From chemistry-request@server.ccl.net Thu Aug 9 09:08:04 2001 Received: from ucidoor.unitedcatalysts.com ([208.23.162.2]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79D83v23566 for ; Thu, 9 Aug 2001 09:08:04 -0400 Received: by ucidoor.unitedcatalysts.com; (8.8.8/1.3/10May95) id JAA20666; Thu, 9 Aug 2001 09:08:03 -0400 (EDT) Received: from 10.1.0.50 by ucismtp02.unitedcatalysts.com (InterScan E-Mail VirusWall NT); Thu, 09 Aug 2001 09:09:53 -0400 (Eastern Daylight Time) Received: from lvlxch01.unitedcatalysts.com ([10.16.100.88]) by lvlmail.unitedcatalysts.com (PMDF V6.0-24 #41777) with ESMTP id <0GHS00FN2XLKKU@lvlmail.unitedcatalysts.com> for chemistry@ccl.net; Thu, 09 Aug 2001 09:03:20 -0400 (EDT) Received: by lvlxch01.unitedcatalysts.com with Internet Mail Service (5.5.2650.21) id ; Thu, 09 Aug 2001 09:08:26 -0400 Content-return: allowed Date: Thu, 09 Aug 2001 09:08:24 -0400 From: "Shobe, Dave" Subject: RE: transition state To: "'CCL'" Message-id: <157A51F55AAAD3119CD70008C7B1629DDAB0F9@lvlxch01.unitedcatalysts.com> MIME-version: 1.0 X-Mailer: Internet Mail Service (5.5.2650.21) Content-type: text/plain; charset=iso-8859-1 This is a good response. A few things I would add (E. Lewars may or may not agree): 1. I would recommend calcHFFC even for HF level calculations (calcFC is equivalent in this case). Alternatively, if you preoptimized at a lower level like HF/3-21G, you can do a freq job at that level and then use readFC instead. Finding a TS is so much easier with an accurate Hessian (and even then it can be difficult!) 2. When a starting geometry for a TS is unavailable, the algorithms mentioned are requested using Opt=QST2 and opt=LST. You can also do a "relaxed scan" of a reaction coordinate, provided you can approximate it with a single geometric parameter, using opt=modredundant. I will give further instructions if needed--the instructions in the manual are kind of overwhelming since there are so many variations in the use of modredundant. --David Shobe -----Original Message----- From: elewars [mailto:elewars@trentu.ca] Sent: Wednesday, August 08, 2001 5:21 PM To: chemistry@ccl.net Subject: CCL:transition state re txyoung, TS calcs: First try to get a reasonable input geometry. I usually guess it (I find this almost always works), build it with an interacive molecule builder like GaussView or Spartan's builder, then do an AM1 or HF/3-21G TS optimization, then submit the AM1 or 3-21G TS structure to your ab initio or DFT calc. With Gaussian I usually use Opt(TS, CalcHFFC, NoEigenTest, MaxCycle=99) The CalcHHFC (calc Hartree-Fock Force constants) is to get a good Hessian if you are doing a correlated-level calc (e.g. MP2); for a HF-level calc omit it. NoEigenTest tells Gaussian not to check the eigenvalues of the Hessian as it goes along; Gaussian TS calcs always seem to fail if you omit NoEigenTest (at least G94 did). MaxCycle=99 is so a difficult job won't fail just because it ran out of optimization cycles. It may be helpful to get an AM1 or HF/3-21G TS using some other program, then feed the Cartesians into Gaussian. If you have problems guesssing a starting geom, Gaussian has algorithms for getting one from the reactant and product structures--look in help under optimization options. E. Lewars ================= tianxiao young wrote: > Dear CCLers, > > Do you know the best way to calculate the optimization transtion state using > Gassian98? Your kind help is most appreciated. > > txyoung > > _________________________________________________________________ > Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp > -= 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 From chemistry-request@server.ccl.net Thu Aug 9 09:47:10 2001 Received: from electra.cc.umanitoba.ca (root@[130.179.16.23]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79DlAv24369 for ; Thu, 9 Aug 2001 09:47:10 -0400 Received: from mira.cc.umanitoba.ca (hultin@mira.cc.umanitoba.ca [130.179.16.8]) by electra.cc.umanitoba.ca (8.9.0/8.9.0) with ESMTP id IAA09375 for ; Thu, 9 Aug 2001 08:47:09 -0500 (CDT) From: Phil Hultin Received: (from hultin@localhost) by mira.cc.umanitoba.ca (8.9.0/8.9.0) id IAA10855 for chemistry@ccl.net; Thu, 9 Aug 2001 08:47:08 -0500 (CDT) Message-Id: <200108091347.IAA10855@mira.cc.umanitoba.ca> Subject: Transition State To: chemistry@ccl.net Date: Thu, 9 Aug 2001 08:47:08 -0500 (CDT) MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit We prefer to do TS optimizations in G98 using either the QST2 or QST3 