From ldw@pchindigo2.IPC.PKU.EDU.CN Wed Jan 4 02:33:44 1995 Received: from pchindigo2.IPC.PKU.EDU.CN for ldw@pchindigo2.IPC.PKU.EDU.CN by www.ccl.net (8.6.9/930601.1506) id CAA21803; Wed, 4 Jan 1995 02:02:04 -0500 Received: by pchindigo2.IPC.PKU.EDU.CN (920330.SGI/911001.SGI) for chemistry@ccl.net id AA05390; Mon, 4 Jul 94 14:52:15 +0800 Date: Mon, 4 Jul 1994 14:52:14 +0800 (SST) From: Lin Dawei X-Sender: ldw@pchindigo2.IPC.PKU.EDU.CN To: ccl Subject: Ask for recent reviews Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear Netters, I want to know recent advance of Simulate Annealing, Molecular Mechanics, Molecular Dynamics, and Monte Carlo methods. Can anyone tell me where can I find recent papers about them? Reviews are the best. If any thing is interesting, I will give a summary. Thanx in advance, Dawei ------------------------------------------------------------------- Dawei LIN Room 115, Building 27 Peking University Beijing 100871 Doctoral Candidate of P.R.China Chemistry Department ldw@pchindigo2.ipc.pku.edu.cn Dawei LIN tel (86)-1-2501490 Institute of Physical Chemistry fax (86)-1-2501725 Peking University Beijing 100871 P.R.China -------------------------------------------------------------------- From Patrick.Bultinck@rug.ac.be Wed Jan 4 04:34:18 1995 Received: from mserv.rug.ac.be for Patrick.Bultinck@rug.ac.be by www.ccl.net (8.6.9/930601.1506) id DAA23720; Wed, 4 Jan 1995 03:41:55 -0500 Received: from allserv.rug.ac.be by mserv.rug.ac.be with SMTP id AA17021 (5.67b/IDA-1.5 for ); Wed, 4 Jan 1995 09:40:59 +0100 Received: by allserv.rug.ac.be (5.0/SMI-SVR4) id AA19128; Wed, 4 Jan 1995 09:41:28 +0100 Date: Wed, 4 Jan 1995 09:41:27 +0100 (MET) From: Patrick Bultinck To: "comp. chem. list" Subject: CPU time vs Total time on RS6000 Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Content-Length: 666 Netters, I have noticed a couple of times now that when I run GAMESS on an RS6000 machine I get quite awful ratios of the CPU time versus the TOTAL time for ab initio calculations using GAMESS (conventional SCF, integral storage). Do other people have this problem too (I'm running only one job at a time, one window active so I'm not doing anything but the calcultion) ? Is this present with other programs too if implemented on RS6k ? Is this a typical RS6k problem ? Last but certainly not least : Can I do something about it ??? Thanks for replies, Send for a summary if you're interested in the replies, Patrick Bultinck, University of Ghent, Belgium From beck@work2.ch-cip.Uni-Koeln.DE Wed Jan 4 05:33:42 1995 Received: from work2.ch-cip.Uni-Koeln.DE for beck@work2.ch-cip.Uni-Koeln.DE by www.ccl.net (8.6.9/930601.1506) id EAA24191; Wed, 4 Jan 1995 04:48:22 -0500 Received: by work2.ch-cip.Uni-Koeln.DE (5.65/CL/1.1) id AA14517; Wed, 4 Jan 1995 10:48:13 +0100 From: beck@work2.ch-cip.Uni-Koeln.DE (Michael Beck tel. 4816) Message-Id: <9501040948.AA14517@work2.ch-cip.Uni-Koeln.DE> Subject: parameters for trans.metals To: CHEMISTRY@ccl.net (mailing list) Date: Wed, 4 Jan 1995 10:48:12 +0100 (MET) Cc: reben@work2.ch-cip.Uni-Koeln.DE (Rupert Rebentisch Tel. 0221/470/4816) X-Mailer: ELM [version 2.4 PL23] Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Content-Length: 374 Dear netters! Has anybody heard about semiempirical NDDO-parameters (AM1, PM3) for transition metals? In particular I'm interested in Chromium- hexacarbonyl and related carbonyl-complexes. I would like to use MOPAC (6, 7, or 93) for the calculation, but if there's an other code available it would be OK. Thanks in advance M.E. Beck Cologne From noy@tci002.uibk.ac.at Wed Jan 4 06:33:44 1995 Received: from uibk.ac.at for noy@tci002.uibk.ac.at by www.ccl.net (8.6.9/930601.1506) id FAA24517; Wed, 4 Jan 1995 05:42:45 -0500 Received: from tci002.uibk.ac.at by uibk.ac.at with SMTP id AA03882 (5.65c/IDA-1.4.4 for ); Wed, 4 Jan 1995 11:42:32 +0100 Received: by tci002.uibk.ac.at (AIX 3.2/UCB 5.64/CFIBK-2e.AIX) id AA25886; Wed, 4 Jan 1995 11:42:29 +0100 From: