From chemistry-request@ccl.net Wed Jul 22 00:45:54 1992 Date: Tue, 21 Jul 92 23:24:59 -0400 From: jle@world.std.com (Joe M Leonard) To: chemistry@ccl.net Subject: MM3 atom types sought Status: R Does anybody have an on-line copy of the MM3 atom types that they could email me? How about a hard-copy they can fax? How about a reference to a list in the literature? I've seen a lot of the the 1989-90 MM3 papers, and they contain information in dribs and drabs... Also, has anybody collected the MM3 literature? I've seen about 23-30 papers since 1989, but am open for suggestions... It's probably rather small, so it should be possible to collect the lot! Thanks in advance, Joe Leonard jle@world.std.com --- Administrivia: The rest of 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 Wed Jul 22 08:39:43 1992 Date: Wed, 22 Jul 92 11:31 GMT From: RWOODS%VAX.MOLECULAR-BIOPHYSICS.OXFORD.AC.UK@OHSTVMA.ACS.OHIO-STATE.EDU To: CHEMISTRY@ccl.net Subject: G88 or other for FPS M64/30 Status: R Does anybody know where I can get a copy of G88 for an FPS M64/30? Preferably through anonymous FTP. While it may not be fast at least I won't have to wait in Queue! Thanks in advance, Robert J. Woods Glycobiology Institute University of Oxford South Parks Road Oxford Ox1 3QU --- Administrivia: The rest of 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 Wed Jul 22 11:41:18 1992 Date: Wed, 22 Jul 92 08:38 CDT From: Subject: QCPE MOPAC v6 for PC To: chemistry@ccl.net Status: R Can anyone tell me how much the 'DOS extender' that's needed for the new QCPE PC version of MOPAC costs? Also, how does one get the NDP FORTRAN compiler? (we'd rather know before we obtain the program). Thanks! Irene Newhouse --- Administrivia: The rest of 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 Wed Jul 22 14:11:29 1992 To: chemistry@ccl.net From: "Larry Wier" Organization: Saint Bonaventure University Date: 22 Jul 92 11:34:03 DST Subject: summary of FORTRAN 77 compiler replies Status: R I recently posted the following message on the list: ------------------ In an attempt to learn more about computational chemistry, I (with help from a friend in Computer Science) recently tried to install MindTool and Psi88 on a Sun SPARC station, but encountered problems when we found that there was not a FORTRAN 77 compiler on the system. Any recommendations on a low cost, FORTRAN 77 compiler that will run on the Suns? Larry Wier ------------------ and I received the following replies. Thank you for helping me. ======================================================= From: Pedro A.M. Vazquez Subject: Re: FORTRAN 77 compiler To: chemistry-request@ccl.net (Larry Wier) Date: Mon, 20 Jul 92 16:13:38 GMT-3:00 Helo Larry: No doubts, get f2c from research.att.com, compile it with the bundled Sun C compiler, then translate the f77 code you need to C. I've done this succesfully with psi88. Pedro ============================================================== Date: Tue, 21 Jul 92 10:01 MET From: Rob Hooft Subject: RE: FORTRAN 77 compiler To: LWIER@sbu.edu X-Vms-To: IN%"LWIER@sbu.EDU",HOOFT Comments: This mail was sent from cluster RUUCLA using PMDF V3.0 There is a free Fortran to C compiler available. Anyone with FTP access can get it easily: ftp research.att.com Name: netlib Password: your@email.adress cd f2c binary get libf77.Z get libi77.Z cd src get all.Z These are "shar" files, uncompress them, and read the file for further instructions. The packages contain README files. My experience is that code that is compiled with f2c and GCC (The Gnu C Compiler) is often only marginally slower than code which is directly compiled with vendor's f77 compilers. One warning: f2c compiles to an unreadable version of C. It is really a Fortran to C compiler, not a translator. It is not easy to modify the intermediate C code. Good luck, Rob Hooft, Department of Crystal and Structural Chemistry, Bijvoet Center for Biomolecular Research, University of Utrecht, The Netherlands. hooft@hutruu54.bitnet (alternative: hooft@chem.ruu.nl) ================================================================= Date: Tue, 21 Jul 92 11:41:08 EDT From: rec@arris.com (Roger Critchlow) Message-Id: <9207211541.AA22951@arris.com> To: "Larry Wier" In-Reply-To: Subject: FORTRAN 77 compiler You can obtain a fortran to c converter for free from netlib. Mail the message: send index from f2c to: netlib@research.att.com for more information. The sources can be obtained by e-mail or ftp. It is most often tripped up by I/O extensions that many people can't