a=20 massively
parallel system.  Why not just run 2 jobs at once, = getting=20 essentially the
same benefit as running one in parallel.  You = also have=20 the flexibility of
being able to run one job and have a processor = free for=20 other tasks.

Jaime Martell, Ph.D.
Assistant = Professor
Department of=20 Chemistry
SUNY Potsdam
Potsdam, NY 13676
Phone:=20 315-267-2268
Fax:   315-267-3170
e-mail: marteljm@potsdam.edu
WWW: http://www2.potsdam.edu/martel= jm/

If you want dual processor and = parallel=20 support for QC calculations on a
Windows PC, try pcGAMESS.  It = can be=20 located at:

http://classic= .chem.msu.su/gran/gamess/index.html

Jim=20 Kress

So maybe this explains why G98W is half the price of=20 G98?
------=_NextPart_000_002B_01C0E2D9.4C5CBAA0-- From chemistry-request@server.ccl.net Tue May 22 12:30:49 2001 Received: from hotmail.com ([209.185.241.99]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MGUmI04015 for ; Tue, 22 May 2001 12:30:49 -0400 Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC; Tue, 22 May 2001 09:30:43 -0700 Received: from 130.120.89.37 by lw3fd.law3.hotmail.msn.com with HTTP; Tue, 22 May 2001 16:30:42 GMT X-Originating-IP: [130.120.89.37] From: "Gustavo Velardez" To: chemistry@ccl.net Subject: Compilation problems Date: Tue, 22 May 2001 16:30:42 -0000 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: X-OriginalArrivalTime: 22 May 2001 16:30:43.0153 (UTC) FILETIME=[8BBDD410:01C0E2DC] Hi everybody, I am trying to compile G98 A.7 in a PC with Red Hat 7.1. I have all the libraries and I followed all the instructions but I have the following error: -------------------------------------------------------------------- bsd/mdutil.c: In function `etime_': bsd/mdutil.c:2361: `CLK_TCK' undeclared (first use in this function) bsd/mdutil.c:2361: (Each undeclared identifier is reported only once bsd/mdutil.c:2361: for each function it appears in.) bsd/mdutil.c: In function `fdate_': bsd/mdutil.c:2374: warning: assignment makes pointer from integer without a cast make: *** [mdutil.o] Error 1 ar rlv ../util.a mdutil.o ar: mdutil.o: No such file or directory endif rm mdutil.o rm: cannot remove `mdutil.o': No such file or directory endif ... ... pgf77 -O2 -tp p5 -Munroll -Mvect=cachesize:524288 -c ml0.f rm -f ml0.f ml0.c make[1]: Leaving directory `/home/gust/g98' pgf77 -O2 -tp p5 -Munroll -Mvect=cachesize:524288 -o g98 ml0.o util.so /applx1/pgi/linux86/lib/libpgftnrtl.a(cnfg.o): In function `__fio_scratch_name': cnfg.o(.text+0x33): the use of `tempnam' is dangerous, better use `mkstemp' ml0.o: In function `MAIN_': ml0.o(.text+0x2b1): undefined reference to `gauabt_' ... ... util.so: undefined reference to `galloc1_' make: *** [g98] Error 1 endif -------------------------------------------------------------------- and the compilation is stopped. What is the problem? Is there some library or something I have forgotten? Thank you for your attention Gustavo Velardez PhD LCAR-IRSAMC. Université Paul Sabatier. 118 route de Narbonne. 31062. Toulouse. France. TE: +33-(0)5-61-55-60-09 FAX:+33-(0)5-61-55-83-17 _________________________________________________________________________ Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com. From chemistry-request@server.ccl.net Tue May 22 17:16:05 2001 Received: from bellatrix.pcl.ox.ac.uk ([163.1.35.134]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4MLG4I07594 for ; Tue, 22 May 2001 17:16:04 -0400 Received: from localhost (ben@localhost) by bellatrix.pcl.ox.ac.uk (8.11.2/8.11.2) with ESMTP id f4MLG2v18579; Tue, 22 May 2001 22:16:02 +0100 Date: Tue, 22 May 2001 22:16:01 +0100 (BST) From: Ben Webb To: Gustavo Velardez cc: Subject: Re: CCL:Compilation problems In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII On Tue, 22 May 2001, Gustavo Velardez wrote: > Hi everybody, > I am trying to compile G98 A.7 in a PC with Red Hat 7.1. I have all the > libraries and I followed all the instructions but I have the following > error: > > -------------------------------------------------------------------- > bsd/mdutil.c: In function `etime_': > bsd/mdutil.c:2361: `CLK_TCK' undeclared (first use in this function) This problem is caused by the change from Linux kernel 2.2 (as used by RedHat 7.0 and earlier) to 2.4 (as used in RedHat 7.1). The fix is simple: add the line "#include " to