non-local vdW and stress_tensor

Wei wei.a... at googlemail.com
Tue Feb 18 17:32:19 UTC 2014


Dear developers,

we are interested in the non-local vdW functional (e.g. vdW-DF2), and we 
can run the testing input files without problem. 
/tests/QS/regtest-dft-vdw-corr/argon*.inp

However, the code (yesterday's SVN) always crashes after 1 or 2 
GEO_OPT/MD steps, if stress_tensor is calculated AND orbital 
transformation (OT) technique is used. Note, for a SCF calculation, 
there is no problem; also GEO_OPT/MD can run well if either 
STRESS_TENSOR or OT is NOT turned on. Since we would like to apply it 
for very large system, OT is important. Therefore, we are wondering if 
the stress_tensor is calculated accurately with non-local vdW functional 
at a fixed nuclear configuration (with OT). Of course it will be great 
if this issue with GEO_OPT/MD can be solved.

Thanks a lot in advance for your help!

The error and input file is provided at the bottom of this mail.

Best Regards,

Wei

Error :

   ----------------------------------- OT 
---------------------------------------

   Step     Update method      Time    Convergence         Total 
energy    Change
------------------------------------------------------------------------------
      1 OT CG       0.15E+00    1.0     0.00011163 -85.0324161489 -8.50E+01

  ****************************************************************************
  *** 17:45:22 ERRORL2 in 
qs_dispersion_nonloc:calculate_dispersion_nonloc ***
  *** processor 0  :: err=-300 condition FAILED at line 
221                ***
  ****************************************************************************


  ===== Routine Calling Stack =====

            12 calculate_dispersion_nonloc
            11 qs_vxc_create
            10 qs_ks_build_kohn_sham_matrix
             9 qs_ks_update_qs_env
             8 scf_env_do_scf_inner_loop
             7 scf_env_do_scf
             6 qs_energies_scf
             5 qs_forces
             4 cp_eval_at
             3 geoopt_bfgs
             2 cp_geo_opt
             1 CP2K
  CP2K| condition FAILED at line 221
  CP2K| Abnormal program termination, stopped by process number 0
--------------------------------------------------------------------------

Input.  few modifications on argon05.inp

&GLOBAL
   PROJECT argon
   RUN_TYPE GEO_OPT
   PRINT_LEVEL LOW
&END GLOBAL
&FORCE_EVAL
   METHOD QS
   STRESS_TENSOR ANALYTICAL
   &DFT
     BASIS_SET_FILE_NAME ../BASIS_MOLOPT
     POTENTIAL_FILE_NAME ../POTENTIAL
     &MGRID
       CUTOFF 200
     &END MGRID
     &QS
       METHOD GPW
     &END QS
     &SCF
       SCF_GUESS ATOMIC
       MAX_SCF 10
       EPS_SCF 1.0e-4
       &OT ON
       &END OT
     &END SCF
     &XC
       &XC_FUNCTIONAL
         &PBE
           PARAMETRIZATION revPBE
           SCALE_C 0.0
         &END PBE
         &VWN
         &END VWN
       &END XC_FUNCTIONAL
       &vdW_POTENTIAL
          DISPERSION_FUNCTIONAL NON_LOCAL
          &NON_LOCAL
            TYPE DRSLL
            VERBOSE_OUTPUT
            KERNEL_FILE_NAME ../vdW_kernel_table.dat
            CUTOFF  80
          &END NON_LOCAL
       &END vdW_POTENTIAL
     &END XC
   &END DFT
   &SUBSYS
     &CELL
       ABC 5.3 5.3 5.3
       ALPHA_BETA_GAMMA 90 90 90
     &END CELL
     &COORD
       SCALED .TRUE.
       Ar 0.1 0.0 0.0
       Ar 0.5 0.5 0.0
       Ar 0.5 0.0 0.5
       Ar 0.0 0.5 0.5
     &END COORD
     &KIND Ar
       BASIS_SET DZVP-MOLOPT-SR-GTH-q8
       POTENTIAL GTH-PADE-q8
     &END
   &END SUBSYS
&END FORCE_EVAL
&MOTION
   &GEO_OPT
     MAX_ITER 2
   &END GEO_OPT
&END MOTION




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140218/d9885cd1/attachment.htm>


More information about the CP2K-user mailing list