noy@tci002.uibk.ac.at (Teerakiat Kerdcharoen) Message-Id: <9501041042.AA25886@tci002.uibk.ac.at> Subject: What makes Gaussian fast ? To: chemistry@ccl.net Date: Wed, 4 Jan 1995 11:42:29 +0100 (NFT) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 1411 Dear cyber-chemists, I have experienced using both GAMESS and Gaussian 92 on AIX platform. Gaussian 92 always runs faster than GAMESS in most case. I have found on a review article where written Gaussian spend the computation at the cost of N**2.7 not N**4 as conventional Hartree-Fock scheme (N = no. of Gaussian primitives). Which special algorithms are tailored into Gaussian ? Could somebody point me out a hint or direct me the the proper references I could read through in more detail. Thanks a lot for any responses. take care, Teerakiat PS. The energy obtained from GAMESS is always lower than from Gaussian. Therefore, I believe that the speed is gained at the sacrifice of some precision. ---------------------------------------------------------------------------- Teerakiat Kerdcharoen (NOY) Institute of General and Inorganic Austrain-Thai Center for and Theoretical Chemistry Chemical Education & Research Innrain 52a, A-6020 Innsbruck Department of Chemistry AUSTRIA Chulalongkorn University Bangkok 10330 THAILAND ----------------------------------------------------------------------------- *** Love is easy to start but hard to stop From gl@ccl.net Wed Jan 4 08:33:44 1995 Received: from coil for gl@ccl.net by www.ccl.net (8.6.9/930601.1506) id IAA25467; Wed, 4 Jan 1995 08:16:41 -0500 Received: by coil (931110.SGI/930416.SGI.AUTO) for CHEMISTRY@ccl.net id AA09706; Wed, 4 Jan 95 14:15:19 +0100 From: "Gerald Loeffler" Message-Id: <9501041415.ZM9704@coil.mdy.univie.ac.at> Date: Wed, 4 Jan 1995 14:15:13 +0100 In-Reply-To: Patrick Bultinck "CCL:CPU time vs Total time on RS6000" (Jan 4, 9:41am) References: X-Mailer: Z-Mail (3.1.0 22feb94 MediaMail) To: Patrick Bultinck , CHEMISTRY@ccl.net Subject: Re: CCL:CPU time vs Total time Content-Type: text/plain; charset=us-ascii Mime-Version: 1.0 On Jan 4, 9:41am, Patrick Bultinck wrote: > Subject: CCL:CPU time vs Total time on RS6000 > Netters, > > I have noticed a couple of times now that when I run GAMESS on an RS6000 > machine I get quite awful ratios of the CPU time versus the TOTAL time > for ab initio calculations using GAMESS (conventional SCF, integral > storage). Do other people have this > problem too (I'm running only one job at a time, one window active so I'm > not doing anything but the calcultion) ? Is this present with other > programs too if implemented on RS6k ? Is this a typical RS6k problem ? > Last but certainly not least : Can I do something about it ??? > > Thanks for replies, > > Send for a summary if you're interested in the replies, > > Patrick Bultinck, University of Ghent, Belgium >-- End of excerpt from Patrick Bultinck Dear Patrick, without knowing AIX to any noteworthy extent, I can tell you the following: If you time your application (GAMESS) with time or timex you should at least get: 1) real time (i.e. wallclock time) 2) user time (i.e. CPU time used for the non-kernel-portion of your application, which usually is what you are interested in) 3) system time (i.e. CPU time used for satisfying requests to the UNIX- kernel made by your application). The sum of user time and system time is the total CPU time spent on behalf of your application. If your application is doing CPU-intensive tasks (like GAMESS is, without any doubt), then the system time should be negligable and the user time should nearly be as high as the real time (given that you are essentially alone on the machine). If your system time on the other hand is unusually high - as seems to be the case with you application - then the application generates many exception conditions, like floating point exceptions, cache misses, thrashing, page faults, ... . On the other hand, intensive disk I/O won't cause a high system time, because the system time is a measure