bsd/mdutil.c (around about line 195 should do). Hopefully, fixing this problem will also fix your other problems. (If not, I can't help you further, because I don't observe these problems on my system!) Ben -- ben@bellatrix.pcl.ox.ac.uk http://bellatrix.pcl.ox.ac.uk/~ben/ "All science is either physics or stamp collecting." - Ernest (1st Baron) Rutherford From chemistry-request@server.ccl.net Tue May 22 23:40:59 2001 Received: from smtp.csc.fi ([193.166.0.99]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f4N3exI16064 for ; Tue, 22 May 2001 23:40:59 -0400 Received: from sampo.csc.fi (sampo.csc.fi [193.166.1.154]) by smtp.csc.fi (8.9.3/8.9.3/CSC) with ESMTP id GAA24071 for ; Wed, 23 May 2001 06:40:57 +0300 Date: Wed, 23 May 2001 06:40:57 +0300 (EEST) From: Maija Lahtela To: Subject: Summary:Gaussian in PC clusters Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi All, Few days ago, I asked the question about Gaussian in PC clusters: I have got several answers and useful hints. Thanks to All CCLers. I may e-mail later to ask you more detailed questions. Yours, Maija Lahtela-Kakkonen From: Szilveszter Juhos Check the Condor job scheduling system at: http://www.cs.wisc.edu/condor/ Furthermore be careful with the RedHat gcc compiler: it is not reliable. If you can manage better to use Debian (for commercial support check: www.progeny.com) Szilva *************************************************** From: walid_2001q@arabia.com Dear Scientist if the problem of yor run is the speed so collecting many Pc May solve it I like To Know How efficient the system become after that I think it will be about 50 % of exepected but if the problem is Hard disk space the collection of many PC will not solve this at last i like to know about your souce of Gaussian 98 under unix since i Work with it under Windows98 And i wish a copy of answers you have recieved since i interested in the cluster computer systema ********************************************************* From: Dr. Robert Spiske can you please give me a litle more detail to your gaussian? We found that it takes 4 Mips R10000 @180MHz to overtake an 1.1MHz Athlon. So i wonder what went wrong. If you compile your gaussian with the normal settings it will be optimized to use 512 kb of cache. Modern Intel and AMD cpu's have only 256 kb cache. Correct settings within the makefile can optimize the performace. gaussian needs an enourmous amount of memory and lots for scratchfile space. therefor gaussian nodes should have a local swapfile (best) or a second fast nfs-network (not so good). Some calculations (i believe it was CCSD) caused so much scratchfile activity that the cpu-usage of the nfs-daemon was bigger than 50%. And 3 of our 4 CPU's did nothing but waiting. A gaussian node should have at least 512MB/CPU if the memory is bigger the rest can serve as file cache allowing the system to continue faster after file i/o. By the way there are problems with nfs within the linux 2.x.x-Kernels that can cause severe problems! To run gaussian in a cluster you will need the LINDA Extensions, we did not do that by now. There are mpi informations within the code for use on HP-machines. Our university has a 32cpu version and we found gaussian to scale up to 4-6 cpu sometimes 8. I dont think yo will not benefit by the use of more cpu's in a single calculation. Hope that helps a little? head aega! Robert *************************************************** From: Doug Fox Maija, Carlos Sosa also copied me on some of this. I did not see if you responded to him with clarification on "slow execution" and in particular "slow wrt to Windows" which is not our experience. We use the same compilers for Windows and Linux (Portland Group) and see very similar performance. So are these the same machines dual booted or at least equivalent machines? What types of calculations are you going for? The scaling for clusters will not go to 100 PC's, this is not a matter of Linda vs. MPI. Rather most problems are not large enough and so you fall below the optimal size on each processor. There are also issues with communications overhead. We support Linda for these clusters and the licensing supports clusters of any size for G98+Linda so it is a flat fee and you can run 12 8 processor calculations, this is done by a number of groups. Also note that parallel