[CP2K-user] [CP2K:20254] OT convergence failure

Pierre-André Cazade pierre.andre.cazade at gmail.com
Wed May 29 15:03:46 UTC 2024
Previous message (by thread): [CP2K-user] [CP2K:20253] OT convergence failure
Next message (by thread): [CP2K-user] [CP2K:20262] OT convergence failure
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi Jurg,

Thank you for your answer. I'll give it a shot.

I am reluctant to share the full input file as my collaborators have
developed new materials they want to keep private for now.

The CELL_OPT phase was completed without any incident. My
collaborators need a stress-strain curve along some axes. During this phase
the calculation crashes for a relatively moderate strain (1%) while running
fine for larger strains, up to 25%.

Unfortunately, It is a regular issue when performing this kind of analysis.

I copy below the part of the input file I can freely share with you:

 &GLOBAL
   PRINT_LEVEL  MEDIUM
   PROJECT_NAME bpp
   RUN_TYPE  GEO_OPT
 &END GLOBAL
 &MOTION
   &GEO_OPT
     OPTIMIZER  LBFGS
     MAX_ITER  1000
     KEEP_SPACE_GROUP  T
     EPS_SYMMETRY     1.0000000000000000E-04
   &END GEO_OPT
 &END MOTION
 &FORCE_EVAL
   METHOD  QS
   STRESS_TENSOR  ANALYTICAL
   &DFT
     BASIS_SET_FILE_NAME BASIS_MOLOPT
     POTENTIAL_FILE_NAME GTH_POTENTIALS
     MULTIPLICITY  1
     CHARGE  0
     &SCF
       MAX_SCF  20
       EPS_SCF     9.9999999999999995E-07
       SCF_GUESS  ATOMIC
       &OT  T
         MINIMIZER  CG
         PRECONDITIONER  FULL_ALL
       &END OT
       &OUTER_SCF  T
         EPS_SCF     9.9999999999999995E-07
         MAX_SCF  100
       &END OUTER_SCF
     &END SCF
     &QS
       METHOD  GPW
     &END QS
     &MGRID
       NGRIDS  5
       CUTOFF     9.0000000000000000E+02
       REL_CUTOFF     6.0000000000000000E+01
     &END MGRID
     &XC
       &XC_FUNCTIONAL  NO_SHORTCUT
         &PBE  T
         &END PBE
       &END XC_FUNCTIONAL
       &VDW_POTENTIAL
         POTENTIAL_TYPE  PAIR_POTENTIAL
         &PAIR_POTENTIAL
           TYPE  DFTD3
           PARAMETER_FILE_NAME dftd3.dat
           REFERENCE_FUNCTIONAL PBE
         &END PAIR_POTENTIAL
       &END VDW_POTENTIAL
     &END XC
     &POISSON
       POISSON_SOLVER  PERIODIC
       PERIODIC  XYZ
     &END POISSON
   &END DFT
   &SUBSYS

Best regards,
Pierre

On Wed, May 29, 2024 at 3:48 PM Jürg Hutter <hutter at chem.uzh.ch> wrote:

