From chemistry-request@www.ccl.net Tue Jan 5 04:45:01 1999 Received: from mail-relay-pune.ernet.in (root@[202.41.82.35]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id EAA08095 Tue, 5 Jan 1999 04:44:58 -0500 (EST) Received: from iucaa (meghnad.iucaa.ernet.in [202.141.155.1]) by mail-relay-pune.ernet.in (8.9.1/8.9.1) with SMTP id PAA02204 for ; Tue, 5 Jan 1999 15:17:09 +0530 Received: from iucaa by iucaa (5.65v3.2/SMI-4.1) id AA09584; Tue, 5 Jan 1999 15:11:45 +0500 Received: by (5.x/SMI-SVR4) id AA01788; Tue, 5 Jan 1999 15:06:37 -0500 Date: Wed, 6 Jan 1999 01:06:36 +0500 (GMT-5) From: Sangeeta Sawant To: Computational Chemistry List Subject: looking for search utility using phi-psi data as query.... Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hello everybody, I am looking for a free/shareware utility which: 1. accepts a query in the form of phi-psi dihedral angles of amino acid residues (sequence may or may not be specified), 2. can search a set of 3D structures of proteins for the given query, 3. can find the matches within a specified range of deviation of the dihedral angles and, 4. list the peptides of matching conformations from the structure data. If such a utility is available, I would be interested in using it. Any help would be welcome and appreciated. Thanks and best wishes for the new year! Sangeeta +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ + Mrs. Sangeeta Sawant + + Research Fellow email : sangeeta@bioinfo.ernet.in + + Bioinformatics Centre phone : +91-20-355039 + + University of Pune fax : +91-20-350087 + + Ganeshkhind + + Pune - 411007 + + India. + +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ From chemistry-request@www.ccl.net Tue Jan 5 07:22:21 1999 Received: from hartree.chem.kuleuven.ac.be (hartree.quantchem.kuleuven.ac.be [134.58.49.1]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id HAA08471 Tue, 5 Jan 1999 07:22:19 -0500 (EST) Received: from localhost (steven@localhost) by hartree.chem.kuleuven.ac.be (8.9.0/8.9.0) with ESMTP id NAA21390 for ; Tue, 5 Jan 1999 13:21:49 +0100 Date: Tue, 5 Jan 1999 13:21:49 +0100 (NFT) From: Steven Creve To: Computational Chemistry List Subject: CCL:QSAR Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi, could anyone point me to software packages (both commercial and non-commercial) for performing QSAR analysis. thank you. steven -------------------------------------------------------------------------- Steven Creve steven.creve@chem.kuleuven.ac.be Labo Quantumchemie Celestijnenlaan 200F 3001-HEVERLEE tel: (32) (16) 32 73 93 BELGIUM fax: (32) (16) 32 79 92 From chemistry-request@www.ccl.net Tue Jan 5 10:26:55 1999 Received: from helix.nih.gov (helix.nih.gov [128.231.2.3]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id KAA09739 Tue, 5 Jan 1999 10:26:54 -0500 (EST) Received: (from mn1@localhost) by helix.nih.gov (8.9.1a/8.9.1) id KAA03873; Tue, 5 Jan 1999 10:26:55 -0500 (EST) Date: Tue, 5 Jan 1999 10:26:54 -0500 From: "M. Nicklaus" To: CHEMISTRY@www.ccl.net cc: "M. Nicklaus" Subject: SUMMARY: Parallel G98 on x86/Linux Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII A couple of weeks ago, I posted the enclosed question about running Gaussian 98 in parallel on x86/Linux. Here's the summary of the responses I got, slightly edited for brevity. >From these few replies it appears that no one has been able so far to get G98 to run properly in this setup. Maybe Gaussian, Inc. could comment on this. Doug? A Belated Happy New Year to everyone. Marc ------------------------------------------------------------------------ Marc C. Nicklaus National Institutes of Health E-mail: mn1@helix.nih.gov Bldg 37, Rm 5B29 Phone: (301) 402-3111 37 Convent Dr, MSC 4255 Fax: (301) 496-5839 BETHESDA, MD 20892-4255 USA http://rex.nci.nih.gov/RESEARCH/basic/medchem/mcnbio.htm Laboratory of Medicinal Chemistry, National Cancer Institute ------------------------------------------------------------------------ ####################### QUESTION ####################### According to Gaussian, Inc. (http://www.gaussian.com/parallel.htm), G98 is available for Shared Memory Parallel execution on [Intel] Linux systems. Does anyone have G98 set up like this, and could report some performance results? I am particulary thinking of a Dual-P II 450MHz system, with G98 compiled with the Portland Group compilers, and using the tuned BLAS libraries as recommended by Gaussian, Inc. (see http://www.gaussian.com/g98_req.htm). This should be about the fastest *inexpensive* G98 implementation on x86/Linux at this time. (One can certainly get higher perfomance by using Linda with >2 CPU's, and/or Xeon processors, but either option would increase the cost substantially.) If someone has this setup and could run some benchmarks for me (should be < 1 h), I'd appreciate. ######################## ANSWERS ######################## From: Drake Diedrich If the SMP mode uses standard System V IPC it might be possible to substitute Distributed IPC (http://wallybox.cei.net/dipc) and run Gaussian in parallel on an arbitrarily large cluster, not just an SMP. I'd have to know something about Gaussian's parallelization strategy to even guess whether it would be efficient though. ------------------------------------------------------------- From: