Molcas Forum

Rafael

Member

Registered: 2016-11-16

Posts: 12

CASPT2 with GASSCF orbitals

Hi,

Is it possible to get CASPT2 energies from GASSCF wave functions?

I used the following input:

&GATEWAY
Basis set
C.cc-pVDZ
C1    -0.5640000  0.000000  0.000000 Angstrom
End of basis
Basis set
O.cc-pVDZ
O1     0.5640000  0.000000  0.000000 Angstrom
End of basis
*------------------------------------------------------------------------------
 &SEWARD &END
End of input
*------------------------------------------------------------------------------
 &SCF &END
PRORbitals
 2 1.d+10
OCCUPIED
 7
OCCNumbers
 2.0 2.0 2.0 2.0 2.0 2.0 2.0
End of input
*--------------------------------------------------------------------------------
>> LINK FORCE $Project.JobIph_S1   JOBIPH
>> LINK FORCE $Project.JobOld_S1x  JOBOLD
>> LINK FORCE $Project.ScfOrb      INPORB
 &RASSCF &END
LUMORB
Symmetry
 1
Spin
 1
Supsym
 1
 1 1
NACTel
14 0 0
GASSCF
 1
 9
 14 14
End of input
*-------------------------------------------------------------------------------
>> LINK FORCE $Project.JobMix_S1x  JOBMIX
 &CASPT2 &END
NoMultistate
Frozen
 0
IMAG
0.05
End of input

But I am getting this error:

()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

                                              &CASPT2

                                   only a single process is used
                        available to each process: 28 GB of memory, 1 thread
()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()()

 Cholesky algorithm in CASPT2 =    1


++    Wave function specifications:
      -----------------------------

      Number of closed shell electrons           0
      Number of electrons in active shells      14
      Max number of holes in RAS1 space          0
      Max number of electrons in RAS3 space      0
      Number of inactive orbitals                0
      Number of active orbitals                  9
      Number of secondary orbitals              19
      Spin quantum number                      0.0
      State symmetry                             1
      Number of CSFs                           540
      Number of root(s) available                1
      Root passed to geometry opt.               1
      A file JOBMIX will be created.
      This is a CASSCF or RASSCF reference function
--

++    Orbital specifications:
      -----------------------

      Symmetry species                           1
                                                 a
      Frozen orbitals                            0
      Inactive orbitals                          0
      Active orbitals                            9
      Secondary orbitals                        19
      Deleted orbitals                           0
      Number of basis functions                 28
--

      Type of Fock operator: STANDARD
      Type of 0-order Hamiltonian: STANDARD WITH DEFAULT IPEA
        "IP-EA" denominator shift =      0.25
      Extra imaginary denominator shift =      0.05
      The input orbitals and the CI vector will be transformed.

  ERROR IN SUBROUTINE GINIT.
   NR OF ACTIVE ORBITALS:                    1
  NR OF ACTIVE ELECTRONS:                   14
         SPIN DEGENERACY:                    1
--- Stop Module:  caspt2 at Fri Apr 21 14:03:31 2017 /rc= _INTERNAL_ERROR_ ---
Non-zero return code - check program input/output
--- Stop Module:  caspt2 at Fri Apr 21 14:03:31 2017 /rc= _INTERNAL_ERROR_ ---

Ps.: I need to use GASSCF because I want to compute single and double core-excited states.

Last edited by Rafael (2017-04-21 13:07:22)

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: CASPT2 with GASSCF orbitals

As far as I know, the CASPT2 program currently supports CASSCF and RASSCF wavefunctions, but not GASSCF.

limanni

Member

Registered: 2015-11-09

Posts: 5

Re: CASPT2 with GASSCF orbitals

Yes. GASPT2 is possible.... http://pubs.acs.org/doi/abs/10.1021/acs.jctc.6b00382

And it is in the OpenMolcas... do not hesitate to ask me if you have questions.

Giovanni

Rafael

Member

Registered: 2016-11-16

Posts: 12

Re: CASPT2 with GASSCF orbitals

Yes. GASPT2 is possible.... http://pubs.acs.org/doi/abs/10.1021/acs.jctc.6b00382

And it is in the OpenMolcas... do not hesitate to ask me if you have questions.

Giovanni

Hi Giovanni,

I used the latest Molcas 8.2 version, so apparently the GASPT2 is not implemented on it. I'll have to wait for a newer version.

Rafael

Member

Registered: 2016-11-16

Posts: 12

Re: CASPT2 with GASSCF orbitals

Yes. GASPT2 is possible.... http://pubs.acs.org/doi/abs/10.1021/acs.jctc.6b00382

And it is in the OpenMolcas... do not hesitate to ask me if you have questions.

Giovanni

I got access to the OpenMolcas 17.0 and I tried the same input I posted above but I got the same error. Is there any special keyword to allow CASPT2 on top of GASSCF?

Thank you for the help

jameslamb

Member

Registered: 2018-05-09

Posts: 1

Re: CASPT2 with GASSCF orbitals

Did anyone find out how to implement GASPT2?

nikolay

Member

From: Stuttgart

Registered: 2016-03-21

Posts: 54

Re: CASPT2 with GASSCF orbitals

You should email limanni

zorkov

Member

Registered: 2016-06-04

Posts: 3

Re: CASPT2 with GASSCF orbitals

I got the same problem in both RASPT2 and GASPT2 calculations. It seems that the error is in src/caspt2/poly0.f. Before calling the subroutine GINIT_CP2, the number of active orbitals in poly0.f is correct. After entering GINIT_CP2, the number of active orbitals becomes to 1 or 2, which leads to that error.  This error also exists in some CASPT2 calculations with very large active space, for example, 7 electrons in 15 orbitals.

Last edited by zorkov (2019-05-05 09:24:04)

andrewshyichuk

Member

Registered: 2020-02-13

Posts: 86

Re: CASPT2 with GASSCF orbitals

I've got the same error today with OpenMolcas 19.11 / d7828899.

So, what is the current status? is it possible to do CASPT2 calculations on top of a GASSCF wavefunction?

ruslan

Member

Registered: 2020-10-29

Posts: 5

Re: CASPT2 with GASSCF orbitals

I've got the same error today with OpenMolcas 19.11 / d7828899.

So, what is the current status? is it possible to do CASPT2 calculations on top of a GASSCF wavefunction?

Yes, I join this question, too. Also, when I try to use >>COPY $PROJECT.JobMix JOB001 after GASSCf calculation I recieve the next mistake:

>>> COPY test.JobMix JOB001
Error: file "/tmp/test/test.JobMix" not found