of CPU activity - and the CPU isn't much involved in I/O. hope that helped, yours gerald -- Gerald Loeffler PhD student in Theoretical Biochemistry Email: gl@mdy.univie.ac.at Phone: +43 1 40480 612 Fax: +43 1 4028525 Mail: University of Vienna Institute for Theoretical Chemistry Theoretical Biochemistry Group Waehringerstrasse 17/Parterre A-1090 Wien, Austria #include "fancy_ASCII_picture" #include "funny_statement" #include "standard_disclaimer" From monteil@univ-angers.fr Wed Jan 4 10:33:52 1995 Received: from babinet.Univ-Angers.Fr for monteil@univ-angers.fr by www.ccl.net (8.6.9/930601.1506) id JAA26715; Wed, 4 Jan 1995 09:38:51 -0500 Message-Id: <199501041438.JAA26715@www.ccl.net> Received: from monteil.univ-angers.fr by babinet.Univ-Angers.Fr with SMTP (1.37.109.4/16.2) id AA17163; Wed, 4 Jan 95 15:38:09 GMT Date: Wed, 4 Jan 1995 15:37:41 +0000 To: chemistry@ccl.net From: monteil@univ-angers.fr (Andre Monteil) X-Sender: monteil@babinet Subject: Summary of Answer to Ln3+ and water Dear Netters, Thank you for your answers to my question : > I am looking for references about Lanthanide ions and water interaction. > Specially I need Lennard-Jones parameters for MD simulation. These are the answers : ****************************************************** Tom Cundari ****************************** The following paper by Ben Hay describes the extension of the MM2 force field to Ln(III) aqua complexes. This paper may have the data you require or point you to the appropriate literature references. Good luck. Hay, B. P. Inorg. Chem. 1991, 30, 1991. ****************************************************** Iosif Vaisman ****************************** AUTHOR(s): Helm, L. Foglia, F. Kowall, T. TITLE(s): Structure and dynamics of lanthanide ions and lanthanide complexes in solution. In: Journal of physics. Condensed matter : an Inst JUN 06 1994 v 6 n 23A supp Page: A137 SICI Code: 0953-8984(19940606)6:23A:suppL.A137:SDLI;1- AUTHOR(s): Frey, Steven T. TITLE(s): Characterization of Lanthanide Complexes with a Series of Amide-Based Macrocycles, Potential MRI Contrast Agents, Using Eu3+ Luminescence Spectroscopy and Molecular Mechanics. Summary: Eu3+ luminescence spectroscopy was used to thoroughly characterize complexes with a series of amide-based macrocycles in aqueous solution, providing stoichiometries, formation constants, and the number of Eu3+ coordinated water molecules. Deductions regarding the composition of the first coordination sphere of the Eu3+ ion are reported on the basis of molecular mechanics calculations. Shown here is a stereoscopic drawing of the energy-minimized structure of Eu2(bis(dtpa-eam)) (H2O)2. In: Inorganic chemistry. JUN 22 1994 v 33 n 13 Page: 2882 AUTHOR(s): Fossheim, R. Dahl, S.G. TITLE(s): Molecular Structure and Dynamics of Aminopolycarboxylates and their Lanthanide Ion Complexes. In: Acta chemica Scandinavica. AUG 01 1990 v 44 n 7 Page: 698 ****************************************************** Bill Ross ****************************** Here is some discussion on ions that may help, abstracted from a document in the tree under http://www.amber.ucsf.edu/amber/amber.html Bill Ross van der Waals parameters: ... Monoatomic ions, however, do not present such analogies in the Amber force field. Therefore a more basic discussion follows. The simplest approach to deriv- ing van der Waals parameters is to somehow match some experi- mental determination of the size of the atom in question. One source of such measurements is diffraction data. For example, the sum of metal and oxygen Pauling radii [2] tends to be 3% smaller than indicated by water-ion neutron and X-ray diffrac- tion data for Li+ and Na+ ions, 2% smaller than for K+, and 1% smaller than for Rb+ and Cs+ [3]. Another source of ion `size' information is crystallographic studies of ion complexes [4]. Since van der Waals parameters consist of two terms, the param- eters that yield a given size are not unique. Another experi- mental datum that can contribute