victor@cti.csic.es I am trying to compile G98 on such a PentiumII 450MHz dual processor . At present I am not been able to succesfully compile the code with the blas-opt libraries tuned for dual processors ( ldblasppro1.1m_11.98.a). It seems that an unsatisfied reference to pthread_create module aborts the linking procedure. I think that either the libpthread.a library is not searched for or G98 does not support parallel execution in that machine. This last point has been suggested to me by Doug Fox , but I do not understand it well. Do you have any suggestion to solve the problem ? On the other hand I have compiled G98 with and without the single processor blas optimized libraries. I have performed both HF and DFT single point calculations and obtained similar CPU times with both executables(!). The %Nproc keyword seems not to work with these executables. Of course, once I solve the dual processor problem I could perform the benchmarking job you asked for. ------------------------------------------------------------- From: JENSAB G98 does not execute in parallel on Intel/Linux systems in our hands. (See the following email to Dr Dough Fox, Gaussian presumably are working on it) Single processor executions however shows quite satisfactory cpu preformance (though Gaussian must have some 32 bit windows95/98/NT compiler). I am not sure both machines were running 100 MHz bus speeds, which should affect the IO intensive gaussian jobs. To: gaussian.com!fox@lorentzian.com (Doug Fox) Subject: Slow g98 on Linux/PGIf77 unable to %NPROC=2 Date sent: Fri, 30 Oct 1998 15:57:30 +0100 Dear Dr. Doug Fox, Our Linux/PortlandGroup F77 compiled Gaussian98 cannot run in parallel. Using a single processor it is somewhat slower than the corresponding G98W running under windowsNT!!?! Here are the Job cpu times (in sec.) for single processors execution on dual pentiumII 266 MHz machines running the DFT test Job from M. C. Nicklaus, R. W. Williams, B. Bienfait, M . Hodocek, J. Chem. Inf. Comput. Sci. Vol 38(5) p 893-905 (1998): # B3LYP/631G(d) and subsequently # B3LYP/631G(d) GEOM=CHECK GUESS=READ Job cpu / sec. g98W/NT 8074 1975 g98/Linux/PGI 8594 2101 + ~ 6 % g94/Linux/f2c 15300 3845 + ~ 91 % We do not observe any difference in Job cpu between using: BLAS = blas.a and BLAS = /usr/local/g98/lib/blas-opt.a /usr/local/g98/lib/blas-f2c.a lsblaspent1.1d_06.98.a (Pentium) and lsblasppro1.1h_08.98.a (PentiumPro) yield similar results, while ldblasppro1.1h_08.98.a (dual PentiumPro) gives compilation errors. The tests were run with no load on the second cpu, where we have observed a quite severe cpu time penalty when running two identical single processor jobs simultaneously (~ 30 - 40 % increase in Gaussian98 reported "Job cpu" times, which are in close to perfect agreement with "wall clock" cpu time). This is why we are very interested in the performance of the parallel version. Given g98 inability to utilize two processors and the reported performance, why are we bothering with the UNIX version?? Are we the only ones having problems with g98 or are reports pouring in? We are however impressed with the g98W which performs really nicely, and THIS program now does compare favorably with traditional 64 bit high-end UNIX workstations and presents a very attractive choice. Yours Sincerely Jens Abildgaard The following extracts from 3 different NPROC=2 tests with different MEM settings (0.5MW, 4.19MW and 4.2MW) leave us, as you may by now expect, completely in the dark. (the route sections for testb and testc are identical to testa) test a: MEM=0.5MW ******************************************** Gaussian 98: x86-Linux-G98RevA.3 2-Sep-1998 28-Oct-1998 ******************************************** %NPROC=2 Will use up to 2 processors via shared memory. %MEM=500000 ---------------- #P B3PW91/STO-3G ---------------- 1/38=1/1; 2/17=6,18=5/2; 3/11=2,25=1,30=1/1,2,3; 4//1; 5/5=2,32=1,38=4,42=-6/2; 6/7=2,8=2,9=2,10=2,28=1/1; 99/5=1,9=1/99; Leave Link 1 at Wed Oct 28 19:35:38 1998, MaxMem= 500000 cpu: 0.3 (Enter /usr/local/g98/l101.exe) --------------------------------------------------------------------- - Geometri Optimeret i B3PW91/6-311G(d,p) Enol uden intramolekylaeaer hyd rogenbindi --------------------------------------------------------------------- - Symbolic Z-matrix: Charge = 0 Multiplicity = 1 C C 1 R1 C 2 R2 1 V2 C 3 R3 2 V3 1 D3 0 C 4 R4 3 V4 2 D4 0 O 2 R5 1 V5 3 D5 0 S 4 R6 3 V6 2 D6 0 H 6 R7 2 V7 1 D7 0 H 3 R8 2 V8 1 D8 0 H 1 R9 2 V9 3 D9 0 H 1 R9 2 V9 3 -D9 0 H 1 R11 2 V11 3 D11 0 H 5 R12 4 V12 3 D12 0 H 5 R12 4 V12 3 -D12 0 H 5 R14 4 V14 3 D14 0 Variables: R1 1.50489 R2 1.36216 R3 1.43593 R4 1.5135 R5 1.33334 R6 1.65523 R7 0.96216 R8 1.08624 R9 1.09284 R11 1.09402 R12 1.0929 R14 1.09347 V2 120.67031 V3 130.03436 V4 114.89229 V5 117.33213 V6 126.32056 V7 109.49169 V8 113.95904 V9 110.80924 V11 111.7517 V12 109.68088 V14 112.9822 D3 180. D4 180. D5 180. D6 0. D7 0. D8 0. D9 59.77432 D11 180. D12 121.55924 D14 0. Leave Link 101 at Wed Oct 28 19:35:39 1998, MaxMem= 500000 cpu: 0.6 (Enter /usr/local/g98/l202.exe) . . . . . (Enter /usr/local/g98/l502.exe) Warning! Cutoffs for single-point calculations used. IExCor= 408 DFT=T Ex=B+HF Corr=PW91 ScaHFX= 0.2000 ScaDFX= 0.8000 0.7200 1.0000 