From paul@simulate.chem.vt.edu Fri Jul 2 35:30:54 1993 From: paul@simulate.chem.vt.edu (Paul Gregory) Message-Id: <9307010616.AA13164@simulate.chem.vt.edu> Subject: IBM SCALABLE POWERPARALLEL SYSTEMS To: chemistry@ccl.net (complist) Date: Thu, 1 Jul 1993 02:16:54 +22305714 (EDT) Hi, Does anyone have an IBM Scalable POWERparallel System, the multiprocessor RS/6000? The planned availability date is Sep. 17, but I assume that some of the major IBM clients might get a sneak preview. V. Paul Gregory Department of Chemistry Internet: paul@simulate.chem.vt.edu Virginia Tech ,/\ Blacksburg, Virginia ,/ `\/`'\ 24061-0212 ,/\/ / ,/ <_ ._. / `\ / / ,/\_ _/' `-`-\ |/ ,/' `_'_' * Blacksburg \_> ' ,/' \__>_/\ +----------------------------------------------+ From ZSYAMP01%EBCESCA1.BITNET@phem3.acs.ohio-state.edu Thu Jul 1 10:47:41 1993 Date: Thu, 01 Jul 1993 09:47:41 +0100 (CET) From: ZSYAMP01%EBCESCA1.BITNET@phem3.acs.ohio-state.edu Subject: Summary on MM/MD To: chemistry@ccl.net Message-Id: <01H00IQMGVAA9H1DH3@phem3.acs.ohio-state.edu> Dear netters, Here is the summary about MM/MD packages i promised 4 months ago (I'm sorry for the delay). It has been 'upgraded' with the opinions of my own users. Although I tried to abstract, it is going to be very large, so here are the main points: - All what I'm going to say is for supercomputers - I have not considered quantum mechanics programs (i.e. GAUSSIAN) - Every computational program (running in a supercomputer) needs a good visualization interface (running in a WS) - It is necessary to to consider three main groups of programs: 'small molecules', 'general purpose', and 'large molecules' ones. - Commercial programs: recommended for simulations; expensive; code not available; good support - Academic programs: recommended for basic research; cheap; code available; support not guaranteed And here comes the summary. Alicia Martinez Supercomputation Center of Catalonia - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: Daniel Ripoll There are three major companies in US with molecular modeling packages in the market: Biosym (Insight-Discover), Molecular Simulations (QUANTA), Tripos Inc.(Sybyl). These three packages provide many basic features that are very similar. All af them will allow you to carry out molecular mechanics and dynamics simulations. They differ mainly in the force field they used for the energy calculations. Each of these programs have their own strength that make them suitable for particular type of applications. Among the ones I know, Sybyl (Tripos Inc) is very good on studies of small molecules. The BIOPOLYMER module has many interesting features useful for protein studies. QUANTA (with Charmm) is very good for studies of protein dynamics. These vendors also provide tools like XPLOR that are extensively used in Macromolecular structure determinations using X-ray crystallography and Nuclear Magnetic Resonance (NMR). Discover (Biosym), as I understand, is a very well developed force field and the molecular dynamics program is very robust and fast. For studies on small molecules, there is another important program you should consider named MACROMODEL. This program is being developed by Prof. Clark Still >from Columbia University). It uses the MM3 force field. The code is being sold at very low price for academic users by Columbia University. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: CVINALS@BNANDP51 (...)I'm most satisfied with Discover (Biosym). It's a complet program (along with Insight) adapted to the beginner (user-friendly) as well as to the expertise. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: Willem Vermin (...) I think that QUANTA/CHARMM is a good choice. As you probably know: QUANTA runs on a workstation, CHARMM can run on a cray or ibm mainframe. (...) And , of course, you have to lsiten to the users of your system. If they really want DISCOVER, I think there is only one thing to do: buy it! - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: UGLIENGO@CH.UNITO.IT (...) I strongly recommend GROMOS CHARM and AMBER just to cover the field of biological molecules. We have the BIOSYM package here in Torino running on a SGI35; we joined the Catalysis and Sorption project two years ago so that we got a large discount. Even so, the package cost a lot of money for our standard. It is very well documented and frequently updated. (...) My point of view is that make no sense to buy DISCOVER alone without the graphic package. It is certainly possible to have it running on a CRAY and get the result on a remote INDIGO running INSIGHT. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: "Gabor I. Csonka" The reply to your question: there is a new software called Hypercem which seems to be very good package. Another thing I can suggest is Macromodel. The Sybill force field of Tripos is also worth mentioning. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: Caterina Ghio 1) INSIGHT and DISCOVER are expensive, but they are very friendly to use. In a routine work they are helpful. 2) SYBYL (Tripos Associates), another visualization software, is slightly cheaper, if it is still so. In any case you can ask for a short period trial for both SYBYL and INSIGHT/DISCOVER. Usually for 2 months. 3) Harel Weinstein (Mount Sinai Hospital, CUNY) prefers using CHARMM and QUANTA for small systems, and AMBER/SYBYL for larger ones. 4) MidasPlus is a cheap package of the Computer Graphics Lab of the University of California, San Francisco. It costs about $300. The upgrade is $50. It is a very good program for visualization and can be used with any simulation software, even though AMBER is suggested. It has no pop-up menus, but the commands can be sharply abbreviated. e-mail NORMA at cgl.ucsf.edu 5) Be sure to buy a workstation allowing you to see stereo pictures. When deal- ing with large systems it is impossible to do without. Both MidasPlus and Insight or SYBYL have the stereo option. We have an SGI 4D-GTXB with stereo. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: Max Sybyl has many good functions and is the best for small molecules (it isn't so good for proteins or nucleic acids). (...) Insight/Discover is better than Sybyl for large molecules and n 'solvent box' simulations; the same applies to Quanta/Charmm. (...) A package which seems also good for visualizations id AVS, distributed by Molecular Simulations. SI), - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: ross@cgl.ucsf.edu (Bill Ross ) The main differences I know between Amber and commercial packages are: Amber: - comes as source code, so you can do your own development. - is probably much cheaper (see prices below). - does not have a released graphics interface yet (one is on the way). Meanwhile, a graphics program that reads and writes Protein Data Bank format files is needed (I believe free ones for SGI are obtainable). - is strong in normal mode and molecular mechanics-based methods but has no quantum, docking, qsar, or other different methods incorporated. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: (Dr.) Dave Winkler I'm a 'small molecule' molecular modeller (i.e. drugs and crop protection chemicals). I find the Sybyl software from Tripos Associates, St. Louis very good. (..) It is a very comprehensive suite of programs which do all of the required functions, and, very importantly, the software is very easy to use. They hve one unique advantage over other similar packages such as Biosym and MSI in that they have ComFA analysis (comparitive molecular field analysis (patented)). They also have a very useful molecular spreadsheet which integrates all of the modelling, database and biological data into a very useful form dor analysis (it is easy to 'drown in data' otherwise). Their Unity molecular database package is the only alternative to the widely used but ancient MACCS database. If cost is a factor, then Chemical Design's CHEM-X molecular modelling package is also very good. They also seem to be a little ahead of the other companies in bringing new, innovative design methods into the market place. You can email them too (Chemdesign@applelink.apple.com). - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: Sophie Creuzet C3NI IBM/CNUSC CHARMm (which was developped by Karplus group) was initially done for proteins and is presently the best you can find for simulations concerning proteins. You'll find hundreds of publications using CHARMm and this is not only a proof that it is scientifically valid but a mine of previous calculations very helpfull for people begining in this field. AMBER (developped by Kollman's group) is the next in my list since its story is very similar to CHARMm's one. But it has originally been developed for nucleic