to parameterization is the free energy of solvation in water or another relevant solvent. However, it is not clear whether the combination of experimen- tal size and solvation free energy determines unique modelling parameters. A further complication arises because R*, epsilon parameters are seperately ``combined'' with those for each of the other atom types to give the interatomic van der Waals potentials. Therefore, it is important to choose parameters consistent with the atom types that the new type will come in contact with. In particular, the TIP water models (as well as other waters) have a spherical van der Waals potential centered on the water oxygen, which is somewhat inflated to enclose the hydrogen atoms in the molecule. Thus a cation that has been parameterized to give a correct radial distance distribution function in such a water model will be ``smaller'' and come in closer contact with neighboring atoms if it is bound in a molecule consisting of Amber atoms. Moreover, different com- bining rules are in use in the modelling community, so that ____________________ [2] Pauling, L. The Nature of the Chemical Bond and the Structure of Molecules and Crystals; Cornell University Press: Ithaca, New York, 1960. [3] Ross, W. S. and Hardin, C. C., J. Am. Chem. Soc. 1994, 116, 6070-6080. (This discussion of van der Waals parameters is based on that work.) [4] Vedani, A., Huhta, D. W., J. Am. Chem. Soc. 1990, 112, 4759-4767 and references therein. ____________________ parameters from one convention must even be adapted to yield the same results in another scheme. Thus it was necessary to adapt the monovalent cation parameters of A"" vist [5] found in parm94.dat and parm91.dat for Amber so that the ion-water com- bined potential gave the same optimal distance as with the com- bining rules used by A"" vist. Matching the ion size in the environment seems to be sufficient in the case with monoatomic cations; the default has traditionally [6] been to use a some- what arbitrary well depth (epsilon) of 0.1 kcal/mol, charac- teristic of a rather nonpolarizable atom, and fit an R* parame- ter (see the Ross and Hardin reference). For simulating the effect of ions in solution on a macromolecular solute, insofar as the ions do not come in contact with the solute, the finer points of van der Waals parametrization should not have much bearing on the outcome. In a vacuum simulation with distance- dependent dielectric to simulate water, an ionic radius should be used which includes the solvation shell, and in this case, given the other approximations, exact van der Waals parameteri- zation is even less important. For the multivalent ions, dif- ferent further approaches may be considered to capture the quasi-bonding electron mobility, including the use of explicit bonds or hydrogen bonding terms (see the Vedani and Huhta reference). ****************************************************** Jan Hrusak ****************************** In the last NATO meeting this summer, there has been a poster to this topic by a group from Lausanne. Contact Thomas Kowall for more information. Institute de Chimie minerale et analytique, Universite de Lausanne BCH 1015 Lausanne ****************************************************** Fernando Luis Barroso da Silva ****************************** Did you read Hitoshi Ohtaki and Tomas Radnai's paper ( "Structure and Dynamics of Hydrated Ions" Chemical Reviews, 1993, vol. 93, n. 3 ) ? Look at page 1183! ****************************************************** Michael Probst ****************************** A paper dealing with aqueous LnCl3 solution (MD - Simulations) is in J.Phys.Chem. 94, p.4672 (1990) A paper with ab inito calculations on the interaction of water with all Lanthanides appeared in Z.Naturforsch. 