0.8100 IRadAn= 0 IRanWt= -1 IRanGd= 0 ICorTp=0 Using DIIS extrapolation. Closed shell SCF: Requested convergence on RMS density matrix=1.00D-04 within 64 cycles. Requested convergence on MAX density matrix=1.00D-02. Requested convergence on energy=5.00D-05. Integral symmetry usage will be decided dynamically. IEnd= 11941 IEndB= 11941 NGot= 500000 MDV= 494540 LenX= 494540 Fock matrices will be formed incrementally for 20 cycles. Cycle 1 Pass 1 IDiag 1: Petite list used in FoFDir. MinBra= 0 MaxBra= 1 Meth= 1. IRaf= 0 NMat= 1 IRICut= 1 DoRegI=T DoRafI=F ISym2E= 1 JSym2E=1. PrsmSu: requested number of processors reduced to: 1 ShMem 1 Linda. E=-0.100792091983499D+04 DIIS: error= 3.65D-01 at cycle 1. T= 1004. NK=0 NO(<0.9)= 0 NV(>0.1)= 0 31.00e < EF 0.00e >EF Err=8.7D-11 RMSDP=2.12D-02 MaxDP=2.07D-01 Cycle 2 Pass 1 IDiag 1: RMSU= 2.07D-02 CP: 9.61D-01 PrsmSu: requested number of processors reduced to: 1 ShMem 1 Linda. Newton Raphson failed in SizSh1. Error termination via Lnk1e in /usr/local/g98/l502.exe. Job cpu time: 0 days 0 hours 0 minutes 0.0 seconds. File lengths (MBytes): RWF= 10 Int= 0 D2E= 0 Chk= 1 Scr= 1 Process number 1 failed to complete. Consistency failure #1 in CalDSu. Error termination via Lnk1e in /usr/local/g98/l502.exe. Job cpu time: 0 days 0 hours 0 minutes 15.9 seconds. File lengths (MBytes): RWF= 10 Int= 0 D2E= 0 Chk= 1 Scr= 1 test b:MEM=4.19MW (Enter /usr/local/g98/l502.exe) Warning! Cutoffs for single-point calculations used. IExCor= 408 DFT=T Ex=B+HF Corr=PW91 ScaHFX= 0.2000 ScaDFX= 0.8000 0.7200 1.0000 0.8100 IRadAn= 0 IRanWt= -1 IRanGd= 0 ICorTp=0 Using DIIS extrapolation. Closed shell SCF: Requested convergence on RMS density matrix=1.00D-04 within 64 cycles. Requested convergence on MAX density matrix=1.00D-02. Requested convergence on energy=5.00D-05. Integral symmetry usage will be decided dynamically. Keep R1 and R2 integrals in memory in canonical form, NReq= 1685467. IEnd= 11941 IEndB= 11941 NGot= 4190000 MDV= 2911014 LenX= 2911014 Symmetry not used in FoFDir. MinBra= 0 MaxBra= 1 Meth= 1. IRaf= 0 NMat= 1 IRICut= 1 DoRegI=T DoRafI=F ISym2E= 0 JSym2E=0. Cycle 1 Pass 1 IDiag 1: Newton Raphson failed in SizSh1. Error termination via Lnk1e in /usr/local/g98/l502.exe. Job cpu time: 0 days 0 hours 0 minutes 0.0 seconds. File lengths (MBytes): RWF= 10 Int= 0 D2E= 0 Chk= 1 Scr= 1 Process number 1 failed to complete. Consistency failure #1 in CalDSu. Error termination via Lnk1e in /usr/local/g98/l502.exe. Job cpu time: 0 days 0 hours 0 minutes 17.6 seconds. File lengths (MBytes): RWF= 10 Int= 0 D2E= 0 Chk= 1 Scr= 1 test c: MEM=4.2MW shmget failed.: Invalid argument From chemistry-request@www.ccl.net Tue Jan 5 10:35:17 1999 Received: from ganimedes.overnet.com.ar (ns1.overnet.com.ar [200.10.96.10]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id KAA09849 Tue, 5 Jan 1999 10:35:14 -0500 (EST) Received: from carpincho.overnet.com.ar (carpincho.overnet.com.ar [200.10.96.19]) by ganimedes.overnet.com.ar (8.8.7/8.8.3) with ESMTP id MAA14017 for ; Tue, 5 Jan 1999 12:50:20 -0300 Received: from nppp235.overnet.com.ar (nppp235.overnet.com.ar [200.16.162.248]) by carpincho.overnet.com.ar (8.8.8/8.8.5) with SMTP id MAA16199 for ; Tue, 5 Jan 1999 12:50:18 -0300 Message-Id: <3.0.6.16.19990105123405.123f782c@pop.overnet.com.ar> X-Sender: dfa953uo@pop.overnet.com.ar X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.6 (16) Date: Tue, 05 Jan 1999 12:34:05 To: chemistry@www.ccl.net From: "Dr. Daniel Glossman" Subject: Pop=Bonding calculation Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Dear netters: I have been doing some G94W calculations on several molecules with the POP=Bonding keyword, but I don't know how to do an interpretation of the results and the Gaussian 94 manual is not very informative in this respect. For example, after the title "... Condensed to atoms (all electrons..." I got a symmetric matrix with diagonal elements equal to zero. 1) What do the numbers mean? 2) Where do they come from? 3) Are they related to the so called bond order analysis? 4) Are there any bibliographic references for the calculation of these numbers? After that I got the total atomic charges. These charges are different from those coming from a traditional Mulliken population analysis in that they are not summed up to the charge of the molecule under study. Again, 1) What do these charges mean? 2) What's the meaning of the summation? 3) Where do these numbers come from? 4) which are the bibliographic references? Thanks in advance for a kind consideration of this request. Best regards Dr. Daniel Glossman ***************************************************************************** Universidad Nacional de Lujan Departamento de Ciencias Basicas e-mail: glossman@overnet.com.ar Casilla de Correo 221 phone: (+54) 1 5533797 (6700) Lujan fax: (+54) 1 5539824 Republica Argentina **************************************************************************** ** From chemistry-request@www.ccl.net Tue Jan 5 10:37:21 1999 Received: from bobino.sefmedia.com (www.sefmedia.com [207.164.6.2]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with SMTP id KAA09874 Tue, 5 Jan 1999 10:37:20 -0500 (EST) Received: from ppp-5200-3103.mtl.total.net (ppp-5200-3103.mtl.total.net [207.139.147.17]) by