I’ve seen that a few people asked related things, but I could find a direct answer. Sorry if I missed the relevant thread.
I am doing a single point calculation that reaches a timeout on the HPC, because the molecule is on the larger side, and I wanted to restart/continue from the fchk file. My input file looks as follows:
memory 96GB
set_num_threads(8)
psi4_io.set_default_path("/scratch/mol")
set maxiter 500
set freeze_core True
set PROPERTIES_ORIGIN [“COM”]
set cubeprop_tasks [‘esp’]
set basis aug-cc-pVTZ
E, wfn = properties(‘mp2’,properties=[‘dipole’],return_wfn=True)
cubeprop(wfn)
fchk(wfn, “mol.fchk”)
oeprop(wfn,“GRID_ESP”,“WIBERG_LOWDIN_INDICES”,“MULLIKEN_CHARGES”)
A fchk was written, so I would like to read and continue form there.
Thx a lot in advance!
Use of oeprop is discouraged. It’s normally better to go through properties.
3a. properties('mp2', properties=['dipole', 'GRID_ESP', 'WIBERG_LOWDIN_INDICES', 'MULLIKEN_CHARGES'], return_wfn=True) is the preferred way of computing properties. Does this not work for you?
3b. Why are you using oeprop? If something in the documentation told you to do this, let us know so we can change it.
Thanks a lot for the quick reply and the suggestions!
I’ll try it according to 2.
Regarding 3, there’s no specific reason, I guess - I’ll try it as you suggested.
Since you are performing an MP2 job, there are two different meanings for ‘restart from fchk’:
(1) read converged HF orbitals from fchk file and start the from MP2 calculation;
(2) read (if possible) the MP2 amplitudes from any Psi4-native file and compute the MP2 correlation energy directly.
where case (2) may be difficult since canonical MP2 amplitudes are computed using orbital energies and two-electrons integrals and thus amplitudes are seldom to be saved into file (as far as I know).
Case (1) is usually simple. For example, one can use the utility fch2psi to generate PSI4 input files (an inp file and an orbital file) from a given .fch(k) file. The utility fch2psi is open-sourced