Molcas Forum

salomon142

Member

Registered: 2023-03-31

Posts: 1

restart an unconverged CASSCF due to walltime limit

Hello, I am new to OpenMolcas.
I am trying to perform a CASSCF+CASPT2 calculation to an organic molecule. I first performed a DFT single point calculation to get the starting orbitals. Then I submitted the CASSCF after copying the .ScfOrb file in the working directory and reading this file trough the following input:

&GATEWAY
 coord= cation.xyz
 Basis= H.ANO-L-VDZ,C.ANO-L-VTZ,N.ANO-L-VTZ
 Group= nosym

&SEWARD
 MEDI Cholesky

&RASSCF
 Title
  ANO-L-pVTZ CAS(12,12) STV2
 NActEl
   12 0 0
 Thrs
   1.0e-06 1.0e-03 1.0e-04
 Inactive
   93
 Ras2
   12
 CiRoot
   7 7 1
 Symmetry
   1
 Spin
   1
 OrbListing
   Nothing
 LumOrb

Is it correct?

Due to the walltime limit of 48 hours, the jobs ended before the wavefunction iterations reach the convergence criteria.
I resubmit the job reading the unconverged JobIph using the modified input:

&GATEWAY
 coord= cation.xyz
 Basis= H.ANO-L-VDZ,C.ANO-L-VTZ,N.ANO-L-VTZ
 Group= nosym

&SEWARD
 MEDI Cholesky

&RASSCF
 FileOrb = $CurrDir/cation_CASSCF.JobIph
 Title
  ANO-L-pVTZ CAS(12,12) STV2
 NActEl
   12 0 0
 Thrs
   1.0e-06 1.0e-03 1.0e-04
 Inactive
   93
 Ras2
   12
 CiRoot
   7 7 1
 Symmetry
   1
 Spin
   1
 OrbListing
   Nothing
 JobIph

However, the calculation seems to start from the scratch. Could you please tell me if there is any way to restart a CASSCF from an unconverged job?

If that is not possible, would it help to submit a job with lower the convergence criteria, and then read the corresponding converged JobIph file with the required criteria?

thanks for the help

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,011

Re: restart an unconverged CASSCF due to walltime limit

Try copying the JobIph as JobOld, and do not specify FileOrb:

>> COPY $CurrDir/cation_CASSCF.JobIph $Project.JobOld
&RASSCF
  Title
  ...
  JobIph