bobino.sefmedia.com (NTMail 3.02.13) with ESMTP id na016809 for ; Tue, 5 Jan 1999 10:37:37 -0500 Received: by zman with Microsoft Mail id <01BE3896.F369CE40@zman>; Tue, 5 Jan 1999 10:34:19 -0500 Message-ID: <01BE3896.F369CE40@zman> From: Bill Hayden To: "'Steven Creve'" , Computational Chemistry List Subject: RE: QSAR Date: Tue, 5 Jan 1999 10:34:18 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from quoted-printable to 8bit by www.ccl.net id KAA09875 Dear Dr. Creve, I noticed you posting on CCL looking for QSAR software. We have a commercial software package, MOE - the Molecular Operating Environment, that contains QSAR applications and an embedded programming language that allows you to quickly add your own descriptors. Below is a brief summary of MOE. MOE - The Molecular Operating Environment Chemical Computing Group Inc. has created a very unique tool for Drug Discovery researchers. MOE's uniqueness is due to SVL - Scientific Vector Language which lays at the heart of the package. SVL is a next generation high-performance data-parallel language. Due to the incorporation of SVL, MOE can be used in any combination of 3 ways: 1. Out-of-the-box: MOE contains a robust set of built-in applications in the areas of High Throughput Screening Data Analysis, Combinatorial Chemistry Library Design, Protein & Homology Modeling, and QSAR as well as traditional modeling and simulation. 2. Customizable functionality: As most of MOE's built-in functionality is coded in SVL and you receive the actual source code for most of the built-in applications , you are able to fully customize MOE ranging from additions to the Graphical User Interface to changes in the specifics of the algorithms. 3. Prototyping: The package includes SVL therefore the user is able to quickly prototype ideas and scientific journal papers at a speed and in a manner not currently available anywhere. In addition, due to SVL, Chemical Computing Group is able to quickly turnout upgrades and new applications to ensure that you have a software package that keeps pace with advances in science. SVL also allows MOE to be a true platform independent package that currently runs native on Sun, SGI, Windows NT, Windows 95, and DEC Alpha NT. Future plans include porting to the parallel platform. This marriage of unique architecture and platform independence makes MOE the ideal corporate deployment vehicle. One that can change and grow as your science changes and grows. To sum up, when purchasing MOE users receive: * All the built-in applications (please see our web page for further details) * Source code for most of the applications. This allows you to change both the way the app looks (GUI) and the way it works (algorithms) * The embedded programming language, SVL, allowing you to create your own applications from ideas or scientific papers * The executables for all the platform we support The license is a floating one whereby you sit the license tokens on a server on the network and check the tokens out as needed on the machine of your choice. I invite you to visit our web site at: http://www.chemcomp.com for more information. Please contact me should you have further questions. hayden@chemcomp.com Thanks, Bill William A. Hayden Vice President Chemical Computing Group Inc. http://www.chemcomp.com hayden@chemcomp.com 514 393 1055 -----Original Message----- From: Steven Creve [SMTP:Steven.Creve@chem.kuleuven.ac.be] Sent: January 5, 1999 7:22 AM To: Computational Chemistry List Subject: CCL:QSAR Hi, could anyone point me to software packages (both commercial and non-commercial) for performing QSAR analysis. thank you. steven -------------------------------------------------------------------------- Steven Creve steven.creve@chem.kuleuven.ac.be Labo Quantumchemie Celestijnenlaan 200F 3001-HEVERLEE tel: (32) (16) 32 73 93 BELGIUM fax: (32) (16) 32 79 92 -------This is added Automatically by the Software-------- -- Original Sender Envelope Address: chemistry-request@www.ccl.net -- Original Sender From: Address: Steven.Creve@chem.kuleuven.ac.be CHEMISTRY@www.ccl.net: Everybody | CHEMISTRY-REQUEST@www.ccl.net: Coordinator MAILSERV@www.ccl.net: HELP CHEMISTRY or HELP SEARCH | Gopher: www.ccl.net 73 Anon. ftp: www.ccl.net | CHEMISTRY-SEARCH@www.ccl.net -- archive search Web: http://www.ccl.net/chemistry.html From chemistry-request@www.ccl.net Tue Jan 5 12:51:37 1999 Received: from ccl.net (atlantis.ccl.net [192.148.249.4]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id MAA11194 Tue, 5 Jan 1999 12:51:36 -0500 (EST) Received: from mailrelay1.lrz-muenchen.de (mailrelay1.lrz-muenchen.de [129.187.254.101]) by ccl.net (8.8.6/8.8.6/OSC 1.1) with ESMTP id MAA19663 for ; Tue, 5 Jan 1999 12:51:34 -0500 (EST) Received: from [129.187.26.175] by mailout.lrz-muenchen.de with ESMTP; Tue, 5 Jan 1999 18:50:29 +0100 Received: from pet.localnet.hh (pet [192.17.5.9]) by localnet.hh (8.9.1/8.9.1) with SMTP id DAA05234; Tue, 5 Jan 1999 03:46:14 +0100 (CET) Message-Id: <199901050246.DAA05234@localnet.hh> Received: by pet.localnet.hh (NX5.67g/NX3.0X) id AA01131; Mon, 4 Jan 1999 08:24:25 +0100 Content-Type: text/plain Mime-Version: 1.0 (NeXT Mail 4.2mach v148) X-Image-Url: http://www.lrz.de/~heller/hheller.tiff X-Face: #K8+M!