QP2 CASSCF Not Converging

[NastaMauger]

Hello,

I am trying to run a CASSCF calculation with QP2 in order to compare my results with those obtained from GAMESS. I’m using the following setup to run it:

qp set Ao_two_e_ints do_ao_cholesky false
qp set Davidson csf_based true
qp set Determinants weight_selection 1
qp set trexio export_ao_one_e_ints  true
qp set Perturbation do_pt2 false
qp set Ao_two_e_ints io_ao_two_e_integrals Write
qp set Casscf_cipsi act_mos_opt true
qp set Casscf_cipsi criterion_casscf gradients
qp set Casscf_cipsi small_active_space false
qp set Casscf_cipsi adaptive_pt2_max false
qp set Casscf_cipsi pt2_min_casscf 0.0
qp set Casscf_cipsi n_big_act_orb 999

The CAS(4,4) calculation converged without any issues. CAS(8,8) on a different (but similar) system also converged successfully. However, on my current system, the CAS(8,8) calculation does not converge and shows strong oscillations.

I noticed that the number of determinants saved by QP2 is also oscillating. To address this, I reduced the value of f (qp set Determinants selection_factor 0.25) so that the number of determinants included in the wavefunction increases more gradually.

Despite this, I am still encountering convergence issues, and the output continues to show oscillatory behavior. I wonder if there is something specific that needs to be set in the input file to help with convergence?

Here is my output : casscf.txt

Thanks!