approaches, rather than the "classical" OPT=TS approach suggested by another CCL'er. You can read about the procedure for submitting these jobs in the Gaussian manual, and there is discussion in the "blue book" by Foresman and Frisch that comes with Gaussian. The problem with OPT=TS is that your guessed structure has to be VERY good or you will fall off of the saddle point. With very simple structures it isn't hard to set up a reasonable guess, but if you are probing more complex systems it can be a problem. The QST2 approach asks for optimized geometries for the starting and ending structures, and then it interpolates between them to obtain a guess. The QST3 approach asks for both starting and ending geometries plus a guess at the TS geometry. Both of these are much more robust than OPT=TS. Do not forget to run a frequency calculation on the geometry that results > from any TS search to check that it displays one and only one imaginary frequency. Also, you must check that the atomic displacments corresponding to this frequency actually do lie along the reaction coordinate you are interested in. You can do this by inspection of the frequency output, but I find it easier to use a program such as MOLEKEL that can animate frequencies. Dr. Philip Hultin Associate Professor Dept. of Chemistry University of Manitoba Winnipeg, MB Canada hultin@cc.umanitoba.ca From chemistry-request@server.ccl.net Thu Aug 9 05:26:16 2001 Received: from hotmail.com ([64.4.21.90]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f799QFv03762 for ; Thu, 9 Aug 2001 05:26:15 -0400 Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC; Thu, 9 Aug 2001 02:26:09 -0700 Received: from 193.227.50.11 by lw14fd.law14.hotmail.msn.com with HTTP; Thu, 09 Aug 2001 09:26:09 GMT X-Originating-IP: [193.227.50.11] From: "tarek elgogary" To: chemistry@ccl.net Subject: Summary; QSAR-Question Date: Thu, 09 Aug 2001 12:26:09 +0300 Mime-Version: 1.0 Content-Type: text/html Message-ID: X-OriginalArrivalTime: 09 Aug 2001 09:26:09.0454 (UTC) FILETIME=[52DE14E0:01C120B5]
Dear CCLers,
thanks a lot for all who reply to my question:
Dear Netteres; QSARER's
I have calculated lots of descriptors for a training set. When trying to
find out the correlation between each descriptor and the experimental
activity (toxicity), one can not find any reasnable correlation for any
descriptor ( always, R^2<0.01
 
Victor Lobanov wrote:
Toxicity is known for being extremely difficult to predict. Typically there multiple different
mechanisms of toxicity and one has to have a correct data set where all compounds are
causing toxicity via the same mechanism or try multiple descriptor models. Exceptions
are cases where toxcity is connected to the transport function, e.g. solubility. In this
case toxic compounds can be differentiated from non toxic by simply predicting their
solubility or absorbtion. Also it is important to known degree of error in the experimental
data, this could be quite high. And finally, there could be just a human error, where one
tries to correlate unrelated things, e.g. experimental data order is different from the descriptor
data order. This could lead to a random correlation, which would be of about the value
you are getting.
Regards,
Victor
Yvonne Martin wrote:
You also need to look for non-linear relationships. A perfect parabola would
have an R2 of 0 from a linear fit.
How did you treat the biological data, e.g. did you tranform it into free
energy-releate units? Usually this doesn't help much, but it couldmake a
difference.
Lastly, simple univariate correlations might not be sufficient. Did you try
equations with more than one term?
Dave.Winkler wrote:
It depends on how you are mapping your descriptors to toxicity. A neural net is a very good general mapping method. If you add Bayesian regularization you will get a very robust QSAR tool (I've attached one of our tox papers showing this). I suggest you use descriptors based on topology eg Randic and Kier and Hall descriptors. Simple descriptors based on the number of atoms of each type (eg number of primary, secondary, tertiary, quaternary carbons, amine nitrogens etc) often work surprisingly well too. Descriptors based on the number of hydrogen donors, acceptors, hydrophobes, anions and cations often work well too. You should use some kind of non-linear mapping method (at least quadratic least squares if not neural nets).
Good luck.
Stephen Molnar wrote:

If a few more details were forthcoming, help might be available.