=vK|\bQ[&P:z)d$Q:i(@36P1!7if~nI%OB\9I>;p5Vs<_YkU;].K(chZJ,k8X96t ]!d?;H/|Jc:Hr-hhbBow![@FkT0xNOVXK|TO%Jg@Xz Date: Mon, 4 Jan 1999 08:24:24 +0100 To: everyone@localnet.hh Subject: Please note change of my HOME email address Reply-To: hheller@gmx.de X-Url: http://www.lrz.de/~heller Hello, I will be changing my provider at the end of January 1999 and this will also mean that my OLD email: heller@altoetting-online.de will no longer be working after that date. Please make a note of my NEW email: hheller@gmx.de which is already effective and which will be the only private email address after January 31st, 1999. My work email address (heller@lrz.de) is unaffected by this change. I wish you all a happy new year 1999 and hope to keep in contact with you in the future. Bye, Helmut --- Servus, Helmut (DH0MAD) ______________NeXT-mail accepted________________ Phone: +49-8671-881665 "Knowledge must be gathered and cannot be given" hheller@gmx.de ZEN, one of BLAKES7 FAX: +49-8671-881665 ------------------------------------------------ Dr. Helmut Heller, Muehldorfer Str. 72, 84503 Altoetting, GERMANY From chemistry-request@www.ccl.net Sun Jan 3 16:08:42 1999 Received: from panix3.panix.com (teQSL9yw46ZlUvrLuifPg9C07ymIh/6a@panix3.panix.com [166.84.1.68]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id QAA21772 Sun, 3 Jan 1999 16:08:42 -0500 (EST) From: kynn@panix.com Received: (from kynn@localhost) by panix3.panix.com (8.8.5/8.8.8/PanixU1.4) id QAA03538; Sun, 3 Jan 1999 16:08:41 -0500 (EST) Date: Sun, 3 Jan 1999 16:08:41 -0500 (EST) Message-Id: <199901032108.QAA03538@panix3.panix.com> To: chemistry@www.ccl.net Subject: Generalizations of Lennard-Jones potential? I'm programming a simple model that requires being able to represent a high degree of chemical heterogeneity among interacting particles. As a first pass, I would like to model point particles interacting via spherically symmetrical potentials, with varying interaction strengths, depending on the "chemical identity" of the particles. I originally thought of using a Lennard-Jones 6-12 (LJ) potential, suitably modified to represent various strengths of interaction, but this has turned out to be a much harder problem than I expected (it may not be obvious why this would be the case, but it has to do with certain of the model's requirements.) Does anyone know of a published study on generalizing the LJ potential roughly along these lines? Thanks in advance, K. Jones P.S. For the pathologically patient, here's a more detailed account of the problem. Fix a characteristic interparticle distance r0. Let LJ(r) stand for a standard Lennard-Jones 6-12 potential, with minimum -1 at r = r0, and let LJR(r) stand for the repulsive part (~ (r0/r)^12) of LJ(r). Let U(m, r) stand for the desired generalization of this potential, where the parameter m is defined as m := U(m, r0). Ideally, we would like U(m, r) to satisfy the following requirements: 1. U(m, r) must be continuous in both m and r, and have continuous derivatives with respect to r of up to no less than order 2. 2. U(-1., r) = LJ(r) (i.e., we must be able to recover LJ(r) as a special case of U(m, r)). 3. If m < 0 (the partially attractive case), U(m, r) must have exactly one extremum, a minimum, at r0. 4. If m < 0, we want U(m, r) to "quickly" approximate |m|*LJ(r), as r->oo. 5. For all m, we want U(m, r) to "quickly" approximate LJ(r) as r->0. 6. U(m, r) must be "not too exorbitant" computationally . I could make the terms in quotes more precise, but this seems premature at this stage. Also, I realize that, most likely, the best candidate will fit this set of requirements very imperfectly. The first solution I tried was to use U(m, r) := |m|*LJ(r) for the partially attractive case (m < 0), and U(m, r) := m*LJR(r) for the purely repulsive case (m > 0). This scheme has several serious drawbacks, notably that it is not continuous at m = 0, and, for small |m|, does not approach LJR(r) quickly enough as r->0, which translates to a repulsive core that is too small. Another partial answer, based on an idea originally proposed by Weeks et al., (see p. 11 of Allen and Tildesley's "Computer Simulation of Liquids") is: For (m <= 0): LJ(r) + 1 + m (r < r0) U(m, r) := |m|*LJ(r) (r >= r0) The most serious objections to this idea is that it doesn't readily generalize to the case m > 0. From chemistry-request@www.ccl.net Mon Jan 4 11:26:43 1999 Received: from carpincho.overnet.com.ar (carpincho.overnet.com.ar [200.10.96.19]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id LAA25947 Mon, 4 Jan 1999 11:26:40 -0500 (EST) Received: from nppp105.overnet.com.ar (nppp105.overnet.com.ar [200.16.162.109]) by carpincho.overnet.com.ar (8.8.8/8.8.5) with SMTP id NAA07747 for ; Mon, 4 Jan 1999 13:40:49 -0300 Message-Id: <3.0.6.16.19990104132430.354797f8@pop.overnet.com.ar> X-Sender: dfa953uo@pop.overnet.com.ar (Unverified) X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.6 (16) Date: Mon, 04 Jan 1999 13:24:30 To: chemistry@www.ccl.net From: "Dr. Daniel Glossman" Subject: Pop=Bonding calculation Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Dear netters: I have been doing some G94W calculations on several molecules with the POP=Bonding keyword, but I don't know how to do an interpretation of the results and the Gaussian 94 manual is not very informative in this respect. For example, after the title "... Condensed to atoms (all electrons..." I got a symmetric matrix with diagonal elements equal to zero. 1) What do the numbers mean? 2) Where do they come from? 3) Are they related to the so called bond order analysis? 