[CP2K:1798] Re: molecule falling apart during md run

Teodoro Laino teodor... at gmail.com
Wed Feb 25 10:32:01 UTC 2009


Dear Abu,

After 20 seconds of rutime.. I see this behavior:

 SCF WAVEFUNCTION OPTIMIZATION

  Step  Update method              Time         Convergence        Total 
energy
  
-----------------------------------------------------------------------------
     1  Mixing/Diag.   0.40E+00      2.069   294.5038702280     
-329.5423433821
     2  Mixing/Diag.   0.40E+00      3.080   159.4206153201     
-329.1328110607
     3  Mixing/Diag.   0.40E+00      3.120   105.4668986783     
-328.0498845729
     4  Mixing/Diag.   0.40E+00      3.182  1095.2138736182     
-319.1742682576
     5  Mixing/Diag.   0.40E+00      3.271  3261.5549040157     
-293.8773371347
     6  Mixing/Diag.   0.40E+00      3.207   743.4708241420     
-265.4738605357
     7  Mixing/Diag.   0.40E+00      3.259   453.2710960401     
-280.2406017615
     8  Mixing/Diag.   0.40E+00      3.213   174.2535774355     
-277.2695854146
     9  Mixing/Diag.   0.40E+00      3.304  2405.2506058034     
-272.8353211907
    10  Mixing/Diag.   0.40E+00      3.304  1820.7298891716     
-267.0406646300
    11  Mixing/Diag.   0.40E+00      3.082  1033.7945891346     
-270.2292200792
    12  Mixing/Diag.   0.40E+00      3.171  1342.9047137583     
-263.2869416906
    13  Mixing/Diag.   0.40E+00      3.268  1397.7632067855     
-269.6408037154
    14  Mixing/Diag.   0.40E+00      3.199  1438.4026792467     
-262.5345384047
    15  Mixing/Diag.   0.40E+00      3.241   747.2931276111     
-268.6729157993
    16  Mixing/Diag.   0.40E+00      3.141   981.6542877413     
-262.7054740768
    17  Mixing/Diag.   0.40E+00      3.202  1071.2871648587     
-268.3759861758


there's no need to go further..
Your wavefunction is never converging!!!!!!!!! If it does not converge 
the forces are wrong and the system, most of the times, explodes..
Please have a look at OT, if you want to converge more easily. There are 
plenty of input files in the
test directory of CP2K (benchmark directory for example).
Anyway my suggestion is that you still need to play a bit more with the 
code.
If the molecule is "falling apart", it is because the number you get are 
trash.

Cheers
Teo


Abu Naser wrote:
> Hi Teo,
>
> Yes, I did have the molecule. I did not include to save space. The 
> molecule was:
>     Ni      14.16100        1.65400       10.32900
>     P       12.82300        3.32000       10.20200
>     P       15.64100        2.85600       11.32900
>     P       13.47400       -0.09700       11.35800
>     P       14.40000        0.51500        8.50300
>     C       11.04600        3.24300       10.66700
>     C       10.33000        2.10500       10.28800
>     C        8.99600        1.94300       10.65100
>     C        8.35700        2.91900       11.40000
>     C        9.05700        4.05200       11.79100
>     C       10.39000        4.21600       11.42000
>     C       12.75600        4.30400        8.63800
>     C       11.59700        4.89500        8.12600
>     C       11.63200        5.62800        6.94200
>     C       12.81700        5.79300        6.25100
>     C       13.97800        5.21600        6.74500
>     C       13.93800        4.47300        7.91500
>     C       13.47300        4.61100       11.37600
>     C       15.00700        4.60600       11.32400
>     C       15.90400        2.54400       13.14100
>     C       16.84800        3.44000       13.92400
>     C       17.35900        3.16600       10.72700
>     C       18.23400        1.91400       10.69200
>     C       11.93700       -0.12700       12.36000
>     C       11.70200        0.96500       13.19900
>     C       10.52600        1.06400       13.93800
>     C        9.55600        0.08300       13.82800
>     C        9.76200       -1.00500       13.00000
>     C       10.95000       -1.11900       12.27500
>                                                            
> With regards,
>
> Abu
>
> 2009/2/25 Teodoro Laino <teodor... at gmail.com 
> <mailto:teodor... at gmail.com>>
>
>
>     Dear Abu,
>     I wander if you have a molecule at all ?!
>     Why is the coordinate section empty?
>
>     please before posting read carefully this post:
>     http://groups.google.com/group/cp2k/web/general-rules-for-posting-help-requests
>
>     Teo
>
>     Abu Naser wrote:
>     > Hi All,
>     >
>     > I have been playing around with cp2k. When I run an MD simulation
>     > using the following input file- molecule is falling apart.
>     > I have been wondering where could be problem.
>     >
>     > With regard,
>     >
>     > Abu
>     >
>     >
>     ##############################################################################
>     > &FORCE_EVAL
>     >   METHOD Quickstep
>     >   &DFT
>     >     &MGRID
>     >       CUTOFF 300
>     >     &END MGRID
>     >     &QS
>     >     &END QS
>     >     &SCF
>     >       SCF_GUESS ATOMIC
>     >     &END SCF
>     >     &XC
>     >       &XC_FUNCTIONAL Pade
>     >       &END XC_FUNCTIONAL
>     >     &END XC
>     >   &END DFT
>     >   &SUBSYS
>     >     &CELL
>     >       ABC 10.0 10.0 10.0
>     >       UNIT ANGSTROM
>     >     &END CELL
>     >     &COORD
>     >
>     >     &END COORD
>     >     &KIND H
>     >       BASIS_SET DZVP-GTH-PADE
>     >       POTENTIAL GTH-PADE-q1
>     >     &END KIND
>     >     &KIND C
>     >       BASIS_SET DZVP-GTH-PADE
>     >       POTENTIAL GTH-PADE-q4
>     >     &END KIND
>     >     &KIND P
>     >       BASIS_SET DZVP-GTH-PADE
>     >       POTENTIAL GTH-PADE-q5
>     >     &END KIND
>     >     &KIND Ni
>     >       BASIS_SET DZV-GTH-PADE-CONFINED
>     >       POTENTIAL GTH-PADE-q18
>     >     &END KIND
>     >   &END SUBSYS
>     > &END FORCE_EVAL
>     > &GLOBAL
>     >
>     >   PROJECT hop
>     >   RUN_TYPE MD
>     >   PRINT_LEVEL MEDIUM
>     > &END GLOBAL
>     > &MOTION
>     >   &MD
>     >     ENSEMBLE NVE
>     >     STEPS 20000
>     >     TIMESTEP 0.05
>     >     TEMPERATURE 300.0
>     >   &END MD
>     > &END MOTION
>     >
>     >
>     ############################################################################
>     > >
>
>
>
>
>
> >




More information about the CP2K-user mailing list