From chemistry-request@ccl.net Thu May 7 13:23:28 1992 Date: Thu, 7 May 92 12:04:08 -0400 From: gt5841a@prism.gatech.EDU (ERNST,MARGOT) Subject: small basis/CI calculations To: chemistry@ccl.net Status: R It seems to be accepted wisdom, that it's not of much value to carry out correlation interaction calculations based on a HF wavefunction with a small (near minimal) basis set. I would like to know more precisely which problems would be caused by such treatment, i.e. how would the presumably overestimated correlation be reflected in quantities such as relative excitation energies, transition d dipole moments and the like. I assume that there would be some general trend, but I feel it is difficult to make an estimate as to what this trend will be. Of course for the case of excitation energies one has to assume that the method that's being used is valid for excited states, and that ground- and excited states would be treated to a comparable level of theory. For the case of ground state properties I would be interested in the effect of the above treatment on the covalency (or polarization) of bonds, i.e. if the HF min. basis wavefunction overestimates polarization, will a CI density lead to an overcorrection? I hope somebody out there will understand what I'm trying to ask and help me out with your experience. By the way, all of that is assuming that the small basis HF method yielded an acceptable geometry for the ground state. Thanks, an inexperienced graduate student I would prefer replies on the net. From chemistry-request@ccl.net Thu May 7 16:26:43 1992 Date: Thu, 7 May 1992 14:31 EDT From: Stephen Urquhart Subject: SIND01 To: chemistry@ccl.net Status: R Dear Fellow Listmembers, I am looking for a source of the code for SIND01 - I am hoping that it resides somewhere convienently on the net. In general, I am looking at semiempirical QM programs that are parameterized for transition metals - specifically for organometalic compounds. My group already has alot of experience with EHMO, but we would like to look at new codes as they become available. Cheers, Stephen Urquhart urquhart@sscvax.cis.mcmaster.ca From chemistry-request@ccl.net Thu May 7 23:30:10 1992 Date: 7 May 92 18:08:00 PST From: MAX VASQUEZ Subject: Experience with Microway's i-860-based card To: chemistry Status: R Dear Netters, A friend from South America asked me for information about a numerical accelerator card available from a company called Microway. It is based on the Intel i-860 chip and is supposed to be attached to a regular 386 or 486 PC-type machine running DOS, UNIX, XENIX, etc. Microway also sells C, C++, and Fortran compilers for the beast. They claim in their advertisements that one can get '80 megaflops of raw numeric output.' If this is anything close to true it would make it faster than many SGI or IBM RS/600 workstations at a much lower price. So, my South American friend wonders if this sounds too good to be true and also wonders if someone out there has had (or heard of) direct experience with this contraption, or with the i-860 chip in general. He (and I) will appreciate your help on this and I will summarize for the net. Thank you very much, Max Vasquez From chemistry-request@ccl.net Fri May 8 00:41:40 1992 Date: Fri, 8 May 1992 10:18:40 SST From: "Choo, Heng Kek" Subject: Gauss86 query To: Computational Chem Grp at Ohio Supercom Ctr Status: R Hello, I'm not a chemist but am running G86 for someone who is. The type of job is: #N HF/AM1 POP=REGULAR The result for the dipole moment is presented only in its x- , y- and z- components as: DIPOLE MOMENT= 0.418443 -0.334582 0.000000 What can I specify so that the 'total' moment be reported as well? I.e. in the form (for a '#N HF/STO-3G POP=REGULAR' job): DIPOLE MOMENT (DEBYE): X= 1.0904 Y= 0.6204 Z= 0.0000 TOT= 1.2545 ***** Please respond directly to me as I'm not subscribed to the ***** group. Thanks in advance. Sincerely Kek