Best Regards

Tarek 


Get your FREE download of MSN Explorer at http://explorer.msn.com
From chemistry-request@server.ccl.net Thu Aug 9 08:15:41 2001 Received: from oxmail.ox.ac.uk ([163.1.2.1]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79CFev19224 for ; Thu, 9 Aug 2001 08:15:41 -0400 Received: from nutmeg.biop.ox.ac.uk ([163.1.16.131]) by oxmail.ox.ac.uk with smtp (Exim 3.12 #1) id 15Uoiy-0004cg-00 for chemistry@ccl.net; Thu, 09 Aug 2001 13:15:40 +0100 Received: from t4-pc2.biop.ox.ac.uk by fennel.biop.ox.ac.uk (MX V4.2 AXP) with SMTP; Thu, 09 Aug 2001 13:13:29 +0100 Date: Thu, 9 Aug 2001 13:13:52 +0100 (BST) From: Guy Coates To: chemistry@ccl.net Subject: SMP Gaussian 98 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi, I am trying to compile an SMP aware version of gaussian98 for use on a 2* PIII system running Linux 2.2.19. I have been able to compile a single CPU version g98 using the portland group F77 compiler. Does anyone know how to compile an SMP version? The original build instruction that accompanied the g98 source code were lost long ago and I can't find a copy on the guassian website. My naive attempt at compiling using the '-mp' compiler flag caused the compilation to abort during the linking stage due to a host of unresolved symbols. As a workaround, I have been able to compile a single CPU version with the multi-threaded version of the ATLAS Blas library. This does take advantage of a second CPU some of the time. Is this as close as I can get to a parallel version? Thanks, Guy Coates ---------------------------------------------------------------- --A fool and his money are soon venture capital. E-mail: guy@biop.ox.ac.uk Tel : +44 (0)1865 275390 (W) +44 (0)7801 710224 (M) Mail : Laboratory of Molecular Biophysics, University of Oxford, Rex Richards Building, South Parks Road, Oxford, OX1 3QU From chemistry-request@server.ccl.net Wed Aug 8 22:51:46 2001 Received: from ns.cwp.co.kr ([210.109.76.1]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f792phv23492 for ; Wed, 8 Aug 2001 22:51:43 -0400 Received: from ns ([172.24.15.92]) by ns.cwp.co.kr (8.9.1a-H1/8.9.1) with SMTP id LAA02902 for ; Thu, 9 Aug 2001 11:50:37 +0900 (KST) Message-ID: <002001c1207e$6ef4f540$5c0f18ac@cwp.co.kr> From: "Sei-Hwan Kim" To: Subject: I'm working with InsightII. Date: Thu, 9 Aug 2001 11:52:44 +0900 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_001D_01C120C9.CD367CA0" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4522.1200 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4522.1200 This is a multi-part message in MIME format. ------=_NextPart_000_001D_01C120C9.CD367CA0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: base64 SGVsbG9+DQoNCkknbSB3b3JraW5nIHdpdGggSW5zaWdodElJLiANCg0KQW55Ym9keSBrbm93cyBo b3cgY2FuIEkgZGlzdGluZ3Vpc2ggdGhlIHN0YWJpbGl0eSBvZiBib3VuZCBsaWdhbmRzLg0KDQpS ZWNlbnRseSwgTXkgZG9ja2luZyByZXN1bHRzIGlzIC0zNDA3IGtjYWwgYW5kIC0zMzgwIGtjYWwg KHRvdGFsIGVuZXJneSkuIA0KDQpDYW4gSSBzcGVhayBtb3JlIHN0YWJsZSDigJMzNDA3a2NhbCB0 aGVuIDMzODBrY2FsPw0KDQpTZWFuLiANCg0Kc2VpaGtpbUBjd3AuY28ua3INCg== ------=_NextPart_000_001D_01C120C9.CD367CA0 Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: base64 PCFET0NUWVBFIEhUTUwgUFVCTElDICItLy9XM0MvL0RURCBIVE1MIDQuMCBUcmFuc2l0aW9uYWwv L0VOIj4NCjxIVE1MPjxIRUFEPg0KPE1FVEEgaHR0cC1lcXVpdj1Db250ZW50LVR5cGUgY29udGVu dD0idGV4dC9odG1sOyBjaGFyc2V0PXV0Zi04Ij4NCjxNRVRBIGNvbnRlbnQ9Ik1TSFRNTCA1LjUw LjQ1MjIuMTgwMCIgbmFtZT1HRU5FUkFUT1I+DQo8U1RZTEU+PC9TVFlMRT4NCjwvSEVBRD4NCjxC T0RZIGJnQ29sb3I9I2ZmZmZmZj4NCjxESVY+PEZPTlQgc2l6ZT0yPg0KPFAgY2xhc3M9TXNvTm9y bWFsIA0Kc3R5bGU9IldPUkQtQlJFQUs6IGtlZXAtYWxsOyBURVhULUFVVE9TUEFDRTogaWRlb2dy YXBoLW51bWVyaWM7IFRFWFQtQUxJR046IGxlZnQ7IG1zby1wYWdpbmF0aW9uOiB3aWRvdy1vcnBo YW4iIA0KYWxpZ249bGVmdD48U1BBTiBsYW5nPUVOLVVTPjxGT05UIGZhY2U9IlRpbWVzIE5ldyBS