From chemistry-request@ccl.net Thu Aug 6 04:37:01 1992 From: Matthew E. Harbowy Subject: the mopac 6.0 / RS 6000 / fortran nightmare ends! To: CHEMISTRY@ccl.net (computational chemists) Date: Wed, 5 Aug 92 23:06:46 EDT Status: R I feel as though I've gone through a secret FORTRAN initiation but I have come to the realization this is only the beginning. Just in case you forget the original cry for help... ------------------------------------------------------------------------ Headaches and frustration! I am attempting to compile MOPAC 6.0 for the IBM RS-6000 and all I get are endless streams of error messages. Is there a kind soul who could provide the correct code, presumably by FTP, which would compile easier of the RISC sytem I use? My compiler is XLF. ------------------------------------------------------------------------ Then a kind soul replied ------------------------------------------------------------------------- From: gene@eastrg2.cray.com (Eugene Fleischmann) I have taken Mopac 6.0 from ouchem.oakland.edu via anonymous ftp. The address is: 141.210.108.5. The directory is: pub/unix Best Wishes, Gene Fleischmann ------------------------------------------------------------------------- and another soul replied ------------------------------------------------------------------------- From: breneman@XRAY.CHEM.RPI.EDU I speak from experience (bitter) which I gained trying to get GAMESS running on our RS/6000 network here at RPI. I think that the problems may be similar, even though I haven't tried compiling Mopac 6.0 on the net yet. Our RS/6000's originally came with the "default" IBM xlf compiler, which atB that time was version 1.0. A slightly later version with a number of bug fixes was known as xlf 1.5. Both of these are totally useless for a lot of well established code (Gaussian90, GAMESS, AIMPAC...the list goes on). I struggled along with this situation by trying to track down each individual error until I discovered that the object code which was generated actually gave significantly erroneous results without runtime errors! I tracked this down to the optimizer. If you turn off optimization completely, and sacrifice a factor of 7-8X in performance, you can get the code to produce correct results, but with a very picky parser. The answer to this was to go to IBM with the problem. I was told that the "high performance" fortran compiler xlf 2.x was the way out of this problem, but that you have to pay real money for this! Eventually, xlf 2.2 was placed on the HESC contract which gives academic institutions low/no fee access to software in the state of New York. The appropriate compiler was then available for $300. This solved almost all of the problems. The remaining minor issues were addressed by turning off optimization on a couple of outer-loop routines which were still causing trouble. Gaussian90 compiled without errors at all under this compiler. Take home lessons: 1) Make sure to compare results of optimized executable with those from a totally unoptimized version (no -O). 2) Get IBM to cough up the compiler which works. Curt Breneman -------------------------------------------------------------------------- Somewhere along the way i came to the belief that the copy of MOPAC 6 which was given to me by my adviser was loaded with pitfalls. So I took the easy way out and got a fresh FTP copy. It still wouldn't compile, but it got closer to completion. I turned off all optimization and just compiled bare-bones xlf -c *.f which got everything to compile, but xlf -o mopac6 *.o died because of two silly timing calls. Well, I trashed them as follows ----------------------------------------------------------------------- I changed date.f to read subroutine date(a) character a*24 a="NO TIMER UTILITY" return end and I changed second.f to read FUNCTION SECOND() DOUBLE PRECISION SECOND C****************************************************** C C SECOND, ON EXIT, CONTAINS THE NUMBER OF CPU SECONDS C SINCE THE START OF THE CALCULATION. C C****************************************************** LOGICAL SETOK CHARACTER*1 X, GETNAM*80 DIMENSION A(2) SAVE SETOK, SHUT DATA SETOK C C IF YOU ARE NOT USING A VAX OR A UNIX COMPUTER, UNCOMMENT THE C NEXT LINE CPU=0.0 C C IF YOU ARE NOT USING A VAX OR A UNIX COMPUTER, REMOVE THE NEXT LINE C Y=ETIME(A) C CPU=A(1) *********************************************************************** with the remainder of the file unchanged. ------------------------------------------------------------------------ this compiled no-problem, and linked no-problem. The .csh provided, however, has stumped me. So I wrote my own: ------------------------------------------------------------------------ #! /bin/csh set job = $1 rm $job.