From owner-chemistry@ccl.net Mon Sep 18 12:21:00 2006 From: "Pablo A. Denis pablod%%fq.edu.uy" To: CCL Subject: CCL: gamess Message-Id: <-32559-060918121217-13832-FA0b0RqRbeMADKOVT1Kp7w(-)server.ccl.net> X-Original-From: "Pablo A. Denis" Date: Mon, 18 Sep 2006 12:12:16 -0400 Sent to CCL by: "Pablo A. Denis" [pablod^fq.edu.uy] Hi CCLers, I am having some problems to install gamess under AIX 5.3 power5 proc, employing the following software: ESSL4.0, XLF10.1 and XLC8.0. 1- The compilation goes without any error messages in compall.log, lked.log compddi.log (the blaswrap is used), however six examples crashed, 5, 14, 19 (GUGA=CITYP) and 23, 25 and 27 (semiempirical pm3 and am1). The error message is the same: DDI process 0: trapped a segmetnation fault (SIGSEGV) Any hint to figure out where is the mistake? 2- I have compiling gamess with level -O3, higher optimization -O4 and -O5 failed. I had problems in the compilation of three fiels, int2d int2f, grd2b. any suggestions? Many thanks in advance. Regards, Pablo pablod*fq.edu.uy University of Uruguay From owner-chemistry@ccl.net Mon Sep 18 13:32:00 2006 From: "R.Arnab r.arnab-#-gmail.com" To: CCL Subject: CCL: [CCL] Building Amorphous Polymer Message-Id: <-32560-060918130130-6929-vDRx32rTofxfz+xZIS8+lQ[a]server.ccl.net> X-Original-From: "R.Arnab" Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 18 Sep 2006 06:58:42 -0500 MIME-Version: 1.0 Sent to CCL by: "R.Arnab" [r.arnab##gmail.com] Hi, I would appreciate if anybody can refer me to any software which is capable of building large amorphous polymer structure. Till date i have been using Cerius2.0 but it fails to generate an amorphous polymer bigger than a certain size (~ 10000 atoms) . Any other work around would also be helpful. Thanking in advance. Thanks Arnab From owner-chemistry@ccl.net Mon Sep 18 14:07:00 2006 From: "R.Arnab r.arnab**gmail.com" To: CCL Subject: CCL: Building Amorphous Polymer Message-Id: <-32561-060918124352-1043-ZJlsyfeNbMsydqnUonUZig..server.ccl.net> X-Original-From: "R.Arnab" Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 18 Sep 2006 05:58:03 -0500 MIME-Version: 1.0 Sent to CCL by: "R.Arnab" [r.arnab/a\gmail.com] Hi, I would appreciate if anybody can refer me to any software which is capable of building large amorphous polymer structure. Till date i have been using Cerius2.0 but it fails to generate an amorphous polymer bigger than a certain size (~ 10000 atoms) . Any other work around would also be helpful. Thanking in advance. Thanks Arnab From owner-chemistry@ccl.net Mon Sep 18 15:31:00 2006 From: "Mary Veronica O Connor moconnor.*_*.westliberty.edu" To: CCL Subject: CCL: Force field parameters for heme iron Message-Id: <-32562-060918134355-24998-KEQd/SIj26adDR6Wr9MtQg*_*server.ccl.net> X-Original-From: "Mary Veronica O Connor" Date: Mon, 18 Sep 2006 13:43:55 -0400 Sent to CCL by: "Mary Veronica O Connor" [moconnor|,|westliberty.edu] I am looking for citations of the latest (or best) papers that give detailed instructions on how to construct force field parameters. Also, can anyone suggest an appropriate training set of molecules to get force field parameters for heme iron? Thanks, Mary From owner-chemistry@ccl.net Mon Sep 18 17:50:01 2006 From: "Wai-To Chan chan{:}curl.gkcl.yorku.ca" To: CCL Subject: CCL: Energy convergence around conical intersection Message-Id: <-32563-060918174400-30110-OUT/MN3pEjiQzPdF+i/wdA]|[server.ccl.net> X-Original-From: Wai-To Chan Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=us-ascii Date: Mon, 18 Sep 2006 17:57:34 -0400 (EDT) MIME-Version: 1.0 Sent to CCL by: Wai-To Chan [chan,,curl.gkcl.yorku.ca] <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< Thank you for the reply. I think I should clarify how i perform the excited state energy calculations. I ran state averaged calculations using zero weight for the ground state. Sometimes, I have to increase the weight of the ground state so as to obtain convergence in the excited state. This happens for the points around the conical intersection. My experience is that what affect convergence most is the initial guess you use. This is a point I am wondering about. What I know is that the initial guess would just help in reaching convergence but no matter what is the initial guess, when the system reach convergence, it always reaches the same point no matter what is the initial guess. THe point confuses me here is that, around the conical intersection, reading different molecular orbitals would lead to different energy (converged of course). This is what we find practically, I am wondering if there is a theoretical background or an explanation for this. >>>>>>>>>>>>>>>>>>>>>>>>> Comments in my previous message were meant to be in general terms. I didn't really have the case of conical intersection in my mind. I have limited experience in excited state CASSCF calculations with GAMESS but I've never tackled conical intersection. Anyway to further elaborate on my comments I actually meant to say that the initial guess in MCSCF not only affect the 'smoothness' of the convergence but the final solution as well. A while