execution works best for CPU bound processes like HF, DFT and even CIS=DIRECT. MP2 energy and gradients are also reasonable because we can use local disk on each node to make disk I/O and CPU in parallel. If you are doing MP4, CBS-QB3, or G2 you are going to hit disk I/O limits and not get great parallel speedups. See the paper in J. Comp. Chem. vol 19 1053, (1998) which gives you a good idea what kind of speedups are likely for various methods. ******************************************************** From: John Bushnell Hi, I saw your post on ccl. We are also planning on setting up a cluster for running Gaussian and other programs, and I am also very interested in what has worked for others, and how well, etc. I was curious what kinds of jobs you had run there for which the performance was dissapointing. Were they HF, DFT, or some higher level of theory? And were they "large" in terms of number of basis sets? I assume that you were running Gaussian compiled with pgf on Linux? I am curious because most single cpu PC timings that I have seen look pretty good, for instance the following recently reported comparison (admittedly very small jobs!). Specifically, the relative speed (for these jobs) of the Pentium/Athlon systems vs. an Origin 2000 (I am assuming an R10000?) seemed good: ********************************************************** From: M. Nicklaus For what it's worth: Here's a recent small series of benchmarks we've run on various Linux systems, plus a Cray and SGI Origin added for comparison. Program: Gaussian 98 Rev. A.7. All executables exactly identical for the Linux systems (copied between machines), compiled w/ PGI v. 3.2 (on Linux), G.98 standard compilation (no tuning/hacking). Jobs run: G.98 test jobs # 1, 28, 94, 155, 194, 296, 302, aggregate time, as reported in output, all runs single-CPU. CPU Speed Chipset Kernel Distro glibc Memory HD/Contr. time min. 256MB MHz, RPM (sec) (all: U-ATA/) P 4 1.5GHz Int. 850 2.4.2-2 RH 7.1 2.2.2-10 RDRAM-800 100, 7200 278 Athlon 1.33 VIA 686B 2.4.0 SuSE 7.1 2.2-7 DDR 266 66, 7200 299 P III 1 GHz Int. 815 2.4.3 (RH 6.1) 2.1.2 SDRAM 133 100, 7200 497 P III 1 GHz Int. 815 2.4.3 (RH 6.1) 2.1.2 SDRAM 133 100, 5400 497 P III 1 GHz Int. 815 2.2.12 RH 6.1 2.1.2 SDRAM 133 100, 7200 646 P III 1 GHz Int. 815 2.2.12 RH 6.1 2.1.2 SDRAM 133 100, 5400 646 P III 866MHz S/Works 2.4.3 (RH 6.2) 2.1.3-15 SDRAM 133 33, 7200 453 P III 866MHz S/Works 2.2.14 RH 6.2 2.1.3-15 SDRAM 133 33, 7200 526 SGI Origin 2000 921 Cray SV-1 2400 "(...)" in the Distro column denotes newer kernel than out-of-the-box. Going from a 2.2 kernel to a 2.4 kernel seems to speed things up considerably, at least on these systems. Hard drive and IDE controller speeds appear to be near irrelevant for these jobs Regarding the mpi version of Gaussian, I believe that is simply a matter of compilation. At least on SGI(IRIX) you can specify the number of processors during compilation. I assume that compilation with the Portland Group compiler on Linux is similar. Yes, definitely summarize if you find out anything interesting. I think a lot of folks are considering PC clusters but facing a lot of uncertainty as far as performance and implementation. - John bushnell@chem.ucsb.edu PS: Linda is expensive! It would be nice to be able to use MPICH if it was efficient. ********************************************************** From: Christoph Steinbeck steinbeck@ice.mpg.de I'm using a 16 node PC cluster for running Gaussian 98 using Linda as the parallel execution environment. Works nicely. It must, however, be noted that for our type of calculation, a reasonable scaling is only achieved up to eight or ten nodes per job. The results of our initial, provisional evaluation are summarized under http://www.ice.mpg.de /departments/ChemInf/edda.html. Follow the last link on the scaling behaviour. The result of this behaviour is that I use the OpenPBS batch queue system to load-schedule jobs between subclusters of eight nodes each. Cheers, Chris ***************************************************** Maija Lahtela-Kakkonen, Application Scientist / Chemistry CSC-Scientific Computing Tekniikantie 15 a D, P.O.Box 405 FIN-02101 ESPOO FINLAND TEL 358-9-4572079 /050-3819506, FAX 358-9-4572302 E-MAIL mlahtela@csc.fi, Internet:www.csc.fi ****************************************************