*.old foreach file (out log arc end) if -e $job.$file mv $job.$file $job.$file.old end setenv FOR005 $job.dat setenv FOR006 $job.out setenv FOR009 $job.res setenv FOR010 $job.den setenv FOR011 $job.log setenv FOR012 $job.arc setenv FOR013 $job.gpt setenv FOR016 $job.syb setenv FOR020 $job.ump time /u/mopac/mopac6 > $job.tim This .csh accepts the name of the file , with .dat as the input deck. It's crude but it works So, for all the fellow FORTRAN initiates who wrote to sympathize, thanks for your support. And here's the solution back at you. fondly matt. ----------------------------------------------------------------------------- Matthew E. Harbowy Cornell U Chemistry Department Ithaca NY 14853 Bharbowy@chemres.tn.cornell.edu (or) ikf@vax5.tn.cornell.edu ----------------------------------------------------------------------------- --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From chemistry-request@ccl.net Thu Aug 6 06:55:56 1992 Date: 06 Aug 92 11:17:37-2300 From: Yves chapron To: CHEMISTRY@ccl.net, harbowy@CHEMRES.TN.CORNELL.EDU Subject: Re: mopac 6.0 Status: R Math, Serge Crouzy made a compilation of MOPAC but the executable program didn't work ! Good luke yves Chapron --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From chemistry-request@ccl.net Thu Aug 6 15:25:37 1992 Date: Thu, 6 Aug 92 14:52 GMT From: RWOODS@VAX.MOLECULAR-BIOPHYSICS.OXFORD.AC.UK To: CHEMISTRY Subject: HP plots from macro-model Status: R Pat: I have had gread success converting post-script to hpgl files with a program called Laser Plotter from Insight. I have not used this program for some time and I have heard rumours that there is a better version available now. Laser plotter was quite reasonably priced so it is worth a call to Insight. Insight Developement Corp. 1024 Country Club Drive, Suite 140 Moraga CA 94556 (415)376-9451 I hope this is the current address. Good luck, Robert J. Woods Glycobiology Institute University of Oxford South Parks Road Oxford OX1 3QU U.K. --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From chemistry-request@ccl.net Thu Aug 6 16:59:35 1992 Date: Thu, 6 Aug 92 10:19:02 GMT-0500 From: bash@next1.scri.fsu.edu (Paul Bash) To: chemistry@ccl.net Subject: Computational Biology Institute at Argonne National Laboratory Status: R NSF-DOE INSTITUTE IN COMPUTATIONAL BIOLOGY A Week-Long Institute at the Advanced Computing Research Facility Argonne National Laboratory December 7-11, 1992 Areas to be covered: Biomolecular Structure Determination Simulation of Enzyme Reactions Protein Folding Molecular Dynamics and Drug Design Sequence Analysis Institute Faculty: Axel Brunger, Yale University Arieh Warshel, University of Southern California Peter Wolynes, University of Illinois Klaus Schulten, University of Illinois Gary Olsen, University of Illinois Eligibility and Selection Criteria: Institute limited to 25 graduate students and/or postdoctoral researchers. Preference given to those likely to advance computational biology research. Applications due September 30, 1992, supported by a letter of recommendation. Note: Participants will receive free lodging for December 6-10 and a stipend for meals and incidental expenses. Travel costs will be reimbursed up to $750. For further information and application materials, write, call, or email: Karen Neumann Mathematics and Computer Science Division Argonne National Laboratory Argonne, Illinois 60439-4844 708-252-7163 neumann@mcs.anl.gov The Institute in Computational Biology is being organized by Paul Bash of Florida State University (904-644-7048,bash@scri.fsu.edu) and Ewing Lusk of Argonne National Laboratory (708-252-7852, lusk@mcs.anl.gov), who can also be contacted for more information. It is supported by the National Science Foundation and by the U.S. Department of Energy. --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From chemistry-request@ccl.net Thu