seem to resist inserting into their otherwise portable programs. I find it happiest when presented with computational subprograms, like linpack or it's ilk. -- rec -- Roger E. Critchlow Jr., rec@arris.com, ARRIS Pharmaceutical Corporation, Cambridge, MA 02139 617.494.0055, 617.494.0482 (fax), ============================================================== Thanks again for the help. Larry Wier ============================== Dr Larry Wier Dept of Chemistry St Bonaventure University St Bonaventure, N.Y. 14778 (716)375-2116 INTERNET:lwier@sbu.edu ============================== --- Administrivia: The rest of 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 Wed Jul 22 16:55:23 1992 Date: Wed, 22 Jul 92 12:44:14 EDT From: doug@star2.cm.utexas.edu (William D Clendening) To: CHEMISTRY@ccl.net Subject: continuous variation of hybrid orbitals Status: R Hi, i'm looking for information on constructing hybrid orbitals in which the hybridization is manipulated by varying parameters. i am interested in generating hybrid orbitals ranging from unhybridized through sp3 in a continuous fashion. i am trying to avoid imposing a particular symmetry on the hybridization space. Thanks for any help, doug clendening doug@star2.cm.utexas.edu cmft557@hermes.chpc.utexas.edu --- Administrivia: The rest of 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 Wed Jul 22 17:05:53 1992 Date: Wed, 22 Jul 1992 17:05:49 -0400 From: jkl@ccl.net To: doug@star2.cm.utexas.edu Subject: Re: continuous variation of hybrid orbitals Status: RO In answer to your question: >i'm looking for information on constructing hybrid orbitals >in which the hybridization is manipulated by varying parameters. >i am interested in generating hybrid orbitals ranging from unhybridized >through sp3 in a continuous fashion. i am trying to avoid imposing >a particular symmetry on the hybridization space. You might look to: %A Werner A. Bingel %A Wolfgang Luttke %T Hybrid Orbitals and Their Applications in Structural Chemistry %J Angew. Chem. Int. Ed. Eng. %V 20 %N 11 %P 899-911 %D 1981 %A Michael C. Bohm %A Peter C. Schmidt %T Electronegativities and Hardness of the Main Group Elements from Density Function Theory: Dependence on the Hybridization of the Chemical Bond %J Ber. Bunsenges. Phys. Chem. %V 90 %P 913-919 %D 1986 %A R. C. Haddon %T GVB and POAV Analysis of Rehybridization and pi-Orbital Misalignment in Non-Planar Conjugated Systems %J Chem. Phys. Lett. %V 125 %N 3 %P 231 %D 1986 %A Henry A. Bent %T An Appraisal of Valence-Bond Structures and Hybridization in Compunds of the First-Row Elements %J Chem. Rev. %V 61 %P 275-311 %D 1961 Jan jkl@ccl.net Dr. Jan K. Labanowski, Senior Research/Supercomputer Scientist/Specialist, etc. Ohio Supercomputer Center, 1224 Kinnear Rd, Columbus, OH 43212-1163 ph:(614)-292-9279, FAX:(614)-292-7168, E-mail: jkl@ccl.net JKL@OHSTPY.BITNET From chemistry-request@ccl.net Wed Jul 22 19:58:40 1992 From: Peter Tonellato Subject: Power Visualization System To: chemistry@ccl.net Date: Wed, 22 Jul 92 16:34:42 CDT Status: R Recently read some rumers concerning I.B.M.'s new PVS (power visualization system) and its use in computational chem/biology. Apparently someone at O.S.C. is involved. Anyone know what is going on and will they please post an overview of the thrust of the project? Thank you, -- Peter J. Tonellato US Mail: Department of Mathematics, Statistics and Computer Science Marquette University Milwaukee, Wisconsin 53233 Voice: 414-288-5228 (Office) 414-288-7573 (Message) FAX: 414-288-5472 (FAX) Arpa: tone@m.mscs.mu.edu --- Administrivia: The rest of 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 Wed Jul 22 21:48:23 1992 Date: Wed, 22 Jul 92 19:52:25 EDT From: m10!frisch@uunet.UU.NET (Michael Frisch) Subject: Re: G90 forces and displacements curiosity To: chemistry@ccl.net Status: R Gernot Frenking asks about a case where Gaussian 90 takes steps of RMS size exactly 0.0000 and therefore the optimization goes exactly nowhere. This is a known problem in Gaussian 90. There are a couple of possible causes, of which the most common is when the program realizes that it has a poor estimate of the Hessian and keeps reducing the allowed step-size until the Hessian improves. Eventually, the steps are scaled back to almost nothing. It can also happen if a really horrible step is taken (straight up in energy) and the program backs up closer and closer to the previous step. Like the first problem, this reflects a poor Hessian. All cases of this type of problem of which we are aware