46a, p.117 (1991) Hope that helps ! ****************************************************** By another way we found an interesting reference which was helpful : P. Auffinger and G. Wipff J. Am. Chem. Soc. (1991) 113, p 5976 Merci et Bonne Annee a tous Andre' \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ \ / \ Andre MONTEIL /|| / \ Laboratoire POMA \/ || / \ Universite d'Angers /\ || / \ 2, bd Lavoisier | |/ /|| / / \ 49045 ANGERS Cedex |__|__/ ||/ / \ / \ monteil@univ-angers.fr / \ Allo : 41 73 53 61 / \ Fax : 41 73 53 52 / \ / /////////////////////////////////////////////////////////////////// From hcj@mazda.wavefun.com Wed Jan 4 13:42:53 1995 Received: from volvo.wavefun.com for hcj@mazda.wavefun.com by www.ccl.net (8.6.9/930601.1506) id NAA01116; Wed, 4 Jan 1995 13:03:20 -0500 Received: from mazda.wavefun.com by volvo.wavefun.com via SMTP (931110.SGI/930416.SGI) for chemistry@ccl.net id AA09529; Wed, 4 Jan 95 10:03:15 -0800 Received: by mazda.wavefun.com (931110.SGI/930416.SGI) for @volvo.wavefun.com:chemistry@ccl.net id AA25908; Wed, 4 Jan 95 10:02:57 -0800 Date: Wed, 4 Jan 95 10:02:57 -0800 From: hcj@mazda.wavefun.com (Harry C. Johnson) Message-Id: <9501041802.AA25908@mazda.wavefun.com> To: chemistry@ccl.net Subject: Spartan Mailing List Reply-To: hcj@wavefun.com I am currently working on developing a Spartan mailing list. The list when put into full operation will be open to everyone with Spartan related questions and information, though it will be moderated. Anyone who is interested in helping me test the list, and thereby becoming the first subscribers should send mail to spartan@wavefun.com. Thanks! -Harry +--------------------------------------------------------------------+ |Harry C. Johnson | /--- ----\ /---/ ---\ ----- /---/ /| / | |Computational Chemist | \ / / / / / / / / / / | / | |Wavefunction Inc. | \ /---/ /---/ /---/ / /---/ / | / | |E-mail: hcj@wavefun.com|---/ / / / / \ / / / / |/ | +--------------------------------------------------------------------+ From jkong@is.dal.ca Wed Jan 4 15:33:51 1995 Received: from is.dal.ca for jkong@is.dal.ca by www.ccl.net (8.6.9/930601.1506) id PAA04541; Wed, 4 Jan 1995 15:23:19 -0500 Received: by is.dal.ca (AIX 3.2/UCB 5.64/4.03) id AA101002; Wed, 4 Jan 1995 16:23:24 -0400 Date: Wed, 4 Jan 1995 16:23:22 -0400 (AST) From: Jing Kong Subject: Multireference vs Multiconfiguration vs basis set To: chemistry@ccl.net Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear netters, I was wondering if optimizing the reference space will help the convergence of multireference CI. I have a case that the HF reference only contributes 0.7 in sum of squares of coef. in a single reference SDCI. I included 17 more configurations (46 spin adopted ones) in the reference space but the contribution from the ref configurations is still too low (0.87). I would like to see it be above 0.9 but I can't add more to reference space due to the limit of hardware. There two questions here I would like to hear from you: (1) should I run a MCSCF first to optimize the reference space? How much benefit I can get? (2) will a bigger basis set help? My basis set is already at triple zeta level. Any advice and pointers to reference are welcome! Thanks! Jing From jle@world.std.com Wed Jan 4 16:33:58 1995 Received: from europe.std.com for jle@world.std.com by www.ccl.net (8.6.9/930601.1506) id PAA05410; Wed, 4 Jan 1995 15:56:14 -0500 Received: from world.std.com by europe.std.com (8.6.8.1/Spike-8-1.0) id PAA07884; Wed, 4 Jan 1995 15:56:06 -0500 Received: by world.std.com (5.65c/Spike-2.0) id AA14716; Wed, 4 Jan 1995 15:56:24 -0500 Date: Wed, 4 Jan 1995 15:56:24 -0500 From: jle@world.std.com (Joe M Leonard) Message-Id: <199501042056.AA14716@world.std.com> To: chemistry@ccl.net Subject: Home computer choices Folks, I'm currently looking at upgrading my home computer (a MacIIci, remember them :-), and have a couple of questions for the Mac/PC users on the net... 1) Is there a way of telling whether the chip in a pentium PC is a "good" or "bad" chip (like the old double-sigma's)? I've seen posts about tests that can be run, but is there any before-purchase things that can be done? 2) Have folks gotten linux