Aug 6 18:18:55 1992 Date: Thu, 6 Aug 92 15:11 GMT From: RWOODS@VAX.MOLECULAR-BIOPHYSICS.OXFORD.AC.UK To: CHEMISTRY Subject: Mopac 6.0 BUG in ESP Status: R Dear Netters: If you are making use of the POTWRT keyword to write ESP's to a file in MOPAC 6, there is a line of code in ESP.FOR that MUST be corrected. Due to a simple omission, the cartesian coordinates are not all converted from Angstroms to Bohrs--just the x- and y-! If this has been previously reported please ignore this message and continue with your reqular programming schedule. The error occurs around line 950 in ESP.FOR. The relevent code (corrected) is below: IF(POTWRT) THEN OPEN(21,STATUS='NEW') WRITE(21,'(I5)') NESP DO 410 I=1,NESP 410 WRITE(21,420) ESP(I),POTPT(1,I)/BOHR,POTPT(2,I)/BOHR, C --- BUG FIX --- C 1POTPT(3,I) C 1POTPT(3,I)/BOHR C ENDIF Robert J. Woods Glycobiology Institute University of Oxford South Parks Road Oxford, OX1 3QU U.K. --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From chemistry-request@ccl.net Thu Aug 6 18:19:33 1992 Date: Thu, 6 AUG 92 09:57:56 From: br1@inel.gov To: chemistry@ccl.net Subject: Q: MNDO level shifting in GAUSSIAN Status: R Dear Netters, I am interested in plotting the HOMO-LUMO gap verses the number of monomer units for several polyphosphazenes. In one case I am dealing with 26 atoms per monomer unit, so semiemperical methods prevail. For specifically, I am using MNDO/STO-3G, as implemented in the GAUSSIAN program suite. I know that SCF virtuals are level shifted by 0.2au by default in overlay 500, but I am unclear as to whether the semiemperical links (400) employ the same trick. A few comparisons have indicated that MNDO virtuals are level shifted by 0.1au, but I am not at all sure of this result. Can anyone out there let me know whether or not this result is correct? Many Thanks, Randy Boehm BR1@INEL.GOV --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From chemistry-request@ccl.net Thu Aug 6 19:34:22 1992 Date: Fri, 12 Jun 92 11:47:58 PDT From: rak@igor.caltech.edu (Robert A. Kumpf) To: chemistry@ccl.net Status: R Hi, We have been trying to run a PMF calcuation for a system of two solutes in a box of TIP3P water using BOSS version 2.7. The first step of the simulation appears to work fine; however, all attempts to continue the simulation using the option ISTART=3 have resulted in values of 'Infinity' being written to the UP file. The ISTART=3 option is the recommended choice for beginning the next step in a free energy perturbation calculation. Has anyone encountered this problem as well? Any suggestions for a fix would be greatly appreciated. Thanks, Bob Kumpf E-mail rak@igor.caltech.edu --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From chemistry-request@ccl.net Thu Aug 6 20:35:20 1992 Date: Thu, 6 Aug 92 14:44:39 -0400 From: rafael@ferry.polymer.uakron.edu (Rafael Garcia-Ramirez) To: CHEMISTRY@ccl.net Subject: Mopac Lives !! Status: R Guys: I have been using Mopac 6.0 for a while now (and I guess I've learned to love it). Nevertheless, there is a particular WARNING message that has always intrigued me. An example of such message, taken >from the output ( .out) file is as follows: ----------------------------------------------------------------------- AM1 EF PRECISE NOINTER /tmp_mnt/home/eby/rafael/sybyl/pbo/pbo2bend2.dat " " GEOMETRY OPTIMISED USING EIGENVECTOR FOLLOWING (EF). SCF FIELD WAS ACHIEVED AM1 CALCULATION VERSION 6.00 05-Aug-92 FINAL HEAT OF FORMATION = 145.40820 KCAL TOTAL ENERGY = -5879.16684 EV ELECTRONIC ENERGY = -43266.97471 EV CORE-CORE REPULSION = 37387.80787 EV WARNING -- GEOMETRY IS NOT AT A STATIONARY POINT IONIZATION POTENTIAL = 8.90687 NO. OF FILLED LEVELS = 85 MOLECULAR WEIGHT = 470.443 SCF CALCULATIONS = 92 COMPUTATION TIME = 27 MINUTES AND 15.990 SECONDS ----------------------------------------------------------------------- I have noticed that this "error" appears sometimes, however exclusively, when a certain geometrical constraint(s) has been placed in the input Z-matrix [thus the geometry optimization is carried out for every atomic coordinate, except for the one(s) corresponding to the given constraint]. I understand that the minimizers in Mopac may not always find the "global minimum" in energy (heat of formation) when the constraint means a significant disturbance from