ago I carried out a series of tests on the sensitivity of the CASSCF solution of a number of systems of varying sizes to different initial guess. I tried using natural orbitals generated from various wavefunctions including broken symmetry UHF, triplet state UHF and RHF-based CIS. I no longer have the files with me. As far as I remember for ozone I obtained the same CASSCF energy using these different initial guesses. This is not the case with the target system in my study. I was trying to find the most reliable guess for a MCSCF excited state calculation of a closed-shell structure in the region of a potential surface showing a RHF ---> UHF instability. I came to the conclusion that the right initial guess to use are the natural orbitals generated from a STABLE UHF solution. That means if there are more than one broken symmetry UHF solution choose the one with lower energy. Apparently this conclusion does not necessarily generalize to other cases. One interesting example is the CASSCF dissociation curve for H2O shown in Jensen's computational chemistry text book (fig 11.5 pg 278). There you will notice that the curve is continous from the equilibrium geometry all the way to the dissociation limit. I thought to myself if the correct CASSCF energy can only be obtained > from a stable UHF solution then I would expect to see some kind of non-physical discontinuity in the CASSCF curve setting in upon elongation of the O--H bond beyond the point where RHF becomes unstable. I suspect that the CASSCF curve shown in this plot almost approaches the all-electron full CI limit since the procedure includes all but only the two core electrons in the CI step. So it doesn't matter if an unstable RHF guess is used over the entire potential curve. For systems of moderate size I would rather play safe and stick with a stable reference wavefunction. Anyway the experience I described above may or not may not be pertinent to your situation. My understanding is that the MCSCF solution like Hartree-Fock is non-linear. The self-consistent field procedure starting from an inappropriate guess might converge to an unstable solution. It seems logical to assume that that CASSCF could converge to different solutions for the same geometry. I can't think of any study addressing the choice of initial guess for excited state MCSCF near the conical intersection. I do remember vaguely that Pulay once said something to the effect that when you want to average two states close in energy you better use the average density in the initial guess. I will try to dig out the comments if you wish. Wai-To Chan From owner-chemistry@ccl.net Mon Sep 18 19:31:01 2006 From: "jerome]![scs.uiuc.edu" To: CCL Subject: CCL: Force field parameters for heme iron Message-Id: <-32564-060918184331-20153-iuDx584H1PjfQPE3ZDf2wA(a)server.ccl.net> X-Original-From: jerome^scs.uiuc.edu Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset="utf-8"; format=flowed Date: Mon, 18 Sep 2006 15:22:53 -0500 Mime-Version: 1.0 Sent to CCL by: jerome*_*scs.uiuc.edu Hi Mary, for details on heme parametrization and relevant references, you can have a look at: Classical Force Field Parameters for the Heme Prosthetic Group of Cytochrome C. F. Autenrieth, E. Tajkhorshid, J. Baudry, and Z. Luthey-Schulten. J. Comp. Chem. (2004) 25(13):1613-1622 All the best, Jerome =========================================== Jerome Baudry, Ph.D. Research Assistant Professor in Chemical Sciences University of Illinois at Urbana-Champaign. School of Chemical Sciences. Noyes lab #150, Box 2-1. Urbana, IL, 61801. (217) 244 3210 jerome^scs.uiuc.edu http://www.scs.uiuc.edu/~jerome/ > -----Original Message----- Mary Veronica O Connor moconnor.^.westliberty.edu writes: > Sent to CCL by: "Mary Veronica O Connor" [moconnor|,|westliberty.edu] > I am looking for citations of the latest (or best) papers that give detailed instructions on how to construct force field parameters. Also, can anyone suggest an appropriate training set of molecules to get force field parameters for heme iron? > Thanks, > Mary From owner-chemistry@ccl.net Mon Sep 18 21:59:00 2006 From: "John McKelvey jmmckel*_*attglobal.net" To: CCL Subject: CCL: Coarse Grained Computing Message-Id: <-32565-060918215530-28936-JNfsBtS0GKXNRP8Fscb4Qg .. server.ccl.net> X-Original-From: John McKelvey Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=ISO-8859-1; format=flowed Date: Mon, 18 Sep 2006 21:52:22 -0400 MIME-Version: 1.0 Sent to CCL by: John McKelvey [jmmckel(0)attglobal.net] Greetings!! I am parameterizing a semiempirical method, and the process is "embarassingly parallel." I once did it in fortran on an SGI box which supports both fork and wait; but, alas, most linux F77 ligraries do not, and I am not a C programmer... So, for an SMP system such as a dual-dual AMD box, what would be the best, and frankly, easiest way to do this? I can either pass only a molecule file name and let the semiempirical code open the file, or I can pass the input file itself. I only want back one floating point number for each file. Thoughts, and suggestions are welcome. Many thanks! John McKelvey