From psh@cxa.dl.ac.uk Sun Aug 15 14:58:30 1993 Date: Sun, 15 Aug 1993 13:58:30 +0100 From: "P. Sherwood" Subject: tools and packages, request for information To: chemistry@ccl.net Message-Id: <28403.9308151258@uk.ac.dl.cxa> Dear netters, A while ago there was a bit of discussion on the question of tools for computational chemistry, it appeared that most effort seems to have been put into the development of packages. I am interested in the extent to which existing packages can be regarded as collections of tools, and in knowing how they are integrated. I would like to compile a summary (which I will post to this list if I get a reasonable response) on "modular" software systems for computational chemistry. I am particularly interested knowing about those incorporating quantum chemistry, both ab-initio and semi-empirical, and/or MM/MD methods. I want to hear about solutions that are built up of distinct code modules which interwork closely by definition of common data structures, and provision of a control structure of some type. I would be grateful if developers and users of such packages would drop me a line directly (to save bandwidth since this basically an old topic) letting me know of their solutions. Please address the following points in any replies: *) functionality catered for. *) any use of standard formats for files (especially those relevant to quantum chemistry data eg integrals, symmetry, basis sets, wavefunctions). *) type of control structures, eg specialised control module that reads input, AVS-style graphical data flow network, unix shell (or similar) interpretive interface, or GUI (as in most modelling packages). *) provision of efficient communications (eg shared-memory segments) between modules. *) interworkability with other systems, eg visualisation. *) can the individual components be distributed/used without the package itself. *) support for heterogeneous network environment and novel architecture computing. *) availability status (commercial, public domain, private) *) do you want your solution publicised in the summary I prepare. Thanks in advance to anyone who can help, Paul ============================================================================== Dr Paul Sherwood, SERC Daresbury Lab, email: p.sherwood@daresbury.ac.uk Warrington, tel: -44-925-603553 WA4 4AD, fax: -44-925-603634 United Kingdom. ============================================================================== From shenkin@still3.chem.columbia.edu Sun Aug 15 10:47:50 1993 Date: Sun, 15 Aug 93 14:47:50 -0400 From: shenkin@still3.chem.columbia.edu (Peter Shenkin) Message-Id: <9308151847.AA12289@still3.chem.columbia.edu> To: "Bill Melchior, Subject: Re: PC worksations? > From chemistry-request@ccl.net Fri Aug 13 19:00:07 1993 > > Apologies if this has been discussed; I just joined the list: > > I'd like to know if anyone has had any experience using a high-speed PC as > a workstation on a Silicon Graphics computer. I.e., if I buy a PC, would > it be worth my while to upgrade memory, terminal, and video board in order > to use it for visualization of molecules following modelling on an SGI > machine? Most SGI 3D visualization software uses SGI's "GL", their family of 3D graphics acceleration hardware and software. Although it is possible to get such hardware/software for a PC through third parties, you might not be able to sit on the SGI and render to the PC; that would require "DGL", (Distributed GL), another piece of system software which might not be available on PC's at all, though it is bundled with all SGI's. If it's not available for the PC, you'll need PC versions of your visualization software. If you are using commercial software that uses the GL, it is unlikely that PC versions are available. In the future, the advent of OpenGL will make what you suggest more feasible; until then, I'd stay away from what you propose unless you're willing to spend a lot of time hacking. If your applications use only XWindows, then you can run an X-emulator on the PC and get acceptable performance -- though still way short of what any self-respecting UNIX system can offer. I do have experience with this configuration, and am reporting this based on my experience. I have no direct experience with PC's equipped with a GL board set -- perhaps someone with such experience can comment further. I suggest you look into SGI's new "Indy" line for low-cost machines that can do exactly what you want. Also, used Personal Irises are available for under $5K on the used computer market, I understand; beware, though: hardware repair and software update contracts are likely to be expensive, as they are pegged to the original cost of the machine. (This makes sense, because newer machines are more highly integrated and simpler to maintain; from the advent of the Indigo on, SGI sells self-maintenance contracts: you trade boards with SGI by FedEx and install them yourself. The PI, which predates the Indigo, is too complicated to allow this; any problem therefore requires a service call. The PI's are very nice machines, though.) In any case, either an Indy or a used PI is likely to be both cheaper and simpler to get to work than a PC configured to run GL. -P. ************************f*u*cn*rd*ths*u*cn*gt*a*gd*jb************************ Peter S. Shenkin, Box 768 Havemeyer Hall, Dept. of Chemistry, Columbia Univ., New York, NY 10027; shenkin@still3.chem.columbia.edu; (212) 854-5143 ********************** Wagner, Beame, Screvane in '93! **********************