b21hbiIgDQpzaXplPTM+SGVsbG9+PC9GT05UPjwvU1BBTj48L1A+DQo8UCBjbGFzcz1Nc29Ob3Jt YWwgDQpzdHlsZT0iV09SRC1CUkVBSzoga2VlcC1hbGw7IFRFWFQtQVVUT1NQQUNFOiBpZGVvZ3Jh cGgtbnVtZXJpYzsgVEVYVC1BTElHTjogbGVmdDsgbXNvLXBhZ2luYXRpb246IHdpZG93LW9ycGhh biIgDQphbGlnbj1sZWZ0PjxGT05UIHNpemU9Mz48Rk9OVCBmYWNlPSJUaW1lcyBOZXcgUm9tYW4i PjxTUEFOIGxhbmc9RU4tVVM+SSdtIA0Kd29ya2luZyB3aXRoIEluc2lnaHRJSS4gPC9TUEFOPjxT UEFOIGxhbmc9RU4tVVMgDQpzdHlsZT0iRk9OVC1TSVpFOiAxMnB0OyBtc28taGFuc2ktZm9udC1m YW1pbHk6IOuwlO2DlTsgbXNvLWZvbnQta2VybmluZzogMHB0Ij48P3htbDpuYW1lc3BhY2UgDQpw cmVmaXggPSBvIG5zID0gInVybjpzY2hlbWFzLW1pY3Jvc29mdC1jb206b2ZmaWNlOm9mZmljZSIg DQovPjxvOnA+PC9vOnA+PC9TUEFOPjwvRk9OVD48L0ZPTlQ+PC9QPg0KPFAgY2xhc3M9TXNvTm9y bWFsPjxGT05UIHNpemU9Mz48Rk9OVCBmYWNlPSJUaW1lcyBOZXcgUm9tYW4iPjxTUEFOIA0KbGFu Zz1FTi1VUz5Bbnlib2R5IGtub3dzIGhvdyBjYW4gSSBkaXN0aW5ndWlzaCB0aGUgc3RhYmlsaXR5 IG9mIGJvdW5kIA0KbGlnYW5kcy48L1NQQU4+PFNQQU4gbGFuZz1FTi1VUyANCnN0eWxlPSJtc28t YmlkaS1mb250LXNpemU6IDEwLjBwdCI+PG86cD48L286cD48L1NQQU4+PC9GT05UPjwvRk9OVD48 L1A+DQo8UCBjbGFzcz1Nc29Ob3JtYWw+PFNQQU4gbGFuZz1FTi1VUz48Rk9OVCBmYWNlPSJUaW1l cyBOZXcgUm9tYW4iIA0Kc2l6ZT0zPlJlY2VudGx5LCBNeSBkb2NraW5nIHJlc3VsdHMgaXMgLTM0 MDcga2NhbCBhbmQgLTMzODAga2NhbCAodG90YWwgZW5lcmd5KS4gDQo8L0ZPTlQ+PC9TUEFOPjwv UD4NCjxQIGNsYXNzPU1zb05vcm1hbD48Rk9OVCBzaXplPTM+PEZPTlQgZmFjZT0iVGltZXMgTmV3 IFJvbWFuIj48U1BBTiANCmxhbmc9RU4tVVM+Q2FuIEkgc3BlYWsgbW9yZSBzdGFibGUgPC9TUEFO PjxTUEFOIGxhbmc9RU4tVVMgDQpzdHlsZT0iRk9OVC1GQU1JTFk6ICdUaW1lcyBOZXcgUm9tYW4n OyBtc28tYXNjaWktZm9udC1mYW1pbHk6IOuwlO2DlSI+4oCTPC9TUEFOPjxTUEFOIA0KbGFuZz1F Ti1VUz4zNDA3a2NhbCB0aGVuIDMzODBrY2FsPzwvU1BBTj48L0ZPTlQ+PC9GT05UPjwvUD48L0ZP TlQ+PEZPTlQgDQpmYWNlPSJUaW1lcyBOZXcgUm9tYW4iPlNlYW4uIDwvRk9OVD48L0RJVj4NCjxE SVY+PEZPTlQgZmFjZT0iVGltZXMgTmV3IFJvbWFuIj48L0ZPTlQ+Jm5ic3A7PC9ESVY+DQo8RElW PjxBIGhyZWY9Im1haWx0bzpzZWloa2ltQGN3cC5jby5rciI+PEZPTlQgDQpmYWNlPSJUaW1lcyBO ZXcgUm9tYW4iPnNlaWhraW1AY3dwLmNvLmtyPC9GT05UPjwvQT48L0RJVj48L0JPRFk+PC9IVE1M Pg0K ------=_NextPart_000_001D_01C120C9.CD367CA0-- From chemistry-request@server.ccl.net Thu Aug 9 11:28:44 2001 Received: from changeofhabit.mr.itd.umich.edu ([141.211.144.17]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79FSiv27307 for ; Thu, 9 Aug 2001 11:28:44 -0400 Received: from chemraeker1 (chemraeker1.Chem.LSA.UMich.Edu [141.211.69.17]) by changeofhabit.mr.itd.umich.edu (8.9.3/3.2r) with SMTP id LAA16226 for ; Thu, 9 Aug 2001 11:28:43 -0400 (EDT) Content-Type: text/plain; charset="iso-8859-1" From: Todd Raeker To: chemistry@ccl.net Subject: CCL: Motherboards and Linux question Date: Thu, 9 Aug 2001 11:22:50 -0400 X-Mailer: KMail [version 1.2] MIME-Version: 1.0 Message-Id: <01080911225000.02016@chemraeker1> Content-Transfer-Encoding: 8bit Hi All, I will be building a 16 node Linux cluster for ab initio calculations and MD simulations from scratch and have narrowed down my choice of micro ATX motherboards to two options. They are the ASUS A7S-VM and the Shuttle MS21N. Does anyone have any experience with either of these? I have no experience with either board and other people I have talked to only the ASUS ATX boards. Thanks Todd. -- Todd Raeker Department of Chemistry University of Michigan (734) 647-2867 Phone (734) 615-6950 FAX From chemistry-request@server.ccl.net Thu Aug 9 11:33:32 2001 Received: from physics.lsu.edu ([130.39.182.96]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79FXWv27397 for ; Thu, 9 Aug 2001 11:33:32 -0400 Received: from localhost (vani@localhost) by physics.lsu.edu (8.9.2/8.9.2) with ESMTP id KAA11019 for ; Thu, 9 Aug 2001 10:32:20 -0500 (CDT) Date: Thu, 9 Aug 2001 10:32:20 -0500 (CDT) From: Vemparala Satyavani To: chemistry@ccl.net Subject: calculation of stress with SHAKE/RATTLE Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi, how does one calculate the stresses when constraint dynamics is used? or is there a way to seperate the constraint forces as two equal and opposite forces so that the stress calculation is simple?