acids, and it's probably the best for that. In fact, both Karplus and Kollman decided to develop parameters for proteins and nucleic acids so today people are using them almost indifferently. However AMBER is almost free and CHARMm rather expensive, I think CHARMm is more popular. I think it's a question of functionalities as well. Moreover, you have a good interface for CHARMm (QUANTA) which doesn't exist for AMBER. Since a graphical interface is very important when you deal with macromolecules this is also a reason why CHARMm is prefered. DISCOVER (developed and sold by BIOSYM) seems to become more and more popular. I can't say anything concerning it's reliability since the number of paper using it is really weak compared to CHARMm or AMBER. The interface (INSIGHT which is among the best) seems to be one of its trump cards. Finally, MSI as well as BIOSYM have very good package for RMN for example, which is also a hudge consumer of computing power. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: "T. Mark Reboul" Inexpensive but useful alternatives you might consider are: MacroModel -- mmod@cuchmc.chem.columbia.edu -- for modeling Grasp -- nicholls@cuhhca.hhmi.columbia.edu -- for visualization These are non-commercial but well-maintained products which are fairly inexpensive, maybe ~$500 each. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: dobbskd@esvax.dnet.dupont.com (...) The best approach is to have a good understanding of the problems that you want to solve over the next few years and then talk to other computational people about the best way to tackle these problems with existing software. It would not be a good idea to try to buy the "best" software (all simulation software falls in this category depending on who you talk to), install it, and then expect your users to do "wonders" with the software (this scenario is greatly pushed by your more visible software vendors). - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: nauss@wrair-emh1.army.mil I have extensive experience with QUANTA and CHARMM. Generally, CHARMM is an adequate program to use. It is very flexible once one knows the commands. (...) I would recommend CHARMM as a stand alone program. QUANTA, however, has problems: (...) the dialog boxes (...) changing the display (...) MSI's hotline (...) Their documentation (...) After having used QUANTA for three years now, I am not pleased with it. I have used an old version of SYBYL from Tripos. I was not pleased with it either as the organization of menus was not intuitive. (...) What I have seen of INSIGHT and DISCOVER I like. The menus have a good feel to them, selections are generally where I expect them to be. Also, INSIGHT seems to be better integrated with the operating system. (...) As far as how powerful DISCOVER is, I am not certain. However, Biosym seems to document their product much better than MSI. Also, Biosym has published their tests of the DISCOVER force field in the open literature. MSI has not although the original CHARMM from Harvard is well documented. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: SOARES@XRAY.BMC.UU.SE If simulation techniques are an auxiliary tool for your work, I would suggest a commercial package, like Discover, Hyperchem, Prosimulate, etc. If however, people are going to do more fundamental research on the field of Molecular dynamics, an open code would be more adequate. (...) In this latter group I would recommend GROMOS or AMBER. We are using GROMOS here in my group and for visualization of trajectories we use the program mdFRODO, which accepts GROMOS outputs. The mdFRODO program is distributed by Dr. Olle Nilsson, and with it is also distributed the NICE package which performs numerical analysis of MD results. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: F0418950@JAGUAR.CSC.WSU.EDU Having used INSIGHTII/DISCOVER for about 3 years now I conditionally recommend it as a good molecular mechanics/dynamics package. (...) I would suggest looking in depth at the capibilites that it has and see if they fit your needs, AS THEY EXIST IN THE PROGRAM ALREADY. If your center plans to go beyond the choices the program gives and develop your own algorithms or install new ones, Insight/Discover is not a good choice (in my opinion). It is too much of a packaged program and adding things to it is very hard. CHARMM on the other hand, is harder to use but extremely adaptable. (...) For research and development, look at CHARMM. For routine use, look at Insight/Discover. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: MARTIN%CMDA I strongly suggest that you get a copy of SYBYL for your SG's. We have just moved from in-house molecular graphics to SYBYL & find that the Windows-based environment & the Sybyl Programming Language make it very easy to interface applications & keep everything straight. - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: We are using AMBER and I think it needs a user-friendly environment to manipulate the results: MIDAS (UCSF) QCPE 623 "MD Display". QCPE 627 "X11-based AMBER MD viewing/analysis program". (...) Maybe Quanta/Charmm would be good to invest in because AMBER works basically with proteins, whilst CHARMM is also good for small molecules (of course it's much more expensive). (...) - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - From: JSCHELDE@STEEVAST.RUG.AC.BE I'm planning to buy Charmm because of its flexibility but I will first try it out. From CHEM8@VAXA.YORK.AC.UK Thu Jul 1 06:25:14 1993 Message-Id: <199307011009.AA16561@oscsunb.ccl.net> Date: Thu, 1 Jul 93 11:09 BST From: "John Waite, Tel 1-7238958, N.H.R.F., Vas. Konstantinou 48, Athens 116-35" To: CHEMISTRY Subject: IMSL and NAG Hi Netters, I would appreciate comments concerning the merits and disadvantages of the computer library systems IMSL and the general NAG library. In particular regarding their ease, or difficulty, in use, their scope (maybe separated into maths. and stats.), their installation, availability of updates, their accessibility and other comments. John Waite CHEM8@VAXA.YORK.AC.UK From CHEM8@VAXA.YORK.AC.UK Thu Jul 1 07:25:01 1993 Message-Id: <199307011025.AA16769@oscsunb.ccl.net> Date: Thu, 1 Jul 93 11:25 BST From: "John Waite, Tel 1-7238958, N.H.R.F., Vas. Konstantinou 48, Athens 116-35" To: CHEMISTRY Subject: FFT formulae for Bxyz and Gxyzz? Hi netters, Does anyone have, or know where to get, either or both code and formulae for 5-point central difference calculation of the triple cross-terms of third and fourth order cartesian tensors? In particular I am interested in calculating, using the Finite Field method, the electric hyperpolarizability components Bxyz and all terms of the form Gxyzz. Hopefully, John Waite chem8@vaxa.york.ac.uk From stewart@mol.ucns.uga.edu Thu Jul 1 02:04:26 1993 From: stewart@mol.ucns.uga.edu (David Stewart) Message-Id: <9307011104.AA19686@mol.ucns.uga.edu> Subject: Re: IBM SCALABLE POWERPARALLEL SYSTEMS To: chemistry@ccl.net Date: Thu, 1 Jul 1993 07:04:26 -0500 (EDT) > > V. Paul Gregory wrote: > Hi, > > Does anyone have an IBM Scalable POWERparallel System, the multiprocessor > RS/6000? The planned availability date is Sep. 17, but I assume that > some of the major IBM clients might get a sneak preview. > > The only site that has one installed is the Cornell Theory Center. I believe that there is only one other one in existence, and that is within IBM. Or so it seems. David Stewart Univ of Georgia From adfernan@undergrad.math.uwaterloo.ca Thu Jul 1 05:48:44 1993 From: "Andrew D. Fernandes" To: CHEMISTRY@ccl.net Subject: reproducibility, codes, and distribution Message-Id: <93Jul1.094847edt.43213-3@undergrad.math.uwaterloo.ca> Date: Thu, 1 Jul 1993 09:48:44 -0400 An interesting question arises when one considers not only the possibility that the authors of a particular paper do not disclose enough information to completely reproduce their results, but the fact that most software authors do not disclose enough to verify that what their program does is what it's supposed to do. Take, for example, Gaussian92 (not that I have any bones to pick with that program, it's just that I have a source code line count for it!). The system consists of ca. 350_000 lines of code, a fair bit of it from revisions that go back to ... when? Short of writing a completely different software package that does exactly the same thing as g92, what guarantee do we have that G92 (or any of Biosym's or AutoDesk's products) actually do what they are supposed to? Even having the source code just doesn't help... who wants to root through 350_000 lines of someone else's code for any reason? Now think back to the last calculation you ran that just didn't quite work for "some reason or another". Was it your