Feature Requests item #2378189, was opened at 2008-12-02 16:15 Message generated for change (Comment added) made by nobody You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=379136&aid=2378189&group_id=23629
Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: Interface Improvements Group: None Status: Open Priority: 5 Private: No Submitted By: Albert DeFusco (adefusco) Assigned to: Nobody/Anonymous (nobody) Summary: Proposed changes to GAMESS-US Reader Initial Comment: Hi there, I am a postdoc with Mark Gordon at Ames Lab and I have been hacking into Jmol recently with the goal of improving its ability to understand output from several methods. Below you will find changes I have made based on the released 11.6 source. I had started using the most recent SVN trunk, but there were some issues regarding antialiasing on the MAC that turned me away. --Additions-- GamessUSReader.java: Can now read EFP1 water molecules for FRAGNAME=[H2ODFT/H2ORHF]. These potentials are currently available in GAMESS. EFP2 potentials are not yet part of the release. This works for any number of EFP molecules and also for optimizations. There are more complete comments in the source. --Changes-- GamessReader.java and GamessUSReader.java I have built support for reading several more types of orbitals (closed shell only). This required changes to GamessReader.java that allow for orbital reading where energy, occupation or symmetry are not present. The changes I made allow for continuous addition of orbitals to the present model. This means that in a single-point run where the "initial guess vectors" are printed along with an RHF (or anything else) for N orbitals, the first N orbitals will be the guess and N+1 to 2N orbitals will be the optimized RHF orbitals. For an MCSCF calculation one may have up to three sets of orbitals (initial guess, MCSCF natural orbitals, MCSCF optimized orbitals). I have left it so that the user would have to be aware of what order the orbitals appear the figure out the mo number to display the desired orbital. (This may not work as expected in all cases.) While I have not tested all possible calculation types, there are some comments about keywords for which the changes work. MoldenReader.java: There were some MOLPRO-generated files that perhaps did not adhere strictly to the documented (albeit vague) standard on the MOLDEN website. I have included several output files which may help understand these changes. The files I want to include are too big to be uploaded here and are available at my website: http://homepage.mac.com/albertdefusco/gamess.tar.bz2 The file is 968 KB. I have had no prior experience with Java and apologize for the flagrant errors I have probably made. ---------------------------------------------------------------------- Comment By: Nobody/Anonymous (nobody) Date: 2009-04-09 13:53 Message: Bob, Although I can't test NBO, this has no adverse affects with my changes. The filter keyword could be quite useful for other orbital types in GAMESS. Albert ---------------------------------------------------------------------- Comment By: Bob Hanson (hansonr) Date: 2009-04-09 13:13 Message: Albert, we did some additional work on the NBO business. Please check this now. Note that both standard MOs and NBOs are returned by default; one can add FILTER "NBO" to the LOAD command to get only NBOs or FILTER "!NBO" to get the others. Bob ---------------------------------------------------------------------- Comment By: Albert DeFusco (adefusco) Date: 2008-12-03 15:49 Message: Bob, I've had a look at the SVN repository and things are working well. The antialising problems with MediaWiki have magically disappeared. With regards to MCSCF Natural Orbitals (wether ALDET or GUGA), the "core" orbitals are printed with their energy while the "active" orbitals have occupancies. I chose to read only the occupancy and any orbital with a negative number is set to 2.0. For this reason, line 80 (rev 10421) would belong with the final "else if" (which currently only reads CI natural orbitals with ORMAS), but more intelligent solutions which try to figure out where the active space starts and read accordingly are possible. An ORMAS example can be found at the link below (the initial guess has also been printed). With this in mind, perhaps the MO TITLEFORMAT default could be updated to "?Eenergy = %E %U" for the case when we have no energy. Because the unit is always set to "a.u.," the energy line is printed even when the energy is null. I suspect just a little logic change will fix it but I haven't been able to track it down in the source. http://homepage.mac.com/albertdefusco/ormas.gamess A correction in GamessUSReader.java at line 420. It should read: //CI NATURAL ORBITALS (from GUGA) using CSF configurations have occupancy and symmetry With example output: http://homepage.mac.com/albertdefusco/guga.gamess I now remember what was strange with the uracil-water optimization. Models 1.1 and 1.7 gets both EIGENVECTORS and MOLECULAR ORBITALS whereas the latter really belongs with model 1.7 only. I had tried to fix this once, but never quite got it work. I just had to be very aware of what was read and where it goes. In dealing with multiple orbital sets, it would be ideal if I could select the mo set to display in the "AtomSetChooser." I understand that this may not be entirely feasible and while I think general users would like a filter to apply when reading the file, perhaps an expert setting to read as it does now will also be helpful. Thank you, Albert ---------------------------------------------------------------------- Comment By: Bob Hanson (hansonr) Date: 2008-12-03 00:32 Message: First, let me say I'm very happy your group is looking at this. Thanks very much for contributing your expertise. Jmol 11.6 is closed to new features, but I've adapted your code and checked it in as 11.7.16_dev. Please try that. Don't know what the "antialiasing on the MAC" that could be a problem -- we have a lot of Mac users, and this problem would be important to resolve if it exists. In any case, I did do some substantial re-writing there, so please check it over and let me know if it's in error. It did read the MOs and H2O from the water uracil file. There were problems reading the GUGA file with the header setting as shown, so please take a good look at that. Also, since there are multiple molecular orbital types, what do you suggest in terms of reading only the ones that would be of particular interest. Perhaps we need a "filter" for this? Bob Hanson ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=379136&aid=2378189&group_id=23629 ------------------------------------------------------------------------------ This SF.net email is sponsored by: High Quality Requirements in a Collaborative Environment. Download a free trial of Rational Requirements Composer Now! http://p.sf.net/sfu/www-ibm-com _______________________________________________ Jmol-developers mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jmol-developers
