Implementing new functionals

[Matt W]

Hi,

I think the probable cause is the MT Poisson solver - it needs box length 
2x the maximum distance between electron density in any given direction - 
so for your box you can only safely go out to ~10 A. 

Probably your actual binding curve is at the smaller distance, but you 
can't see it for the huge noise from the spurious results at long distance.

Matt

On Thursday, May 1, 2014 4:02:44 AM UTC+1, August Melcher wrote:
>
> I'm working on producing a neon dimer potential energy curve, and I keep 
> getting results that look like this. I don't seem to understand how 
> coordinates work because I keep seeing the binding part of the curve at 
> what I expected to be the largest separation. Here is a sample input of 
> what I thought would be a 1.5A separation. The difference in the x 
> coordinates is increased correspondingly for the rest of the points. See 
> attached for what the curve looks like.
>
> &FORCE_EVAL
>
 

>   METHOD Quickstep
>   &DFT
>     BASIS_SET_FILE_NAME 
> /global/u1/a/amelcher/soft/cp2k/cp2k/tests/QS/EMSL_BASIS_SETS
>     POTENTIAL_FILE_NAME 
> /global/u1/a/amelcher/soft/cp2k/cp2k/tests/QS/POTENTIAL
>     &MGRID
>       CUTOFF 1000
>       NGRIDS 5
>       REL_CUTOFF 50
>     &END MGRID
>     &QS
>       METHOD GAPW
>       EPS_PGF_ORB 1.0E-12
>       EPS_FILTER_MATRIX 0.0e0
>     &END QS
>     &POISSON
>       PERIODIC NONE
>       PSOLVER MT
>     &END
>     &SCF
>       EPS_SCF 1.0E-8
>       SCF_GUESS ATOMIC
>       MAX_SCF 40
>       &OUTER_SCF
>         EPS_SCF 1.0E-8
>         MAX_SCF 50
>       &END OUTER_SCF
>       &OT ON
>       &END
>     &END SCF
>     &XC
>       &XC_FUNCTIONAL
>         &BECKE97
>             PARAMETRIZATION  wB97X-V
>             SCALE_X 1.0
>             SCALE_C 1.0
>         &END
>       &END XC_FUNCTIONAL
>       &HF
>         &SCREENING
>            EPS_SCHWARZ 1.0E-14
>         &END
>         &INTERACTION_POTENTIAL
>            POTENTIAL_TYPE MIX_CL
>            SCALE_COULOMB   0.833
>            SCALE_LONGRANGE 0.167
>            OMEGA 0.30
>         &END
>         &MEMORY
>           MAX_MEMORY 128
>         &END
>         #FRACTION 0.167
>       &END
>       &vdW_POTENTIAL
>          DISPERSION_FUNCTIONAL NON_LOCAL
>          &NON_LOCAL
>            TYPE RVV10
>            PARAMETERS 6.3 0.0093
>            VERBOSE_OUTPUT
>            KERNEL_FILE_NAME 
> /global/u1/a/amelcher/soft/cp2k/cp2k/tests/QS/rVV10_kernel_table.dat
>            CUTOFF  80
>          &END NON_LOCAL
>       &END vdW_POTENTIAL
>    &END XC
>   &END DFT
>   &SUBSYS
>     &CELL
>       ABC 20.0 15.0 15.0
>       PERIODIC NONE
>     &END CELL
>     &COORD
> Ne       13.25 0.0000 0.0000
> Ne       11.75 0.0000 0.0000
>     &END COORD
>     &KIND Ne
>       BASIS_SET aug-cc-pVQZ
>       LEBEDEV_GRID 590
>       RADIAL_GRID 250
>       POTENTIAL ALL
>     &END KIND
>   &END SUBSYS
> &END FORCE_EVAL
> &GLOBAL
>   PROJECT wb97xv15
>   PRINT_LEVEL LOW
>   RUN_TYPE ENERGY
>   &TIMINGS
>     THRESHOLD 0.000000001
>   &END
> &END GLOBAL
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140501/b4b9bcf3/attachment.htm>