[molpro-user] CPMCSCF changes density matrices

Aaron Virshup aaron at spawn.scs.uiuc.edu
Fri Nov 4 17:38:54 GMT 2005


Hi molpro users,

I have encountered a strange problem with my density matrices when doing
CPMCSCF calculations in multi. Specifically, cpmcscf gradient
calculations overwrite the density matrix for the state for which they
are calculated.

Further, it appears that the density matrix is always overwritten with
the same matrix. For example, after a CAS(1x2) SA-2 calculation for
atomic hydrogen (multi;state,2;occ,2;closed,0;), the density matrices
are

Ground State:
   0.18269672   0.28443330
   0.28443330   0.44282296

Excited State:
    1.58205585  -1.44615575
   -1.44615575   1.32192960

After doing multi again with "cpmcscf,grad,state=1.1;" the density
matrices become

Ground State:
    0.88237628  -0.58086122
   -0.58086122   0.88237628

Excited State:
    1.58205585  -1.44615575
   -1.44615575   1.32192960

Doing multi a third time with "cpmcscf,grad,state=2.1" gives

Ground State:
    0.18269672   0.28443330
    0.28443330   0.44282296

Excited State:
    0.88237628  -0.58086122
   -0.58086122   0.88237628


I'm not sure what to make of this, but be careful if you try to access
density records after a cpmcscf calculation.

Aaron
-- 
============================================================
Aaron Virshup                                 Martinez Group
Department of Physics                    Phone: 217-244-7383
University of Illinois at Urbana-Champaign
1110 W. Green St.
Urbana, IL 61801                   aaron at spawn.scs.uiuc.edu
============================================================




More information about the Molpro-user mailing list