> Hi
> so there is some deeper problem that I cannot see without more information.
>
> There is another option of OT that isn't performing a Cholesky
> decomposition.
> &OT
>     ALGORITHM  IRAC
>
> It might just delay the problem, but it is worth a try.
> If this fails you need to send a test example.
>
> regards
> JH
>
> ________________________________________
> From: cp2k at googlegroups.com <cp2k at googlegroups.com> on behalf of
> Pierre-André Cazade <pierre.andre.cazade at gmail.com>
> Sent: Wednesday, May 29, 2024 4:22 PM
> To: cp2k at googlegroups.com
> Subject: Re: [CP2K:20252] OT convergence failure
>
> Hi Jürg,
>
> I tried with EPS_DEFAULT=1.e-12 but it still failed and I got the same
> error message.
>
> Regards,
> Pierre
>
> On Wed, May 29, 2024 at 10:58 AM Jürg Hutter <hutter at chem.uzh.ch<mailto:
> hutter at chem.uzh.ch>> wrote:
> Hi
> could be a problem of numerical accuracy. Try to reduce EPS_DEFAULT.
> regards
> JH
>
> ________________________________________
> From: cp2k at googlegroups.com<mailto:cp2k at googlegroups.com> <
> cp2k at googlegroups.com<mailto:cp2k at googlegroups.com>> on behalf of
> pierre.an... at gmail.com<mailto:pierre.an... at gmail.com> <
> pierre.andre.cazade at gmail.com<mailto:pierre.andre.cazade at gmail.com>>
> Sent: Wednesday, May 29, 2024 11:52 AM
> To: cp2k
> Subject: [CP2K:20250] OT convergence failure
>
> Dear CP2K users,
>
> Regularly, when straining crystal structures, I get convergence failure of
> the Orbital Transformation procedure.
>
>
>  *******************************************************************************
>  *   ___
>      *
>  *  /   \
>     *
>  * [ABORT]
>      *
>  *  \___/         Cholesky decomposition failed. Matrix ill conditioned ?
>     *
>  *    |
>     *
>  *  O/|
>     *
>  * /| |
>     *
>  * / \
>  cp_dbcsr_cholesky.F:120 *
>
>  *******************************************************************************
>
>
>  ===== Routine Calling Stack =====
>
>            15 cp_dbcsr_cholesky_decompose
>            14 qs_ot_get_derivative
>            13 ot_mini
>            12 ot_scf_mini
>            11 qs_scf_loop_do_ot
>            10 qs_scf_new_mos
>             9 scf_env_do_scf_inner_loop
>             8 scf_env_do_scf
>             7 qs_energies
>             6 qs_forces
>             5 cp_eval_at
>             4 cp_opt_gopt_step
>             3 geoopt_lbfgs
>             2 cp_geo_opt
>             1 CP2K
>
>
> What puzzles me is that this happens for very small longitudinal strain of
> the crystal (less than 1%), while larger strains on the same crystal
> converges perfectly fine.
>
> Any insights about this issue and to overcome it. It has been plaguing my
> calculations for w a while now and it is impeding my research.
>
> Best regards,
> Pierre
>
> --
> 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+unsubscribe at googlegroups.com<mailto:
> cp2k%2Bunsubscribe at googlegroups.com><mailto:
> cp2k+unsubscribe at googlegroups.com<mailto:
> cp2k%2Bunsubscribe at googlegroups.com>>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/cp2k/5b0a3083-7895-4273-80f1-f95a070526f8n%40googlegroups.com
> <
> https://groups.google.com/d/msgid/cp2k/5b0a3083-7895-4273-80f1-f95a070526f8n%40googlegroups.com?utm_medium=email&utm_source=footer
> >.
>
> --
> 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+unsubscribe at googlegroups.com<mailto:
> cp2k%2Bunsubscribe at googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/cp2k/ZR0P278MB0759BE41D919AA77AF3C64C59FF22%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM
> .
>
> --
> 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+unsubscribe at googlegroups.com<mailto:
> cp2k+unsubscribe at googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/cp2k/CAFnf9gq5trNSnP6HQUMzqBsMcPyyOxpCv89_zoTVYw1yCnQVNQ%40mail.gmail.com
> <
> https://groups.google.com/d/msgid/cp2k/CAFnf9gq5trNSnP6HQUMzqBsMcPyyOxpCv89_zoTVYw1yCnQVNQ%40mail.gmail.com?utm_medium=email&utm_source=footer
> >.
>
> --
> 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+unsubscribe at googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/cp2k/ZR0P278MB07599C1DC6C06B7063436ACF9FF22%40ZR0P278MB0759.CHEP278.PROD.OUTLOOK.COM
> .
>

-- 
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+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAFnf9gqxMrX9rsDKfXs9Wjc9HstCgr3Q0-4ohB_6EbjTWyDfUg%40mail.gmail.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20240529/c3d68417/attachment-0001.htm>
Previous message (by thread): [CP2K-user] [CP2K:20253] OT convergence failure
Next message (by thread): [CP2K-user] [CP2K:20262] OT convergence failure
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
More information about the CP2K-user mailing list