(am using the RATTLE routine provided in Allen & Tildesley book) Thanks vani From chemistry-request@server.ccl.net Thu Aug 9 14:05:56 2001 Received: from post1.inre.asu.edu ([129.219.110.72]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79I5uv30254 for ; Thu, 9 Aug 2001 14:05:56 -0400 Received: from conversion.post1.inre.asu.edu by asu.edu (PMDF V6.0-24 #47346) id <0GHT00501BJXLG@asu.edu> for chemistry@ccl.net; Thu, 09 Aug 2001 11:04:45 -0700 (MST) Received: from webmail2.asu.edu (webmail2.asu.edu [129.219.10.28]) by asu.edu (PMDF V6.0-24 #47346) with ESMTP id <0GHT00581BJX98@asu.edu>; Thu, 09 Aug 2001 11:04:45 -0700 (MST) Received: (from nobody@localhost) by webmail2.asu.edu (8.9.3/8.9.3) id LAA25948; Thu, 09 Aug 2001 11:04:39 -0700 Date: Thu, 09 Aug 2001 11:04:39 -0700 (MST) From: Benjamin.Moritz@asu.edu Subject: TZVP Basis Set X-Originating-IP: 149.169.38.28 To: chemistry@ccl.net, help@gaussian.com Message-id: <997380279.3b72d0b7aa40d@webmail2.asu.edu> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit User-Agent: IMP/PHP IMAP webmail program 2.2.3 Dear All, I was wondering if anyone can point me in the direction on where I can find literature about the TZVP basis set. I need info on what type of basis set it is. Thank you. ******************************************************************************** Benjamin J. Moritz Work: 480-965-8509 College of Engineering and Applied Sciences Mailto:bmoritz@asu.edu Dept. of Chemical & Materials Engineering PO Box 876006 Tempe, AZ 85287-6006 ******************************************************************************** Powered by Compu-Global-Hyper-Meganet From chemistry-request@server.ccl.net Thu Aug 9 14:27:13 2001 Received: from post2.inre.asu.edu ([129.219.110.73]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79IRDv30685 for ; Thu, 9 Aug 2001 14:27:13 -0400 Received: from conversion.post2.inre.asu.edu by asu.edu (PMDF V6.0-24 #47347) id <0GHT00601CLBMD@asu.edu> for chemistry@ccl.net; Thu, 09 Aug 2001 11:27:11 -0700 (MST) Received: from webmail2.asu.edu (webmail2.asu.edu [129.219.10.28]) by asu.edu (PMDF V6.0-24 #47347) with ESMTP id <0GHT0064VCLBBW@asu.edu>; Thu, 09 Aug 2001 11:27:11 -0700 (MST) Received: (from nobody@localhost) by webmail2.asu.edu (8.9.3/8.9.3) id LAA28640; Thu, 09 Aug 2001 11:27:12 -0700 Date: Thu, 09 Aug 2001 11:27:12 -0700 (MST) From: Benjamin.Moritz@asu.edu Subject: One more on TZVP X-Originating-IP: 149.169.38.28 To: chemistry@ccl.net, help@gaussian.com Message-id: <997381632.3b72d6004fe73@webmail2.asu.edu> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1 Content-transfer-encoding: 8bit User-Agent: IMP/PHP IMAP webmail program 2.2.3 I forgot to ask one other question. I have been doing calculations on NOx species adsorbed onto copper sites using the TZVP basis set. Was this the most efficient basis set to use with copper? I've been getting good results with it, but am skeptical about using TZVP. Was I right in choosing TZVP for copper? Any feedback will be greatly appreciated. Thank you. ******************************************************************************** Benjamin J. Moritz Work: 480-965-8509 College of Engineering and Applied Sciences Mailto:bmoritz@asu.edu Dept. of Chemical & Materials Engineering PO Box 876006 Tempe, AZ 85287-6006 ******************************************************************************** Powered by Compu-Global-Hyper-Meganet From chemistry-request@server.ccl.net Thu Aug 9 15:16:29 2001 Received: from mail.bancorp.ru (root@[195.239.131.24]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79JGSv31926 for ; Thu, 9 Aug 2001 15:16:28 -0400 Received: from d-pm4-120.bancorp.ru (d-pm4-120.bancorp.ru [195.239.151.120]) by mail.bancorp.ru (8.11.3/8.11.3/relay.domain) with ESMTP id f79JGPI32736 for ; Thu, 9 Aug 2001 23:16:26 +0400 Date: Thu, 9 Aug 2001 23:13:05 +0400 From: Gregory Shamov X-Mailer: The Bat! (v1.45) Educational Reply-To: Gregory Shamov X-Priority: 3 (Normal) Message-ID: <804576578.20010809231305@bancorp.ru> X-Confirm-Reading-To: Disposition-Notification-To: gas5x@bancorp.ru To: chemistry@ccl.net Subject: Bandwidth or latency? Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Hello All, We have the Linux cluster, on Fast Ethernet. The scalability of FastEthernet clusters is not too good. We have no enough funding to switch to Myrinet or SCI. There are technologies able to decrease Ethernet network latency by factor two or three, such as M-VIA and GAMMA. And there is a way to increase network bandwidth by using more than one Ethernet NICs per cluster node (network trunking). The question is: what is more important network parameter for commonly used parallel quantum-chemistry programs like a GAMESS, NWChem, etc - bandwidth or latency? Changes in what parameter will affect the total performance stronger? Thank you in advance. -- Best regards, Gregory Shamov mailto:gas5x@bancorp.ru From chemistry-request@server.ccl.net Thu Aug 9 16:31:43 2001 Received: from ivory.trentu.ca ([192.75.12.103]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79KVhv02907 for ; Thu, 9 Aug 2001 16:31:43 -0400 Received: from trentu.ca ([204.225.13.50]) by trentu.ca (PMDF V5.2-32 #37862) with ESMTP id <01K6XQADX60E001BQ2@trentu.ca> for chemistry@ccl.net; Thu, 9 Aug 2001 16:31:14 EDT Date: Thu, 09 Aug 2001 16:37:20 -0400 From: elewars Subject: Re: CCL:transition state To: "Shobe, Dave" , chemistry@ccl.net Message-id: <3B72F47F.7F882092@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 References: <157A51F55AAAD3119CD70008C7B1629DDAB0F9@lvlxch01.unitedcatalysts.com> Yes, CalcFC may well help for a TS opt (but probably not for a PES-minimum opt), as it is likely to give a better Hessian than the default molecular mechanics-type Hessian. A PES scan could be helpful in difficult cases EL ====== "Shobe, Dave" wrote: > This is a good response. A few things I would add (E. Lewars may or may not > agree): > > 1. I would recommend calcHFFC even for HF level calculations (calcFC is > equivalent in this case). Alternatively, if you preoptimized at a lower > level like HF/3-21G, you can do a freq job at that level and then use readFC > instead. Finding a TS is so much easier with an accurate Hessian (and even > then it can be difficult!) > > 2. When a starting geometry for a TS is unavailable, the algorithms > mentioned are requested using Opt=QST2 and opt=LST. You can also do a > "relaxed scan" of a reaction coordinate, provided you can approximate it > with a single geometric parameter, using opt=modredundant. I will give > further instructions if needed--the instructions in the manual are kind of > overwhelming since there are so many variations in the use of modredundant. > > --David Shobe > > -----Original Message----- > From: elewars [mailto:elewars@trentu.ca] > Sent: Wednesday, August 08, 2001 5:21 PM > To: chemistry@ccl.net > Subject: CCL:transition state > > re txyoung, TS calcs: > > First try to get a reasonable input geometry. I usually guess it (I find > this > almost always works), build it with an interacive molecule builder like > GaussView or Spartan's builder, then do an AM1 or HF/3-21G TS optimization, > then > submit the AM1 or 3-21G TS structure to your ab initio or DFT calc. > With Gaussian I usually use Opt(TS, CalcHFFC, NoEigenTest, MaxCycle=99) > > The CalcHHFC (calc Hartree-Fock Force constants) is to get a good Hessian if > you > are doing a correlated-level calc (e.g. MP2); for a HF-level calc omit it. > NoEigenTest tells Gaussian not to check the eigenvalues of the Hessian as it > goes along; Gaussian TS calcs always seem to fail