From owner-chemistry@ccl.net Sat Apr 14 05:36:00 2007 From: "Frank Neese neese(0)thch.uni-bonn.de" To: CCL Subject: CCL: ORCA --- how to read in MOs for restart? Message-Id: <-34046-070414051315-28850-OR8sA4BaQwT9nErVHbPVbA[#]server.ccl.net> X-Original-From: Frank Neese Content-Transfer-Encoding: 8bit Content-Type: text/plain; charset="iso-8859-1"; format=flowed Date: Sat, 14 Apr 2007 11:12:37 +0200 Mime-Version: 1.0 Sent to CCL by: Frank Neese [neese-,-thch.uni-bonn.de] Dear Marianna, it will likely not be possible to feed the MOs > from your program into ORCA. At this point it would be a major operation since the format of the gbw file is too complicated to be accessed > from outside. I am considering to provide an easier interface but this has not yet been done. Yes, semiempirical calculations do not use direct SCF since there are so few integrals that it makes most sense to store them. If this requires a lot of storage then your system must be truly huge (we have even finished DFT calculations with more than 700 atoms on a PC - do you have many thousand atoms?). I have difficulties to imagine that integral storage should be the bottleneck in your calculation. For very large systems it is the diagonalization that strongly dominates the computation time and in this situation it might be worth a try to switch to the Newton-Raphson converger. We should have a mailing list but it is presently still a bit disfunctional. It will certainly be improved in the future. Hope that helps, all the best, Frank ---------------------------------------------------------------- Prof. Dr. Frank Neese Lehrstuhl für Theoretische Chemie Universität Bonn Wegelerstr. 12 D-53115 Bonn e-mail: neese/./thch.uni-bonn.de http://www.thch.uni-bonn.de/tc/ Tel.: 0228/732351 (secretary: Ms Kronz) FAX: 0228/739064 ----------------------------------------------------------------