From chemistry-request@ccl.net Fri Aug 23 11:12:46 1991 Date: Fri, 23 Aug 91 11:03:24 EDT From: G. Ravishanker To: R2VJV%AKRONVM@vm1.cc.UAKRON.EDU Subject: IBM-6000 ... Status: R Hi There was already a message explaining the various features of an IBM RISC-6000. I thought I would add a little bit more to it. We at Wesleyan moved to RISC 6000's a year ago, primarily because we needed machines with a lot of floating point performance to run our MD programs. We are very happy with the performance of the Powerstation 530's that we have. We use Polygen Quanta, which now has moved upto a release version from the Beta version. Most of the other commercial Molecular Modeling packages are now available for the RS6000. The graphics on the RISC is noticeably slower than the same on an IRIS 4D-GT/70, but not too different compared to a Personal Iris. Most of our programs (graphics as well as non-graphics) has been succesfully ported with minor modifications. The font-manager library available on Silicon Graphics is not supported on the RISC machine, but the font access goes through calls to X-windows based routines. Other than that and a couple of minor differences, the Silicon Graphics GL based programs work just fine on the IBM RS6000. The current price reductions (especially on Powerstation 550) makes these machines very desirable. The OS is stable for the kind of things we do. Networking is OK. We still have some problems in this area and we are in the process of narrowing down whether it is the IBM RS6000 or the PC gateway that is causing the problem. We are currently testing a beta version of VNMR (2-D and 3-D NMR data processing software from Varian) on the RS6000 and the fourier tranformation goes at unbelievable speed. We have 3 machines, mostly running MD simulations and analysis. We are likely to acquire two more one for a Quantum Chemistry group and one for departmental use. Ravi From chemistry-request@ccl.net Fri Aug 23 11:16:38 1991 Date: Fri, 23 Aug 91 11:07:26 EDT From: G. Ravishanker To: Leif.Laaksonen@vtt.fi Status: R Hi We use the routine called LSQSTR from GROMOS distribution to compare structures. I have cut the comments from the subroutine and am attaching it below: COMMENT LSQSTR ROTATES THE ATOMS WITH COORDINATES X ABOUT THE ORIGIN C C SUCH THAT THE FUNCTION E = 0.5 * SUM OVER ALL NR ATOMS OF C C W*(X-XP)**2 HAS A MINIMUM. HERE W DENOTES THE WEIGHT FACTORS AND C C XP ARE THE REFERENCE COORDINATES. FOR A DESCRIPTION OF THE C C METHOD SEE A.D. MCLACHLAN, J. MOL. BIOL. 128 (1979) 49. C C IF THE SUBROUTINE FAILS, IT IS RETURNED WITH A MESSAGE AND C C IROT=0. C The paper referenced here has an appendix explaining the mathematics of the proposed algorithm. Ravi From chemistry-request@ccl.net Fri Aug 23 12:21:12 1991 From: rs0thp@rohmhaas.com (Dr. Tom Pierce) Subject: Re: IUPAC and Theoretical/Computational Chemistry To: states@ncbi.nlm.nih.gov (David States) Date: Fri, 23 Aug 91 11:53:25 EDT Status: R > > Dan Calef writes: > > "What could the IUPAC do to help theoretical chemistry?" > > Davis States replies: > Standard data description semantics would be a good to place start. > Ideally we would like to be able to pipe the output of one program > directly into another program without losing data or accuracy. To > exchange data between programs (or databases) it is necessary to agree > on both the syntax for the data and a set of semantics. > > .... > > The hard part of the problem is agreeing on semantic rules. When I say > the energy was x, do I mean calories (gram), calories (kg), kcal/mole, > joules, ergs, electron volts ... If units are included, do kcal/mole, > Kcal/mole, kcal/Mole, KCAL/MOLE all mean the same thing? How do you > describe an electron wavefunction? When a program expects kcals/mole, > and it gets ergs, the results are not going to be pretty. > > Basically folks, I think this is an idea who time has NOT come. Designing standards especially using a Standards Setting Committee (IUPAC) is something that should be attempted after competing standards have been developed, not before we know what should be standardized and what should be ignored. I see two sets of standards(methods) that have not yet been developed. One standard would be needed for submitting articles to journals. In this case I think that the people who publish journals should be out front on this and design standards that would work. The second set of standards that are not yet developed are how data is stored and communicated between programs. Admittedly there are competing methods. But we can chose to use/write interfaced to MOPAC or G8X files or to Molecular Design's format or many other formats. Some of the data is similar between these formats, (XYZ) but data fields are meaningless such as commands for a particular program. The need for standards today simply does not exist in Computational Chemistry. Perhaps it exists between the database vendors (CHem Abstracts/Molecular Design ), but I think it is too early to start imposing standards. -------------------------------------------------------------------- Sincerely, Thomas Pierce Computer Applications Research | rs0thp@rohmhaas.com Internet < Keep your transactions separate from your relationships - Safe Computing!