if you omit NoEigenTest > (at > least G94 did). MaxCycle=99 is so a difficult job won't fail just because it > ran > out of optimization cycles. > > It may be helpful to get an AM1 or HF/3-21G TS using some other program, > then > feed the Cartesians into Gaussian. > > If you have problems guesssing a starting geom, Gaussian has algorithms for > getting one from the reactant and product structures--look in help under > optimization options. > > E. Lewars > > ================= > tianxiao young wrote: > > > Dear CCLers, > > > > Do you know the best way to calculate the optimization transtion state > using > > Gassian98? Your kind help is most appreciated. > > > > txyoung > > > > _________________________________________________________________ > > Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp > > > > -= 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 From chemistry-request@server.ccl.net Thu Aug 9 16:41:35 2001 Received: from ivory.trentu.ca ([192.75.12.103]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79KfZv03149 for ; Thu, 9 Aug 2001 16:41:35 -0400 Received: from trentu.ca ([204.225.13.50]) by trentu.ca (PMDF V5.2-32 #37862) with ESMTP id <01K6XQMROWFC0017CI@trentu.ca> for chemistry@ccl.net; Thu, 9 Aug 2001 16:41:13 EDT Date: Thu, 09 Aug 2001 16:47:19 -0400 From: elewars Subject: Re: CCL:Transition State To: Phil Hultin , chemistry@ccl.net Message-id: <3B72F6D6.96F40FDF@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 References: <200108091347.IAA10855@mira.cc.umanitoba.ca> Hello, I usually submit a guess to Spartan and request an AM1 TS opt; this very rarely fails: Spartan's TS algorithm does not seem to need a *very* good guess. I then submit the AM1 TS to a HF/3-21G TS opt., usually with Spartan The resulting structure almost always works for a higher-level TS opt with G98 (I suspect the AM1 structure would be about as good for a starting geometry). Of course there may be families of TS's that do not lend themselves to initial guessing and require the automated procedure. EL ==== Phil Hultin wrote: > We prefer to do TS optimizations in G98 using either the QST2 or QST3 > approaches, rather than the "classical" OPT=TS approach suggested by > another CCL'er. > > You can read about the procedure for submitting these jobs in the Gaussian > manual, and there is discussion in the "blue book" by Foresman and Frisch > that comes with Gaussian. > > The problem with OPT=TS is that your guessed structure has to be VERY good > or you will fall off of the saddle point. With very simple structures it > isn't hard to set up a reasonable guess, but if you are probing more > complex systems it can be a problem. The QST2 approach asks for optimized > geometries for the starting and ending structures, and then it > interpolates between them to obtain a guess. The QST3 approach asks for > both starting and ending geometries plus a guess at the TS geometry. Both > of these are much more robust than OPT=TS. > > Do not forget to run a frequency calculation on the geometry that results > > from any TS search to check that it displays one and only one imaginary > frequency. Also, you must check that the atomic displacments > corresponding to this frequency actually do lie along the reaction > coordinate you are interested in. You can do this by inspection of the > frequency output, but I find it easier to use a program such as MOLEKEL > that can animate frequencies. > > Dr. Philip Hultin > Associate Professor > Dept. of Chemistry > University of Manitoba > Winnipeg, MB Canada > hultin@cc.umanitoba.ca > > -= 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 From chemistry-request@server.ccl.net Thu Aug 9 12:25:54 2001 Received: from mail2.zrz.tu-berlin.de ([130.149.4.14]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79GPsv28531 for ; Thu, 9 Aug 2001 