4) Are there any bibliographic references for the calculation of these numbers? After that I got the total atomic charges. These charges are different from those coming from a traditional Mulliken population analysis in that they are not summed up to the charge of the molecule under study. Again, 1) What do these charges mean? 2) What's the meaning of the summation? 3) Where do these numbers come from? 4) which are the bibliographic references? Thanks in advance for a kind consideration of this request. Best regards Dr. Daniel Glossman ***************************************************************************** Universidad Nacional de Lujan Departamento de Ciencias Basicas e-mail: glossman@overnet.com.ar Casilla de Correo 221 phone: (+54) 1 5533797 (6700) Lujan fax: (+54) 1 5539824 Republica Argentina **************************************************************************** ** From chemistry-request@www.ccl.net Mon Jan 4 13:30:01 1999 Received: from degusse.IRO.UMontreal.CA (degusse.IRO.UMontreal.CA [132.204.24.51]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id NAA00261 Mon, 4 Jan 1999 13:30:01 -0500 (EST) Received: from iro.umontreal.ca (dhcp27076.IRO.UMontreal.CA [132.204.27.76]) by degusse.IRO.UMontreal.CA (8.9.1/8.9.1) with ESMTP id NAA01092; Mon, 4 Jan 1999 13:29:52 -0500 (EST) Message-ID: <3691076C.FF72F069@iro.umontreal.ca> Date: Mon, 04 Jan 1999 13:24:45 -0500 From: Francois Major Organization: Universite de Montreal X-Mailer: Mozilla 4.05 [en] (Win95; I) MIME-Version: 1.0 To: CHEMISTRY@www.ccl.net, amber@cgl.ucsf.edu Subject: Postdoctoral positions available at LBIT Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit The Laboratory for Informatics and Theoretical Biology (LBIT) is seeking two postdoctoral trainees: one in the field of RNA 3-D modeling and one in QSAR. The LBIT is located at the University of Montreal, Montreal, Quebec, Canada. The salaries and conditions are conformed to the MRC of Canada. Interested Ph.D. in a related field, such as CS, Biol. Sc., Biochem., Chem., etc., with experience in the use of computer softwares (molecular dynamics, MM/QMM, docking, etc.) or in programming (C++) are invited to submit their CV to: Dr. Francois Major major@iro.umontreal.ca -- Department of Computer Science Universite de Montreal C.P. 6128, Succ. Centre-Ville Montreal, Quebec H3C 3J7 Canada (514)343-7091 (514)343-5834 fax www.iro.umontreal.ca/~major From chemistry-request@www.ccl.net Mon Jan 4 14:08:30 1999 Received: from gate-sl1.mdli.com (ns2.mdli.com [208.200.221.3]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id OAA03262 Mon, 4 Jan 1999 14:08:29 -0500 (EST) Received: (from smap@localhost) by gate-sl1.mdli.com (8.8.8/8.8.8) id LAA11303; Mon, 4 Jan 1999 11:00:52 -0800 (PST) Received: from puffin.mdli.com(191.254.19.10) by gate-sl1.mdli.com via smap (V2.1) id xma011299; Mon, 4 Jan 99 11:00:50 -0800 Received: from shepherd.mdli.com by puffin.mdli.com (8.8.5/BCH1.0) id LAA08643; Mon, 4 Jan 1999 11:07:56 -0800 (PST) Received: by shepherd.mdli.com with Internet Mail Service (5.5.2232.9) id ; Mon, 4 Jan 1999 11:09:00 -0800 Message-ID: <8E13FBBA26A9DA119ED400A0C9AB2CE55A7856@shepherd.mdli.com> From: Guenter Grethe To: CHMINF-L , Computational Chemistry List , Diversity LIst , Medicinal Chemistry List Cc: Gerald Maggiora , Guenter Grethe , Jacques Weber , Janet Ash , John Blackmore , Johnny Gasteiger , Peter Willett , Rainer Moll , Vincent van Geerestein Subject: 5th Chemical Structures Conference - Abstract Submission Date: Mon, 4 Jan 1999 11:08:59 -0800 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2232.9) Content-Type: text/plain Happy New Year! !999 has arrived and the Fifth International Conference on Chemical Structures (June 6-10, 1999) in Noordwijkerhout, The Netherlands, is rapidly taking shape. This is just a short reminder that the deadline for submitting abstracts for oral presentations and/or posters is January 15, 1999. Details can be found at: http://ChemWeb.com/conference/5iccs/5iccs.html Best regards, Guenter Grethe Conference Chair From chemistry-request@www.ccl.net Mon Jan 4 20:46:10 1999 Received: from ulys.POSTE.UMontreal.CA (ulys.POSTE.UMontreal.CA [132.204.2.41]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id UAA06109 Mon, 4 Jan 1999 20:46:09 -0500 (EST) Received: from [132.204.150.83] (cerca201.CERCA.UMontreal.CA [132.204.150.201]) by ulys.POSTE.UMontreal.CA with ESMTP id UAA02761 (8.8.7/IDA-1.6); Mon, 4 Jan 1999 20:46:09 -0500 (EST) Date: Mon, 4 Jan 1999 20:46:09 -0500 (EST) X-Sender: salahub@ulys.poste.umontreal.ca Message-Id: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: chemistry@www.ccl.net From: Dennis Salahub Subject: CERCA - CADD Symposium, Montreal, Apr. 14 - 16 Cc: symposium@cerca CERCA - Centre for Research on Computation and its Applications Symposium "New Perspectives for Computer-Aided Drug Design" Montreal, April 14-16, 1999 Keynote session: John Erickson, NCI; Peter