Page 1 of 1

RESPONSE under CI calculation : ERROR (MOLLAB) MOLECULE label not found

Posted: 08 Nov 2019, 03:15
by ankit7540
Hello,
I am trying to obtain ground state properties for O2, starting with mcscf (CAS 12) and possibly moving to CI.
In this regard, it seems that the wavefunction part of the calculation runs fine, but problems occur for RESPONSE module for CI.
( for mcscf, RESPONSE runs ok and the values of polarizabilities seems to be as expected, file attached).

In the end of the output file from CI calculation, I get,

Code: Select all

 


                 .------------------------------------------------.
                 | Starting in Static Property Section (ABACUS) - |
                 `------------------------------------------------'


 
     Date and time (Linux)  : Fri Nov  8 13:11:26 2019
     Host name              : varsh                                   

 *** ERROR (MOLLAB), MOLECULE label SIR IPH  not found on unit  11
 File name: /mnt/raid0/scratch/dalton/DALTON_scratch_ankit/p_CI_O2_8319/SIRIFC              


 QTRACE dump of internal trace stack

 ========================
      level    module
 ========================
          5    MAGINT      
          4    ABACTL      
          3    ABACUS      
          2    DALTON      
          1    DALTON main 
 ========================



 --- DUMP OF LABELS ON UNIT   11
 Rec. no.     1  ********   8Nov19   13:11:26  CIRESPON
 Rec. no.     9  ********   8Nov19   13:11:26  CIDIAG2 
 Rec. no.    11  ********   8Nov19   13:11:26  SIRFLAGS
 Rec. no.    13  ********   8Nov19   13:11:26  CREFDETS
 Rec. no.    15  ********   8Nov19   13:11:26  EODATA  

        15 records read from file.

  --- SEVERE ERROR, PROGRAM WILL BE ABORTED ---
     Date and time (Linux)  : Fri Nov  8 13:11:26 2019
     Host name              : varsh                                   

 Reason: ERROR (MOLLAB) MOLECULE label not found on file

  Total CPU  time used in DALTON:   5.24 seconds
  Total wall time used in DALTON:   1.07 seconds


 QTRACE dump of internal trace stack

 ========================
      level    module
 ========================
          5    MAGINT      
          4    ABACTL      
          3    ABACUS      
          2    DALTON      
          1    DALTON main 
 ========================

 
This is after some output from RESPONSE regarding polarizabilities, whose values are incorrect (significantly different from that of mcscf output and previous expt reports). This problems seems to be related to this viewtopic.php?t=527 earlier question. In my case, I make sure to run each calculation with its own scratch directory.

Full output files are attached.

Re: RESPONSE under CI calculation : ERROR (MOLLAB) MOLECULE label not found

Posted: 10 Nov 2019, 17:58
by hjaaj
Dear Ankit,

Both response calculations are correct. The CI calculation fails in the **PROPERTIES module which you have asked for with the .RUN ALL input. (I can see it is confusing the way Dalton stops; CI is not implemented in the **PROPERTIES module so Dalton should stop with an error message instead of trying to read MCSCF information which does not exist.)

The calculated polarizabilities are terrible for CI because CI corresponds to a sum over states and you have only included the valence states from 2s and 2p MOs in the CI.

Re: RESPONSE under CI calculation : ERROR (MOLLAB) MOLECULE label not found

Posted: 12 Nov 2019, 07:10
by ankit7540
Dear Prof. Jensen,
Thank you for the reply.
If I understand correctly, I should now include 3s, 3p and possibly 3d orbitals, in the active space for the CI calculation. I guess that should improve the polarizability for CI method.