Author Topic: -proper command doesn't give properties.  (Read 9172 times)

slackenerny

  • Jr. Member
  • **
  • Posts: 12
  • Karma: +0/-0
-proper command doesn't give properties.
« on: June 12, 2008, 04:41:04 PM »
I have a finished mp2 calculation, where I want to get some orbitals. I tried to add the pointval commands to the control file and then run 'ricc2 -proper'. Sadly it does not produce any plt files. Even if I keep the control file unchanged and run ricc2 -proper he does not rewrite the files for TD and SD. On the other hand I don't get an error message either. The file just ends with:

     *************************************************************
     *                                                           *
     *                RUNNING IN ANALYSIS MODE                   *
     *                ========================                   *
     *                                                           *
     *************************************************************



    ------------------------------------------------------------------------
         total  cpu-time :  17.67 seconds
         total wall-time :  17.67 seconds
    ------------------------------------------------------------------------


   ****  ricc2 : all done  ****


    2008-06-12 16:31:19.305

 ricc2 ended normally


And that's it. It is possible that one file or another was lost (correctly: Not copied back from the computing node). Might that be an explanation?

best wishes!

P.S.: The same goes for the $pop keyword. He just ignores it...

mpjohans

  • Full Member
  • ***
  • Posts: 26
  • Karma: +0/-0
    • .
Re: -proper command doesn't give properties.
« Reply #1 on: June 14, 2008, 06:05:45 PM »
Hello!

Yes, -proper seems broken in ricc2, in the way that it doesn't automatically do what it claims to. But you can do the same manually. If you've calculated the stuff needed for the density, for example using:
  $ricc2
    mp2
    geoopt model=mp2
you should have a file called something like "mp2-gsdn-1a1-000-total.cao" in your directory. You can transform that into a .plt file by adding the following to control:
  $anadens
    calc totaldensity from
    1d0 mp2-gsdn-1a1-000-total.cao

Then you'll get totaldensity.plt by running "ricc2 -fanal" or "ricc2 -proper", without any extra work.

More info in the manual: http://www.turbomole-gmbh.com/manuals/version_5_10/DOK_HTML/node206.html

Have a nice day,
    Mikael J.
    http://www.iki.fi/~mpjohans
 

christof.haettig

  • Global Moderator
  • Sr. Member
  • *****
  • Posts: 291
  • Karma: +0/-0
    • Hattig's Group at the RUB
Re: -proper command doesn't give properties.
« Reply #2 on: June 16, 2008, 12:49:03 PM »
Hello,

first:   ricc2 supporsts the -proper flag only since TM V 5.10, older versions only accept the -fanal option and require the
           $anadens data group in the control file

since TM V5.10 the -proper flag should work also for ricc2, but not that
           --  it requires the calculation of a correlated density (if you just request a MP2 calculation without a gradient,
                there is nothing to plot for ricc2, which you can't plot via  dscf...) and as there is (as for the other programs)
                no further input for the -proper option, the state for which the density is written to the interface files
                must be uniequely defined via the 'geoopt model=xxx'  option
           --  you can give all the options for the generation of plot files already in the calculation of the gradient, but if
                you use 'ricc2 -proper'  after the gradient calculation it requires (as rimp2) that the density file (gsdens for
                ground states and exdens for excited states) was kept

Thus, you either used TM V5.9.1 (or an older version) or the density files where not present because they have either not been calculated (not geoopt flag) or were lost...

Christof

PS: more details can be found in the section 'visualization of densities...' in the CC2 chapter of the docu.
           
« Last Edit: June 16, 2008, 12:56:37 PM by christof.haettig »