vibration frequency intensity

[jia... at vt.edu]

Dear All,

I recently try to calculate the vibration frequency using the following 
code:

I used the fully optimized geometry with the criterion:

   &GEO_OPT

      TYPE MINIMIZATION

      MAX_DR    3.0E-03

      MAX_FORCE 0.00005

      RMS_DR    1.5E-03

      RMS_FORCE 3.0E-04

      MAX_ITER 2000

      OPTIMIZER CG

      &CG

        &LINE_SEARCH

          TYPE 2PNT

        &END LINE_SEARCH

      &END CG

   &END GEO_OPT

   &SCF

      ....

      EPS_SCF    1.0E-6

      ....

   &END SCF


When do the vibration_analysis I change the &SCF  &EPS_SCF to 1.0E-8  

  &VIBRATIONAL_ANALYSIS

   INTENSITIES

   NPROC_REP 16

   DX 0.001

   &MODE_SELECTIVE

     &INVOLVED_ATOMS

       INVOLVED_ATOMS 1 2 3

     &END INVOLVED_ATOMS

   &END MODE_SELECTIVE

   &PRINT

     &PROGRAM_RUN_INFO ON

     &END

   &END

 &END


Submit the job using 2 nodes(1node has 32 processors).

1. I found that with different NPROC_REP value (16, 32,64) it gives similar 
frequency but with quite different intensity. How does the NPROC_REP 
affects the intensity?

2. In the output,  

....

MS| ITERATION STEP        1

MS| TRACKED MODE      313.813098cm-1  NOT  CONVERGED 

......

 FREQUENCY AND CONVERGENCE LIST

       FREQUENCY    INTENSITY    MAXVAL CRITERIA    NORM CRITERIA   
 CONVERGENCE

  VIB|    58.079   582.068111      0.263E-06          0.443E-06           
YES

  VIB|   348.828   591.399560      0.101E-05          0.303E-05           NO

  VIB|   579.845     1.641373      0.249E-05          0.419E-05           NO


If it is not converged, it means the system doesn't have this frequency? 



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20170227/bd1c263d/attachment.htm>