Kollman, UCSF; Yvonne Martin, Abbott;John Moult, Maryland; Harel Weinstein,Mount Sinai Invited and Contributed Sessions I - Solvation and Environmental Effects:Jia-Li Gao, SUNY, Buffalo;Gerhard Hummer, Los Alamos;Charlie Brooks, Scripps; Richard Friesner, Columbia II - Bioinformatics and Structural Predictions: Roland Dunbrack, Fox Chase Cancer Center;Janice Glasgow, Queen's University;Andrej Sali, Rockefeller University; Richard Griffey, Isis Pharmaceuticals Details and registration: http://www.cerca.umontreal.ca/cadd Version francaise http://www.cerca.umontreal.ca/cmao Please let us know of any other lists that might like to receive this information. The CERCA/Pharmaceutical chemistry group: John Gunn, Francois Major, Benoit Roux, Dennis Salahub From chemistry-request@www.ccl.net Tue Jan 5 19:06:13 1999 Received: from mailrelay1.cc.columbia.edu (cu41754@mailrelay1.cc.columbia.edu [128.59.35.143]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id TAA14420 Tue, 5 Jan 1999 19:06:12 -0500 (EST) Received: from still3.chem.columbia.edu (still3.chem.columbia.edu [128.59.112.36]) by mailrelay1.cc.columbia.edu (8.8.5/8.8.5) with ESMTP id TAA10911 for <@smtp.columbia.edu:chemistry@www.ccl.net>; Tue, 5 Jan 1999 19:06:09 -0500 (EST) Received: (from shenkin@localhost) by still3.chem.columbia.edu (980427.SGI.8.8.8/970903.SGI.AUTOCF) id TAA27043 for chemistry@www.ccl.net; Tue, 5 Jan 1999 19:06:10 -0500 (EST) From: "Peter Shenkin" Message-Id: <9901051906.ZM27041@still3.chem.columbia.edu> Date: Tue, 5 Jan 1999 19:06:10 -0500 In-Reply-To: kynn@panix.com "CCL:Generalizations of Lennard-Jones potential?" (Jan 3, 4:08pm) References: <199901032108.QAA03538@panix3.panix.com> X-Mailer: Z-Mail (3.2.3 08feb96 MediaMail) To: chemistry@www.ccl.net Subject: Re: CCL:Generalizations of Lennard-Jones potential? Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii On Jan 3, 4:08pm, kynn@panix.com wrote: > Subject: CCL:Generalizations of Lennard-Jones potential? > I'm programming a simple model that requires being able to represent a > high degree of chemical heterogeneity among interacting particles. As > a first pass, I would like to model point particles interacting via > spherically symmetrical potentials, with varying interaction > strengths, depending on the "chemical identity" of the particles. > > I originally thought of using a Lennard-Jones 6-12 (LJ) potential, > suitably modified to represent various strengths of interaction, but > this has turned out to be a much harder problem than I expected (it > may not be obvious why this would be the case, but it has to do with > certain of the model's requirements.) > > Does anyone know of a published study on generalizing the LJ > potential roughly along these lines? ... You say *what* you want to do, but not *why* you want to do it. Are you looking for greater accuracy vis-a-vis high-quality experimental or q-mechanical results? Are you trying to circumvent overflows in monte-carlo simulations when the particles get too close? Or what? Don't forget that your accuracy with respect to the Word of God (quantum mechanics :-) ) is always limited by the neglect of 3-and-more-body effects. Still, Tom Halgren, in one of his MMFF papers (J. Comput. Chem., but I don't have the ref. handy) did a study that led him to the use of a 7-14 potential with rather complicated combining rules. Whether or not you like his solution, he makes careful comparisons between simple molecular-mechanics nonbonded potentials and highly accurate pairwise- central potentials that have been derives for specific simple systems. Also, Tom and others do "buffering" of one kind or anotheer. The simplest such scheme is to replace r in the LJ equation by r+a, where a is a small constant whose value is chosen to have little effect at "typical" distances, but to avoid computational pathologies if/when r should become close to zero. In our era of IEEE arithmetic this particular motivation is no longer of such great concern, but of course the buffering also changes the functional form and can be adjusted to achieve a better fit to accurate data. Hope this helps, -P. -- *** "Freedom's just another word for nothing left to lose." (B. Yeltsin)*** *Peter Shenkin; Chemistry, Columbia U.; shenkin@columbia.edu (212)854-5143* * MacroModel URL: http://www.cc.columbia.edu/cu/chemistry/mmod/mmod.html * From chemistry-request@www.ccl.net Tue Jan 5 19:18:06 1999 Received: from relay2.UU.NET (relay2.UU.NET [192.48.96.7]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id TAA14508 Tue, 5 Jan 1999 19:18:05 -0500 (EST) Received: from uucp1.uu.net by relay2.UU.NET with SMTP (peer crosschecked as: uucp1.uu.net [192.48.96.81]) id QQfwxl00458; Tue, 5 Jan 1999 19:18:05 -0500 (EST) Received: from m10.UUCP by uucp1.uu.net with UUCP/RMAIL ; Tue, 5 Jan 1999 19:18:05 -0500 Received: from holly by m10.lorentzian.com; Tue, 5 Jan 1999 19:12:28 -0500 Received: (from frisch@localhost) by holly (950413.SGI.8.6.12/950213.SGI.AUTOCF) id TAA05627 for CHEMISTRY@www.ccl.net; Tue, 5 Jan 1999 19:15:13 -0500 From: frisch@lorentzian.com (Mike Frisch) Message-Id: <199901060015.TAA05627@holly> Subject: Re: CCL:G:SUMMARY: Parallel G98 on x86/Linux To: uunet!helix.nih.gov!mn1@uunet.uu.net (M. Nicklaus), chemistry-request@ccl.net