12:25:54 -0400 Received: from mailszrz.zrz.tu-berlin.de ([130.149.4.11]) by mail2.zrz.tu-berlin.de with esmtp (exim-3.31) id 15Usd2-0003jk-00; Thu, 09 Aug 2001 18:25:48 +0200 Received: from natrium.chem.tu-berlin.de ([130.149.169.143]) by mailszrz.zrz.tu-berlin.de with esmtp (exim-3.31) id 15Usd1-0000Qh-00; Thu, 09 Aug 2001 18:25:47 +0200 Mime-Version: 1.0 X-Sender: chwu0531@mailszrz.zrz.tu-berlin.de Message-Id: In-Reply-To: <200108091347.IAA10855@mira.cc.umanitoba.ca> References: <200108091347.IAA10855@mira.cc.umanitoba.ca> Date: Thu, 9 Aug 2001 18:25:47 +0200 To: Phil Hultin From: Christoph van =?iso-8859-1?Q?W=FCllen?= Subject: Re: CCL:Transition State Cc: Chemistry@ccl.net Content-Type: text/plain; charset="iso-8859-1" ; format="flowed" Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by server.ccl.net id f79GPsv28532 >isn't hard to set up a reasonable guess, but if you are probing more >complex systems it can be a problem. The QST2 approach asks for optimized >geometries for the starting and ending structures, and then it >interpolates between them to obtain a guess. The QST3 approach asks for If you form a bond, the "optimized geometry" of the educt may be very far away. In this case a constrained optimization with the "bond" distance fixed to value of, say, 3 Angstrom for a C-C-bond to be formed, should be done. -- +---------------------------------+-------------------------------------+ | Prof. Christoph van Wüllen | Tele-Phone (+49) (0)30 314 27870 | | Technische Universität Sekr. C3 | Tele-Fax (+49) (0)30 314 23727 | | Straße des 17. Juni 135 | eMail | | D-10623 Berlin, Germany | Christoph.vanWullen@TU-Berlin.De | +---------------------------------+-------------------------------------+ From chemistry-request@server.ccl.net Thu Aug 9 17:27:31 2001 Received: from thorin.eas.asu.edu ([149.169.26.91]) by server.ccl.net (8.11.0/8.11.0) with SMTP id f79LRUv04077 for ; Thu, 9 Aug 2001 17:27:30 -0400 Received: by thorin.eas.asu.edu id AA11376; Thu, 9 Aug 2001 14:27:29 -0700 Date: Thu, 9 Aug 2001 14:27:29 -0700 (MST) From: Donghai Sun To: Chemistry Subject: ADF question Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII hi, how to visualize the frequency by Molekel from the output file from ADF? Now i am using ADF through NCSA and get the standard output of it. Since i tried Gaussian and could visualize the frequency from it but failed in ADF. Does anybody may give me a hand on this? Also, when i use the Linear transit method to get the initial guess for geometry optimization calculation, sometimes i get decent result from the linear transit method but sometimes i fail. Does somebody may do me a favor on the way to make reasonable guess for cluster, like ZCuNO2H2NH...? Thanks you very much! donghai Donghai Sun Department of Bio, Chemical and Materials Engineering Arizona State University Tel: 480-965-8509 (O) (FAX) 480-446-8666 (H) Email: sun@thorin.eas.asu.edu From chemistry-request@server.ccl.net Thu Aug 9 12:10:16 2001 Received: from server.mldnet.com (root@[212.56.192.20]) by server.ccl.net (8.11.0/8.11.0) with ESMTP id f79GACv28194 for ; Thu, 9 Aug 2001 12:10:13 -0400 Received: from ppp-212.56.193.108.mldnet.com (ppp-212.56.193.108.mldnet.com [212.56.193.108]) by server.mldnet.com (8.9.3/8.9.3) with ESMTP id TAA10510 for ; Thu, 9 Aug 2001 19:10:06 +0300 Date: Thu, 9 Aug 2001 19:15:12 +0300 From: Mike Peleah X-Mailer: The Bat! (v1.53d) Reply-To: Mike Peleah X-Priority: 3 (Normal) Message-ID: <10811901603.20010809191512@yahoo.com> To: chemistry@ccl.net Subject: INDO parameters for Ag MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dear chemistry, I have a coumpound containing Ag which I need to calculated by INDO1/S mathod. Could anybody tell me if parameters for Ag in ZINDO or INDO1/S are available? Best regards, Mike mailto:MikePeleah@yahoo.com