CI Convergence problem

[Arun Venkathanathan <arun()at()chem.ucla.edu>]

 Dear CCLers,
 I am doing a SDCI calculations and have been using Davidson's
 diagonalization procedure to obtain convergence. I have recently made some
 modifications to the CI expansion (by using a truncated CI vector and
 approximate Hamiltonian to one of the contributions of sigma vector) and
 tried to run the CI calculations again using the Davidson's
 diagonalization and I find that the convergence now is very slow.
 Can anypne suggest me how to overcome this convergence problem and also if
 there are any possible Fortran codes available to accelerate convergence.
 Thanks in advance.
 - Arun
 ______________________________________________________________________
 Arun Venkathanathan                      | Tel: (310) - 825 - 6589 (O)
 Department of Chemistry & Biochemistry   |      (310) - 206 - 5931 (O)
 University of California, Los Angeles    |      (310) - 841 - 5167 (R)
 Los Angeles, CA 90095, USA  	         | Fax: (310) - 206 - 4038
 ----------------------  E-mail: arun()at()chem.ucla.edu   ------------------