running on Pentiums - or are they merely running 486 code that doesn't know or care the difference? I assume there are different instructions in a Pentium, or am I full of it with this? 3) Are there graphics cards that should be used and/or avoided (in Dos/Windows and in linux)? Several vendors mention Diamond Stealth cards, for example - do these work with Linux? Win3.1? etc... 4) Have people used CD-ROM's that are external and swappable between PC's and Mac's (obviously a SCSI device)? Any comments on who's to use? 5) Are there fortran compilers for the Mac (There seem to be several good ones for the PC)? If so, who are they, and do they have anything resembling the tools on a workstation? I'll also take recommendations for C compilers for the Mac... 6) Are the Windows development environments "good enough" to keep one from having to reboot the machine when the code being developed crashes (possibly stupid question #106)? How about on the Mac? Yes, I'm rather used to workstations... 7) Finally are there PC vendors that I should use and or avoid? Thanks for your comments (well, the comments I expect to get :-)! I'd love to get a big SGI or IBM at home, but I'd rather not be killed in my sleep... Joe Leonard jle@world.std.com From Karl.F.Moschner@urlus.sprint.com Wed Jan 4 17:45:01 1995 Received: from sprintf.merit.edu for Karl.F.Moschner@urlus.sprint.com by www.ccl.net (8.6.9/930601.1506) id QAA06122; Wed, 4 Jan 1995 16:38:58 -0500 X400-Received: by mta merit in /PRMD=internet/ADMD=telemail/C=us/; Relayed; Wed, 4 Jan 1995 12:01:45 -0500 X400-Received: by /ADMD=TELEMAIL/C=US/; Relayed; Wed, 4 Jan 1995 11:52:15 -0500 X400-Received: by /PRMD=SMXFL2/ADMD=TELEMAIL/C=US/; Relayed; Wed, 4 Jan 1995 11:59:58 -0500 X400-Received: by /PRMD=LANGATE/ADMD=TELEMAIL/C=GB/; Relayed; Wed, 4 Jan 1995 11:59:00 -0500 Date: Wed, 4 Jan 1995 11:59:00 -0500 X400-Originator: Karl.F.Moschner@urlus.sprint.com X400-Recipients: non-disclosure:; X400-MTS-Identifier: [/PRMD=LANGATE/ADMD=TELEMAIL/C=GB/;Wed Jan 4 11:59:49 199503] X400-Content-Type: P2-1984 (2) Content-Identifier: RE: CPU vs Total From: Karl.F.Moschner@urlus.sprint.com Message-ID: <"Wed Jan 4 11:59:49 199503*/I=F/G=Karl/S=Moschner/OU=4267EDUR/O=TMUS.URL/PRMD=LANGATE/ADMD=TELEMAIL/C=GB/"@MHS> To: Patrick.Bultinck@rug.ac.be, CHEMISTRY@ccl.net Subject: RE: CPU vs Total Time If you're system is basically running single jobs, which is assumed throughout this discussion, it is reasonable to assume that the difference between the CPU time and the total time, also referred to as clock or wall time, is due to disk I/O, input/output. There is some overhead for memory access but it's negligible compared to I/O. The I/O overhead generally arises from either: updating job files, the output your generating; reading/writing temporary or scratch files; and/or, swapping, that is accessing virtual memory which is disk based. You'll have to take a closer look at your programs operations and monitor your system activity to identify the principal culprit but generally, anything you can do to improve I/O should speed up that portion of the job. The simple solutions: General - boost I/O Swapping - install more RAM Minimize job output - only ask for what you really need Scratch files - boost I/O and may also be minimized by selecting appropriate job options, e.g., for very I/O demanding Gaussian 92 MP2 calculations with frequencies, a "frequently asked question": Direct MP2=Stingy Opt=CalcHFFC FREQ=NoRaman NOTE: Minimizing scratch file size will also help you stay within current common "workstation" UNIX operating system (SGI Irix 5.2/6.0, IBM AIX, HP UX but not DEC OSF/1) file limitations, about 8-14 GB (gigabytes) for a file system and 2 GB for a single file. There are a few methods for boosting I/O. Some are simple with minimal additional expense while others may be major. Keep in mind that this only reduces I/O proportionately and will not effect CPU time when determining cost-benefit. To boost I/O, implement file