> From chemistry-request@ccl.net Fri Aug 23 12:49:59 1991 Date: Fri, 23 Aug 91 11:34:36 CDT From: rowlan@nyssa.ctd.anl.gov To: chemistry@ccl.net Subject: Molecular Modeling Toolkits Status: R X-Office Phone: (708) 972-7587 X-Graphics Lab Phone: (708) 972-5776 Date: Fri, 23 Aug 91 11:34:36 -0500 From: rowlan@nyssa.ctd.anl.gov Please Help, Does anyone know of any AVS molecular modeling modules or any similar vendor or public-domain software ? Please respond directly to rowlan@mcs.anl.gov Mega Thanks, --------------------------------------------------------------------- John Rowlan Phone: (708) 972-7587 Scientific Visualization Facility Email: rowlan@mcs.anl.gov Argonne National Laboratory Fax: (708) 972-5986 --------------------------------------------------------------------- From chemistry-request@ccl.net Fri Aug 23 15:05:02 1991 Date: Fri, 23 Aug 1991 14:52 EDT From: "DOUGLAS A. SMITH" Subject: Re: Molecular Modeling Toolkits To: rowlan@nyssa.ctd.anl.gov Status: R AVS (Applications Visualization Software) is a Stardent Computer, Inc. software system which is now licensed on many platforms, including DEC, Cray, Convex, and IBM. Commercial development of AVS tools, including the Chemistry Viewer, is being done by Molecular Simulations Inc. of Sunnyvale, CA. Their various addresses are included at the bottom of this message. Non-commercial AVS tools for chemistry are being handled in several fashions. The North Carolina Supercomputer Center is going to be the national AVS depository, but I don't have a name for a contact there yet. There are also 5 beta test sites for the AVS Chemistry Developer's Kit, which allows one to build new AVS modules. These sites include: Molecular Simulations, Inc., Kubota (Japan), Scripps, and the University of Toledo (i.e. me). I don't recall the fifth site, but it might be Batelle. the person at Stardent in charge of Chemistry and AVS is Joe Leonard, who appears frequently on the Chemistry mailer. He can also be reached at jle@stardent.com. I will be happy to talk with anyone about AVS in more detail. Right now, my group is working on converting our file format interconversion program FITIT (recently announced on this mailer - or send mail to me for more information) into AVS. We are also working on a PHI-PSI driver/grapher with links to several modeling packages. Doug Smith Assistant Professor of Chemistry The University of Toledo Toledo, OH 43606-3390 voice 419-537-2116 fax 419-537-4033 email fax0236@uoft02.utoledo.edu Molecular Simulations, Inc. Corporate Headquarters: 796 North Pastoria Ave. Sunnyvale, CA 94086 (408)732-9090 -- ask for Mike Rivers, VP for sales (408)732-0831 -- fax Eastern Regional Sales Office: 214 Carnegie Center Suite 207 Princeton, NJ 08540 (609)520-0877 -- ask for Glen Churchfield (609)520-0240 -- fax From chemistry-request@ccl.net Fri Aug 23 15:33:15 1991 Date: Fri, 23 Aug 91 15:10:15 EDT From: George R Famini To: chemistry@ccl.net Subject: Re: IUPAC and Theoretical/Computational Chemistry Status: R wait a minute... isn't this exactly what ASTM is trying to do? I believe they have a subgroup on standardized molecular data. Also, there's a British group doing the same thing... I disagree with Tom Pierce, the time is now for standards to at least begin evolving. It sure would make it easier for those of us that use lots of different computational chemistry packages to readily move between them. I don't think it'll happen anytime soon (the ASTM group has been meeting for several years, I think), but it sure would be nice. George Famini Life is like a ferrari, it goes too fast. Of course, you probably couldn't afford it anyway-- Garfield. From chemistry-request@ccl.net Fri Aug 23 19:43:19 1991 Date: Thu, 22 Aug 91 22:59:10 EDT From: Joe Leonard To: chemistry%ccl.net%apgea.army.mil@crdec7.apgea.army.mil Subject: GAMESS users on Titans Status: R This time I'm looking for anybody who's using GAMESS on Stardent 750/3000's (the P3 processors). This is the NDSU version, not the Daresbury (UK). Please contact me at the address below, or Jim Jensen at jojensen@apgea.army.mil Thanks in advance!@ Joe Leonard Stardent Computer jle@Stardent.COM From chemistry-request@ccl.net Fri Aug 23 20:04:06 1991 Date: Fri, 23 Aug 91 18:43 CDT From: Subject: IUPAC & theoretical chemistry To: chemistry@ccl.net Status: R I'd like to come down on the side of those advocating *someone's* working on standardizing aspects of computations. I agree with those who say that it is time to do so with input formats: it seems every time I need to do another type of computation there is a large learning curve for the input. I sometimes get the impression they are deliberately being obtuse, so as to prevent people from doing computations. The writers of GAMESS are to be commended in having added to their latest version the capability to use MOPAC or Gaussian style input. I also agree on the units used on output! Nobody likes mks, but it *is* a standard, so how about reporting output in mks units, in addition to the authors' favorite units, whatever they may be? Irene Newhouse