[molpro-user] Restart of MRCI+Q numerical hessian

[Denis DUFLOT]

Dear molpro users,

I have to evaluate the numerical hessian of a rather big molecule of Cs 
symmetry at MRCI+Q level.
To avoid exceeding time limit, I am testing the maxtask keyword in the 
frequency options.

My test example is HCO in minimal basis set. First I ask for maxtask= 5:

  memory,128,m
  file,2,test.wfu,new
  symmetry,z
  basis=STO-3G
  angstrom
   OC1=                 1.27874174 ANGS
   HO5=                 1.12360464 ANGS
   HOC3=               66.09668796 DEGREE
    geometry={
    c
    o    1 oc1
    h    2 ho5         1 hoc3
   }
  rhf
  multi
  ci

  {frequencies,variable=energd,maxtask=5}

the calculations does end correctly  with :

  Computing numerical hessian using default procedure for command CI
  Symmetry elements for unique displacements: Z
  Using no symmetry in wavefunction calculations
  Running default procedure: MULTI000  CI
  Numerical Hessian for variable ENERGD; current value: -111.73122022 AU
  Recomputing wavefunction at reference geometry
  Numerically approximating hessian using central energy differences
  Task list generated. Total number of displacements:     48

     5 tasks completed, CPU=1m 7s Elapsed=7s

  Task-limit exceeded. Restart information written to record  5500.2

So far, so good. Now to restart, my first guess is to modify the input 
something like this
{frequencies,variable=energd,restart=5500.2,maxtask=15}

which leads to :

  Computing numerical hessian using default procedure for command CI

  Symmetry elements for unique displacements: Z

  Using no symmetry in wavefunction calculations

  Running default procedure: MULTI000  CI
  Inconsistent active variables in restart: CX1A CX1
  Inconsistent active variables in restart: CY1A CY1
  Inconsistent active variables in restart: CZ1A CZ1
  Inconsistent active variables in restart: OX2A OX2
  Inconsistent active variables in restart: OY2A OY2
  Inconsistent active variables in restart: OZ2A OZ2
  Inconsistent active variables in restart: HX3A HX3
  Inconsistent active variables in restart: HY3A HY3
  Inconsistent active variables in restart: HZ3A HZ3
  Inconsistent zvarsym in restart: CX1A CX1
  Inconsistent zvarsym in restart: CY1A CY1
  Inconsistent zvarsym in restart: CZ1A CZ1
  Inconsistent zvarsym in restart: OX2A OX2
  Inconsistent zvarsym in restart: OY2A OY2
  Inconsistent zvarsym in restart: OZ2A OZ2
  Inconsistent zvarsym in restart: HX3A HX3
  Inconsistent zvarsym in restart: HY3A HY3
  Inconsistent zvarsym in restart: HZ3A HZ3
  ? Error
  ? Restart error
  ? The problem occurs in htask_read

This seems to be related to the symmetry breaking from Cs to C1 in the 
freq job. I have tried various
modifications of my input using the options of the freq keyword and/or 
deactivating symmetry and/or changing coordinates type but no luck.
The wfu file seems to be read correctly at the beginning of the restart. 
Especially the wfu contains :

  CX1             =     -1.38694144
  CY1             =     -0.05478494
  CZ1             =      0.00000000
  OX2             =      1.02942983
  OY2             =     -0.07681422
  OZ2             =      0.00000000
  HX3             =      0.18681067
  HY3             =      1.87213864
  HZ3             =      0.00000000

which corresponds to the 5th grid point in the 5th task, as expected...
No one seems to have reported this kind of problem on this list so any 
advice would be appreciated.

Best regards
Denis


-- 
*****************************************************************

Dr Denis Duflot
Maître de Conférences HDR/Associate Professor
Université Lille 1, Sciences et Technologies
Laboratoire de Physique des Lasers, Atomes et Molécules
UMR CNRS 8523 Bâtiment P5
F-59655 Villeneuve d'Ascq Cedex
France
e-mail : Denis.Duflot at univ-lille1.fr
https://www.researchgate.net/profile/Denis_Duflot/
Tél : (33) 3 20 43 49 80
Fax : (33) 3 20 43 40 84
*****************************************************************
IMAMPC 2013 CONFERENCE :
http://imampc2013.univ-lille1.fr
*****************************************************************