the equilibrium geometry and, therefore, such a message would be likely to appear. However, some- times the minimum found clearly corresponds to a "true" minimum in energy state (judging by the perfectly smooth trend observed in the ENERGY-vs- CONSTRAINT VALUE curve) and the message still appears !! ... Can somebody PLEASE comment on the possible implications behind this and / or a possible way to fix it ??. You can answer directly to me or, if it's easier (I'm ready to accept any public remarks about my IGNORANCE ), to the board. Thanks a lot .... Rafael G. Ramirez ------------------------------------------------------------------ email : rafael@eby.polymer.uakron.edu phone: (216) 972-5810 usmail: Institute of Polymer Science FAX : (216) 972-5290 The University of Akron, Akron, OH 44325-3909 U. S. A. ------------------------------------------------------------------ --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From chemistry-request@ccl.net Thu Aug 6 22:14:15 1992 Date: Thu, 6 Aug 92 18:30:12 EDT From: bruc@horta.bms.com (Robert Bruccoleri) Subject: MM2 FORCEFIELD PARAMETER DATABOOK To: PA13808 Status: R > MM2 FORCEFIELD PARAMETER DATABOOK > Now for the bad news. The book is published by MARTECH ,541 Tokyo Club > bldg. Kasumigaseki 3-2-6 Chiyoda-ku Tokyo 100 Japan.and is priced at > ONE THOUSAND($1000.00)USA DOLLARS.Also it has a condition ofsale that it > not be"lent,hired out or otherwise lent out or otherwise circulated " U.S. Copyright law does not give a copyright holder the right to prevent lending a book. Currently, rentals can be restricted only for phonorecords and computer software. > Oh yes reading the fine print I see that when publishing work using > parameters from the book "He(she) shall not make public the parameter it > itself"(More on that later). Again, U.S. Copyright law does not restrict rights to information, just the expression of ideas. Assuming that anyone can buy the book, and that you do not have to sign a non-disclosure agreement to get the book, their publication has made the parameters public, and anyone can use them (at least in countries which have copyright laws like the U.S.) There was a recent, relevant case that went to the Supreme Court which said that the names, addresses, and phone numbers in a phone book can be freely copied and used by a rival phone company as long as the information is not presented in the same way. It came up in a case where a company wanted to publish a directory covering geographically close towns that were served by six different phone companies. +-----------------------------------------+----------------------------+ | Robert E. Bruccoleri, Ph.D. | Bristol-Myers Squibb | | Senior Research Investigator | Pharmaceutical Research | | Macromolecular Modeling | Institute | | bruc@bms.com | Room H.3812C | | Phone: (609) 252-6165 | P.O. Box 4000 | | Fax: (609) 252-6030 | Princeton, NJ 08543 USA | +-----------------------------------------+----------------------------+ | Member of the League for Programming Freedom | +-----------------------------------------+----------------------------+ --- Administrivia: This message is automatically appended by the mail exploder. CHEMISTRY@ccl.net --- everybody; CHEMISTRY-REQUEST@ccl.net --- coordinator only OSCPOST@ccl.net : send something from chemistry; FTP: www.ccl.net --- From jkl@ccl.net Thu Aug 6 23:23:11 1992 Date: Thu, 6 Aug 1992 23:23:01 -0400 From: jkl@ccl.net To: chemistry@ccl.net Subject: Gaussian Mailing List Status: R I am forwarding info about Gaussian list: jkl@ccl.net ------------------------------------------------------------------- This is a reminder that we at Gaussian have assembled an electronic mailing list for those interested in news regarding the Gaussian series of programs. The mailing list will help us keep users updated on new versions of the program, improvements and enhancements to the program, and upcoming Gaussian workshops and user meetings. We invite any interested parties to subscribe to the list. To do so, please contact us at the e-mail address or phone number(s) given below, and we will be happy to include your name on the list. We also welcome any comments or suggestions you might have. Sincerely, David J. Moses Phone - 412-621-2050 e-mail - info@gaussian.com Gaussian, Inc. FAX - 412-621-3563