Date: Tue, 5 Jan 1999 13:07:05 -0500 (EST) In-Reply-To: from "M. Nicklaus" at Jan 5, 99 10:26:54 am X-Mailer: ELM [version 2.4 PL24] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: frisch@lorentzian.com M. Nicklaus writes: > > A couple of weeks ago, I posted the enclosed question about > running Gaussian 98 in parallel on x86/Linux. Here's the > summary of the responses I got, slightly edited for brevity. > > >From these few replies it appears that no one has been > able so far to get G98 to run properly in this setup. > > Maybe Gaussian, Inc. could comment on this. Doug? > > A Belated Happy New Year to everyone. > > Marc > There are a number of misconceptions in the replies posted with this message. Let me explain the situation: 1. The Linux version can use SMP parallelism for HF and DFT using the usual system-V shared memory and unix fork/join calls. This does not require the parallel blas library. It does require the following: a. A machine with more than one procesor and Linux compiled with SMP support, and with the kernel parameter for maximum shared memory segment set to some large value (the maximum amount of memory you'll want to give G98, typically 3/4 or more of the amount of physical memory on the machine). b. An input deck for a reasonably large job (i.e., not water STO-3G!) with %nproc set to requiest more than one processor. %nproc works by default with the Linux version (even on machines with only 1 CPU), but the default amount of shared memory allowed by the kernel is not sufficent to run anything, so jobs with %nproc will fail until the kernel is rebuilt with this limit increased. 2. We have not attempted to build with the Portland compiler and the parallel blas library. We plan to use the thread features of the Portland compiler to make our own parallel calls to the (serial) blas library routines, but this is not in the current revision of G98. 3. Some post-SCF jobs spend most of their time in matrix multiplies, which current use the serial blas routines and thus are not yet parallelized. 4. Running efficiently in parallel in 2 CPUs requires twice as much memory bandwidth a running on 1 CPU. Server vendors like SGI, IBM, and Digital/Compaq devote considerable effort and cost to having enough bandwidth to keep several processors going. A system with two Intel processors but only enough bandwidth to keep one busy is not going to run twice as fast in parallel, nor will it run two separate jobs in the same elapsed time as a single job. The observed performance will depend strongly on the details of the hardware. 5. For networks of Linux machines we do support parallelism in HF and DFT calculations using Linda, just as for networks with other models of workstations. Linda is required only for network parallelism, not for SMP paralleism. 6. The use of the tuned blas library produces a substantial (serial) performance improvement in those jobs which spend a lot of time doing matrix multiplies. It doesn't make much difference for jobs which spend their time doing integrals. So it will make a big difference for QCISD(T) but not for HF. 7. We use the same compiler (Portand) and blas libraries on both Windows/NT and Linux, and performance is similar. The differences in reported times are a few percent and probably reflect Linux reporting system time overhead (as other unix versions do) while this isn't included on a per-process basis under NT. Elapsed times are sometimes a bit better under Linux because of better I/O, but the best summary of the situation is that there isn't a signficant performance diffence for running a single, stand-alone job under one OS or the other. Mike Frisch From chemistry-request@www.ccl.net Tue Jan 5 22:38:37 1999 Received: from balihai.uchicago.edu (balihai.uchicago.edu [128.135.136.49]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with SMTP id WAA15300 Tue, 5 Jan 1999 22:38:37 -0500 (EST) Received: from localhost by balihai.uchicago.edu via SMTP (950413.SGI.8.6.12/931108.SGI.ANONFTP) for id VAA08390; Tue, 5 Jan 1999 21:39:42 -0600 Date: Tue, 5 Jan 1999 21:39:42 -0600 (CST) From: "Fred P. Arnold" To: Computational Chemistry List Subject: NBO/NRT and CpM systems Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hello, When using NBO on transition metal systems with Cp, the standard routine prints a structure with a mixture of single and 3-center bonds (C-C-M), at least for early metals such as Ti. When attempting to use the Natural Resonance Theory section of the NBO program, how does one generate a NRTLST entry (manual specification of reference structure), since the 3Cbond keyword is inoperable? Should I pick one localized structure, say a cyclopentadiene with two localized double bonds, and a lone pair on the carbon not involved, pointing towards the metal, or is there a better way? This question arises since the default routine is incapable of finding a non-ionic structure on my system (CpL2TiR), and therefore NRT fails with a thud, and without writing even the ionic structure as a starting point. Thank you for your time. -fred Frederick P. Arnold, Jr. "...But if you give theorists enough Advanced Research Systems time, they can predict anything." 5640 S. Ellis Ave Dr. M. S. Turner, U. Chicago. Chicago, IL 60637