[molpro-user] CASSCF calculation error

Cristopher Camacho ccamacho at linuxmail.org
Sun Feb 4 17:25:51 GMT 2007


I am trying to perform a CASSCF calculation with a big active space (18 active orbitals and 14 active electrons). As reference I am using the CAS wavefunction produced with 12 active orbitals and 14 active electrons, but my calculation stops with the following error

Wavefunction dump at record             2142.2

 Convergence thresholds  0.10E-01 (gradient)  0.10E-05 (energy)  0.10E-02 (step length)

 Number of orbital rotations:     834     (  56 Core/Active  404 Core/Virtual   0 Active/Active  374 Active/Virtual)
 Total number of variables: 126603458


 ITER. MIC  NCI  NEG     ENERGY(VAR)     ENERGY(PROJ)   ENERGY CHANGE     GRAD(0)  GRAD(ORB)   GRAD(CI)     STEP       TIME

 ****  SINGULARITY IN SCHMIDT ORTHOGONALISER
  ISTATE =                         1

 ERROR EXIT
 CURRENT STACK:      MAIN

I know the Schmidt orthonormalization is performed to produces the orthonormal basis to optimize CI coefficients by doing some rotations in the CASSCF calculation, but does anybody know how to overcome this problem? Is there any option in molpro to control this?

Thank you very much.

Cheers,

Cristopher

=
Mat Concepts - Manufacture and Install
Specializing in mat programs for commercial real estate properties, buildings, retail, grocery stores, hotels, etc. Logos and custom designs. Your one source.


-- 
Powered by Outblaze




More information about the Molpro-user mailing list