system, or a somewhat flakey algorith-to-program translation? It seems that any program intended publication should have some sort of formal verification that it is doing exactly what it should be doing. Some programming languages encourage this sort of thing; others discourage it. But to write modularly-verifyable code is not an easy thing. -Andrew. Andrew D. Fernandes, Deptartment of Chemistry University of Waterloo From skeinath@ncsa.uiuc.edu Thu Jul 1 04:19:58 1993 Date: Thu, 1 Jul 93 09:19:58 CDT From: skeinath@ncsa.uiuc.edu (Steven E. Keinath) Message-Id: <9307011419.AA05314@saturn.ncsa.uiuc.edu> To: CHEMISTRY@ccl.net Subject: Force fields for graphite Dear Netters: Does anyone know of literature reference(s) to a good force field for treating the graphite crystal? There appears to be favorable pi-orbital overlap between crystal layers, and our experinece up this point has been that the more standard force fields cannot predict the graphite crystal structure. Marc Mansfield and Steve Keinath Michigan Molecular Institute Midland, MI 48640 USA From rc_davis@srs.gov Thu Jul 1 06:38:27 1993 Date: Thu, 1 Jul 1993 10:38:27 -0400 (EDT) From: "Ricardo C. Davis" Subject: chemistry references (fwd) To: Computational Chemistry List Message-Id: Some info passed on from a fellow collegue here at SRS. Another reason why network access is beneficial to science! Ricardo C. Davis Internet: rc_davis@srs.gov Building 773-42A, SRTC/SCS, Westinghouse Savannah River Co. P.O. Box 616, Aiken, SC 29802 (803)725-5172 ---------- Forwarded message ---------- >Date: Wed, 30 Jun 93 14:40:09 PDT >Reply-To: net-happenings%is.internic.net@internet.srs.gov >Originator: net-happenings@is.internic.net >Sender: net-happenings%is.internic.net@internet.srs.gov >From: Gleason Sackman > >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >Forwarded by Gleason Sackman, net-happenings moderator >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > >---------- Text of forwarded message ---------- >Date: Wed, 30 Jun 1993 11:46:42 CDT >From: harkanyi k >To: Multiple recipients of list PACS-L >Subject: chemistry references > >----------------------------Original message---------------------------- >Dear Netters, > >As of June 28, 1993, a new chemistry library resource "Chemistry >Information" is available for your use in ftp-able form. > >"Chemistry Information" is an electronic reference source. Its >purpose is to provide solutions to frequently occuring chemistry >questions through the use of library resources. It is a valuable >reference tool, especially when librarians are not on duty. Its >audience is the undergraduate student; however some of the >information can also be beneficial to graduate students. Not only >chemistry but also engineering, physics, nursing, nutrition, and >health science students can utilize this resource. > >"Chemistry Information" covers nomenclature, compound >identification, properties, structure determination, toxicity, >synthesis, registry numbers, and synthesis. For each component it >lists the most appropriate reference resources (online catalog, >databases, indexes, journals, government documents, handbooks, >tables, and/or treatises). > >"Chemistry Information" also provides definition of terms, meanings >of acronyms, full form of abbreviations, call numbers, and location >of materials. > >How to acquire "Chemistry Information"? Ftp to ucssun1.sdsu.edu; >directory pub/chemras; use binary; use mget command and the * wild >card symbol to transfer all files in the chemras directory (mget >*.*). Please read the read.me file first. > >Since the contents are San Diego State University library >resources, feel free to adapt them to your collection. If you have >any comments, please direct them to the author Katalin Harkanyi >(harkanyi@ucssun1.sdsu.edu) after Labor Day. > >If you have any difficulty downloading the material or with the >programming, please contact Robert Carande >(carande@ucssun1.sdsu.edu) after Jyly 31st. > >Enjoy! > Katalin > >Katalin Harkanyi >Chemistry and Engineering Librarian >University Library, Science Division >San Diego State University >(harkanyi@ucssun1.sdsu.edu) From ZPZCM@jazz.ucc.uno.edu Mon Jul 1 04:01:23 1993 Date: 01 Jul 1993 10:01:23 -0600 (CST) From: "ZORAN P. ZDRAVKOVSKI" Subject: Supercomputer time grants To: chemistry@ccl.net Message-Id: <01H00VPUTULI8WY0G8@jazz.ucc.uno.edu> Hi