systems, including swap space (virtual memory), which: Use disks with higher transfer rates Use higher bandwidth buses (MBs = megabytes per second, maximum): Fast-Wide Differential SCSI-2 20 MBs common on high-end systems Fast-Wide SCSI-2 10 MBs very common SCSI 5 MBs almost defunct Implement disk striping across multiple SCSI buses to get near aggregate speedups, e.g., 2 * F/W Diff. SCSI-2 = 40 MBS max. Striping on a single SCSI bus still limits you to the bus maximum. NOTE: Large scratch files generally will readily exceed cache memories available in some storage subsystems, i.e. RAID subsystems, essentially eliminating their benefit in this specific case. Implement special I/O subsystems, e.g., HIPPI, but it can be expensive. Good luck! Karl F. Moschner Karl.F.Moschner@urlus.sprint.com Unilever Research U. S. From steve@carbo.cc.binghamton.edu Wed Jan 4 19:33:52 1995 Received: from carbo.cc.binghamton.edu for steve@carbo.cc.binghamton.edu by www.ccl.net (8.6.9/930601.1506) id SAA08459; Wed, 4 Jan 1995 18:55:22 -0500 Received: by carbo.cc.binghamton.edu (931110.SGI/931108.SGI.ANONFTP) for chemistry@ccl.net id AA01466; Wed, 4 Jan 95 18:45:46 -0500 Date: Wed, 4 Jan 1995 18:43:02 -0500 (EST) From: Steven Schafer Subject: WWW site To: chemistry@ccl.net Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII This is just to announce that the State University of New York at Binghamton Chemistry WWW server has been moved to: http://chemiris.cc.binghamton.edu:8080 The old address will remain up for a few more days. Steven Schafer S.U.N.Y. Binghamton Chemistry Department Binghamton, New York From jle@world.std.com Wed Jan 4 22:33:54 1995 Received: from europe.std.com for jle@world.std.com by www.ccl.net (8.6.9/930601.1506) id WAA10588; Wed, 4 Jan 1995 22:32:21 -0500 Received: from world.std.com by europe.std.com (8.6.8.1/Spike-8-1.0) id WAA13679; Wed, 4 Jan 1995 22:32:15 -0500 Received: by world.std.com (5.65c/Spike-2.0) id AA28125; Wed, 4 Jan 1995 22:32:33 -0500 Date: Wed, 4 Jan 1995 22:32:33 -0500 From: jle@world.std.com (Joe M Leonard) Message-Id: <199501050332.AA28125@world.std.com> To: chemistry@ccl.net Subject: Question about cooperative multitasking... Folks, me again... Let me ask you all another question - how does one write applications for Windows or Macs to use cooperative multitasking? As somebody who's worked since the Good Old Days on machines with an interval timer (virtual memory, too, for the most part), how do you write code for systems without one? Is this "magically" done by the compiler/builder or does one have to insert "release-cpu-now" function calls throughout compute-intensive code? I've just realised, that if I want to program at home, I might be in for a REAL shock over this issue... Thanks in advance! Joe jle@kaiwan.com From ch11mh@surrey.ac.uk Wed Jan 4 22:39:57 1995 Received: from mail-server.surrey.ac.uk for ch11mh@surrey.ac.uk by www.ccl.net (8.6.9/930601.1506) id VAA10275; Wed, 4 Jan 1995 21:42:41 -0500 Received: from central.surrey.ac.uk by mail-server.surrey.ac.uk with SMTP (PP); Thu, 5 Jan 1995 02:34:37 +0000 Received: by central.surrey.ac.uk (1.37.109.8/16.2) id AA22724; Thu, 5 Jan 1995 02:33:37 GMT Date: Thu, 5 Jan 1995 02:33:36 +0000 (GMT) From: Mr Martin J Hargreaves To: chemistry@ccl.net Subject: WWW Site Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Just a quick message to announce the WWW server for the Department of Chemistry, University of Surrey, UK. The server can be reached at... http://www.chem.surrey.ac.uk/ I hope you find it interesting, it was put together by myself and another final year Computational Chemist from a scrounged old PC and Linux in a couple of days. Comments to ch11mh@surrey.ac.uk please. Regards, Martin. ---------------------------------------------------------------- | Martin Hargreaves, ch11mh@surrey.ac.uk| | Undergraduate Computational Chemist | | WWW Server Admin http://www.chem.surrey.ac.uk| ----------------------------------------------------------------