Requesting Help with - "CPASSERT failed error"

[Johnee Britto]

Hello CP2K community,

I am trying to optimize the geometry of a perovskite containing roughly 400 atoms

using the implicit solvent model in CP2K. I found there are two implicit solvent models available in CPK: SCRF (self-consistent reaction field) and the SCCS (continuum solvation model). 

For the SCRF, I am not able to run the calculations due to some CPASSERT failed error (see attached screenshot), which might be due to some input problem which I have no clue about. Could anyone help me solve this problem?

Here is the GEO_OPT input for SCRF.

&GLOBAL
  PROJECT      test
  PRINT_LEVEL  MEDIUM
  RUN_TYPE     GEO_OPT
  FLUSH_SHOULD_FLUSH  T
&END GLOBAL
&MOTION
   &GEO_OPT
    TYPE MINIMIZATION
    MAX_DR    1.0E-03
    MAX_FORCE 1.0E-04
    RMS_DR    1.0E-03
    RMS_FORCE 1.0E-04
    MAX_ITER 1000
    OPTIMIZER  BFGS
  &END GEO_OPT
&END MOTION

&FORCE_EVAL
   METHOD Quickstep
  &DFT
    BASIS_SET_FILE_NAME   path-to-basis/BASIS_MOLOPT
    POTENTIAL_FILE_NAME   path-to-potential/GTH_POTENTIALS
    CHARGE  0
    MULTIPLICITY 1
    &MGRID
      NGRIDS 4
      CUTOFF 350
      REL_CUTOFF 40
    &END MGRID
    &QS
    METHOD GPW
    EXTRAPOLATION_ORDER 3
    EPS_DEFAULT 1.0E-12
    &END QS

  &SCRF

    EPS_OUT 37.5
    LMAX  3
      &SPHERE
       RADIUS 10
      &END SPHERE
    &END SCRF

    &SCF
      SCF_GUESS    RESTART
      EPS_SCF      1.0E-6
      MAX_SCF      50
      &OUTER_SCF
        EPS_SCF                    1.0E-6
        MAX_SCF                    100
      &END OUTER_SCF
      &OT ON
        PRECONDITIONER             FULL_SINGLE_INVERSE
        MINIMIZER                  CG
        N_HISTORY_VEC               7
      &END OT
       &PRINT
        &RESTART
         LOG_PRINT_KEY             T
        &END RESTART
      &END PRINT
    &END SCF

      &XC
      &XC_GRID
       XC_SMOOTH_RHO              NN10
       XC_DERIV                   NN10_SMOOTH
      &END XC_GRID
      &XC_FUNCTIONAL              PBE
      &END XC_FUNCTIONAL
       &VDW_POTENTIAL
         POTENTIAL_TYPE  PAIR_POTENTIAL
         &PAIR_POTENTIAL
           TYPE  DFTD3
            REFERENCE_FUNCTIONAL PBE
            PARAMETER_FILE_NAME ./dftd3.dat
         &END PAIR_POTENTIAL
       &END VDW_POTENTIAL
    &END XC
  &END DFT
  &SUBSYS
    &TOPOLOGY
     COORD_FILE_NAME test.xyz
     COORD_FILE_FORMAT  XYZ
     &CENTER_COORDINATES     T
     &END CENTER_COORDINATES
    &END TOPOLOGY
    &PRINT
    &END PRINT
    &CELL
      A   75.00 00.0000 00.0000
      B   00.0000 75.00 00.0000
      C   00.0000 00.0000 75.00
      ALPHA_BETA_GAMMA  90.00 90.00 90.00
      PERIODIC XYZ
    &END CELL

For the SCCS solvent model, I am able to run the calculations just by replacing the $SCRF block by

    &SCCS
    RELATIVE_PERMITTIVITY 37.5
    &END SCCS

However, the SCF convergence is pretty slow. Can anyone suggest to me on how I can accelerate the SCF step using the SCCS model? I even tried changing the EPS_SCCS to 1.0E-04, and it is still very slow. 

Best regards,

Johnee

[Krack Matthias]

Hi Johnee

 

The SCRF method is most likely not maintained since much more than a decade and thus it is possibly not fully functional any longer.

The SCCS method adds an additional self-consistency cycle for the polarization potential during each SCF iteration step, once a specific SCF convergence threshold is reached. This introduces a computational cost for the implicit solvent. While using less stringent SCCS parameter settings can speed up calculations, it also increases the noise in the atomic forces, leading to less stable structure relaxations and MD simulations.

 

Best

 

Matthias

 

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/08d98cc2-f84f-4d69-813d-55ef61dc0616n%40googlegroups.com.

[Johnee Britto]

Dear Matthias,
Thank you so much for your time and the valuable information. It really helps.

Warm regards,
Johnee

You received this message because you are subscribed to a topic in the Google Groups "cp2k" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cp2k/VKGXymgqO3o/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cp2k+uns...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/ZRAP278MB082751B42D39D9CD385E58D6F4BC2%40ZRAP278MB0827.CHEP278.PROD.OUTLOOK.COM.