netters, We are presently doing transition state calculations of some cyclic polar additions using semi-empirical and ab initio methods. The semi-empirical results seem to correlate well with known experimental data. We are using a PC and Silicon Graphics computer with little memory and disk space for the ab initio caluclations, so we are confined to RHF / STO-3G and 3-21G theory and basis set calculations. We would like to submit these to higher basis set and theory calculations, so the polar nature would be better described than with the theory level we are using. Does anyone know whether we can use any of the supercomputer centers for such calculations? Are there any "supercomputer time" grants available? All suggestions and comments are greatly appreciated. Zoran Zdravkovski Department of Chemistry University of New Orleans New Orleans, LA 70148 (504) 286-7090 fax (504) 286-6860 e-mail ZPZCM@UNO.EDU From cramer@staff.tc.umn.edu Thu Jul 1 06:17:06 1993 Message-Id: <0012c330e03019921@staff.tc.umn.edu> From: "Christopher J Cramer-1" Subject: Fullerenes To: chemistry@ccl.net Date: Thu, 1 Jul 93 11:17:06 CDT I would be grateful to receive Cartesians or Z-matrices for any and all fullerene isomers, in particular C60, C70, C76, and C78. Should a site not already exist with this information, I'll be happy to compile it and give it to Jan for ftp purposes. Many thanks to respondents, who will save me hours of drawing. CJC -- Christopher J. Cramer University of Minnesota Department of Chemistry 207 Pleasant St. SE Minneapolis, MN 55455-0431 (612) 624-0859 From case@scripps.edu Thu Jul 1 02:16:48 1993 Date: Thu, 1 Jul 93 09:16:48 PDT From: case@scripps.edu (David Case) Message-Id: <9307011616.AA16063@lanczos.Scripps.EDU> To: chemistry@ccl.net Subject: Re: reproduciblity, codes, and disclosure mercie@cumc.cornell.edu wrote: > > Most individuals that wish to receive source code probably fall in three > categories: 1) code scavengers, 2) code "extenders". My candidates for the missing "third category": People who (1) want to know what the code "really" does, as opposed to what the documentation says it does; (2) would like to run programs on machines other than those supported by the developers; (3) want to make sure that they can continue to do research after buying a new machine or upgrading their operating system; (4) find the documentation confusing and need to see what input the code is really expecting to find, or how it is used; (5) wish to obtain more detailed output than the program would ordinarily provide; (6) encounter bugs and wish to correct them; (7) wish to publish results using the code in question. In short, there are lots of folks besides "scavengers" and "extenders" who have a legitimate interest in having access to source code. ...dave case From shepard@dirac.tcg.anl.gov Thu Jul 1 07:58:49 1993 Date: Thu, 1 Jul 93 12:58:49 CDT From: shepard@dirac.tcg.anl.gov (Ron Shepard) Message-Id: <9307011758.AA01594@dirac.tcg.anl.gov> To: chemistry@ccl.net Subject: Re: IBM SCALABLE POWERPARALLEL SYSTEMS >V. Paul Gregory wrote: > >> Hi, >> >> Does anyone have an IBM Scalable POWERparallel System, the multiprocessor >> RS/6000? The planned availability date is Sep. 17, but I assume that >> some of the major IBM clients might get a sneak preview. >> >> > The only site that has one installed is the Cornell Theory Center. I > believe that there is only one other one in existence, and that is > within IBM. Or so it seems. > > David Stewart > Univ of Georgia I believe that there are several test sites. Argonne is one of them, as is Cornell. The Math and Computer Science Divison at Argonne just hosted a workshop for interested users for its machine. As you would expect, it is straightforward to port codes to the machine that currently run on networked workstations. P4, TCGMSG, Chameleon, and some of the other usual parallel development tools have been ported already; other tools, including MPI, are in the works. Configurations consist of up to 128 nodes, each of which is essentially a 370 workstation (i.e. 120 MFLOPS peak). The ANL machine currently has 32 nodes, each of which has 128MB of memory and 1GB of local disk space. In addition to ethernet connections between the nodes, there is an "omega switch" (which was just recently installed at ANL) which has higher bandwitch and lower latency (the