function properly in Gaussian 92. As I've said before here, we welcome reports of failing cases in the current version of Gaussian, so we can make sure the problems are rectified the next time around. The workaround for these types of problems is to restart the optimization, with an anlytic Hessian if practical. Thus # opt=calcfc geom=check guess=read will start a new optimization with an analytical Hessian using the last point of the previous optimization (actually, right before the last step). If the analytic Hessian is too costly, you can use geom=modify to request numerical differentiation with respect to selected variables. In Gaussian 92, if the analytical Hessian is not practical, you may want to try opt=cartesian. In most cases, the default of internal coordinates is more efficient, but if you have problems then you may have one of the cases in which optimization in cartesian coordinates is better. A typical example would be a system with a couple of fused rings with few or no double bonds -- this is a floppy molecule which is not efficiently described using z-matrix internal coordinates. Of course, if the optimization has gone astray, an intermediate point may be better than the last point done before the optimization. In Gaussian 92, you can request the structure at the Nth point of the previous optimization via: geom=(check,step=N) Note a) that this only works when using a checkpoint file from an incomplete of failed optimization and b) only a linearly independant set of low energy structures are retained during an optimization, so not all structures are available. The lowest energy point should always be available. Mike Frisch ------- --- Administrivia: The rest of 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 Wed Jul 22 23:53:05 1992 Date: Wed, 22 Jul 92 21:28:46 EDT From: m10!frisch@uunet.UU.NET (Michael Frisch) Subject: Re: gaussian wavefunctions To: chemistry@ccl.net Status: R Dawn Friedman asked: I'm a bit mystified by the output of the STABLE keyword in Gaussian 90. Has anyone used this keyword to test the stability of their HF (especially UHF) wavefunctions? Once you've determined that your wavefunction _is_ unstable with respect to one or more factors, and used the stable=opt option to get a better wavefunction, what do you do next? For example, if the unstable wavefunction was the result of an geometry optimization, is it possible to get a better geometry from the better wavefunction in some way? What do you do with the better wavefunction other than admire it? And what do the various forms of instability "mean" in physical terms, if anything? Thanks, Dawn (friedman@tammy.harvard.edu) There are several uses for stability calculations, for example: 1. In symmetric systems, you can check if your wavefunction has the correct (ground state) symmetry. This type of instability is one example of "internal" instability -- a lower energy solution of the same type (real vs. complex orbitals, restricted vs. unrestricted). 2. In systems which are difficult to converge and require methods such as SCF=QC or SCF=DM to force convergence, one can converge to a saddle point (in Fock space) which is nearer to the starting guess than the optimal wavefunction. A stability test checks for this. 3. For a closed-shell systems, and RHF-->UHF instability indicates that the system is not well described in terms of doubly-occupied orbitals and consequently there is a lower-energy UHF solution. Just going to UHF may or may not be sufficient to produce good results -- depending on what properties are being computed, a more sophisticated methods, such as QCISD or MCSCF, may be appropriate. A geometry optimized using a wavefunction with an internal instability is definately suspect, if the wavefunction is found to be unstable at the final structure, the geometry should be re-optimized using the lower energy wavefunction resulting from the stability calculation. This is especially true of complicated open-shell systems, in which there may be several possible UHF solutions, with different solutions being optimal in different regions of the nuclear potential energy surface. Two good references on SCF stability and its implications are: Seeger and Pople, JCP 66, 3045 (1977) and H. B. Schlegel and J. J. McDouall, in "Computational Advances in Organic Chemistry", C. Ogretir and I. G. Csizmadia, eds., Kluwer Academic, the Netherlands, 1991, pages 167-185. Mike Frisch ------- --- Administrivia: The rest of 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 ---