actual performance numbers change daily). The hardware looks interesting. Hans Lischka and I hope to have parts of the COLUMBUS MRSDCI code running on it this summer, along with some direct-SCF and PT codes that are being developed with Robert Harrison, Rick Kendall, and some other folks at PNL. Regards, Ron Shepard shepard@tcg.anl.gov From jle@world.std.com Thu Jul 1 13:58:38 1993 Date: Thu, 1 Jul 1993 17:58:38 -0400 From: jle@world.std.com (Joe M Leonard) Message-Id: <199307012158.AA12004@world.std.com> To: chemistry@ccl.net Subject: Stability testing of codes... I don't know the specifics of anybody else, but I'm sure everybody does just about the same thing... Regarding how folks know that codes work - they're tested against test suites that have been accreting "since the beginning of time". Consistancy checks are a means of verifying that version n is "correct", in that it reproduces the results of version n-1, or the set of experimental data, or... Computer science authors have been writing the last 20+ years about how programs cannot be " programs cannot be proven correct, etc. All we can hope for is that significant bugs will be found by testing and what's not found is not significant. Then there's the tried and true method of sending stuff out and waiting for the yell(s) from the customer base. Works every time :-) You should see some of the crap floating about the net regarding PC and MAC games and what those folks think about upgrade policy and testing... Joe Leonard jle@world.std.com From afj@chem.ucla.edu Thu Jul 1 10:37:24 1993 Date: Thu, 1 Jul 93 17:37:24 -0700 From: afj@chem.ucla.edu (Andy Jacobson) Message-Id: <9307020037.AA26045@uclachem.chem.ucla.edu> To: chemistry@ccl.net Subject: Looking for info on twisted charge transfer states. Hi Folks, Can anyone out there give me any leads on MD calculations relating to twisted intramolecular charge transfer (TICT) states? I'm particularly interested in cases of solvent effects on the interconversion between twisted and planar states in aryl-NR2 groups. Literature references would be a great start. Thanks to all, -A.J. Andy Jacobson UCLA Dept. Pharmacology UCLA School of Medicine Los Angeles, CA 90024-1721 From raman@bioc01.uthscsa.edu Thu Jul 1 14:33:27 1993 From: raman@bioc01.uthscsa.edu (C.S.RAMAN) Message-Id: <9307020033.AA02017@bioc01.uthscsa.edu> Subject: Re: Fullerenes To: cramer@staff.tc.umn.edu (Christopher J Cramer-1) Date: Thu, 1 Jul 1993 19:33:27 -0500 (CDT) > > I would be grateful to receive Cartesians or Z-matrices for any and all > fullerene isomers, in particular C60, C70, C76, and C78. Should a site not To subscribe to the Buckminster Fullerene mailer send mail with null subject field and the body of the text containing the word "INTRO" or "HELP" to: bucky@sol1.lrsm.upenn.edu This exploder is really fantastic if you are interested in Fullerenes and contains a huge bibliography with references dating back to B.C. :) To get this file, send the keyword "BIBLIO" to the same address. This service is maintained by Jack Fischer's group at UPENN. 2. There is an anon FTP site devoted to this topic and is maintained by the Arizonal fullerene consortium: physics.arizona.edu login:ftp passwd: 3. There is a new journal called "Journal of Fullerene Science and Technology" whose first issue rolled out in Feb. 1993. 4. The library at U. Arizona maintains a huge database of fullerene related papers and books published todate; telnet sabio1.library.arizona.edu login: sabio choose other databases from the menu and follow the simple directions. The FTP site mentioned above (#2) does have a program called PCBIB that allows one to search through the database for specific keywords and you might want to download this. Currently it works on PC platforms. Hope this information is useful to you! Cheers -raman -- C.S.Raman raman@bioc01.uthscsa.edu - Internet UNIX Programming & Administration 70412.2354@compuserve.com - CIS SPARC & SGI Systems raman@hermes.chpc.utexas.edu - CHPC Department of Biochemistry craman@launchpad.unc.edu UTHSCSA 7703 Floyd Curl Dr. (210) 567-6623 [Tel] San Antonio, TX 78284-7760 (210) 567-6595 [Fax] ****************************************************************************** If a man's wit be wandering, let him study the Mathematics -Francis Bacon ******************************************************************************