The specified OLD file cannot be opened

[Yiyang]

Dear Matt,

Thank you for your correction! Now the script finally works.

For this test system, the cut off distance is set very small, due to the 
short lattice parameter in c direction (2.5A).
For a real run, I will use a cutoff distance at least 6A (< 0.5*L_cell).

If dispersion correction is needed in the XC section, I would try to use

--------------------------------------------------------------------
&VDW_POTENTIAL
     POTENTIAL_TYPE                             PAIR_POTENTIAL
     &PAIR_POTENTIAL
         TYPE                                               DFTD3 
         PARAMETER_FILE_NAME            dftd3.dat 
         REFERENCE_FUNCTIONAL         B3LYP
         #CALCULATE_C9_TERM              TRUE
         #LONG_RANGE_CORRECTION   FALSE
         R_CUTOFF [angstrom]                  16.0
     &END PAIR_POTENTIAL 
&END VDW_POTENTIAL
-------------------------------------------------------------


I wish this piece of script is appropriate for the pe.inp script.


Besides, I find that for this periodic system, I cannot use
  POTENTIAL_TYPE   COULOMB
in the HF section, which results in very long OT optimization process and 
never converges due to my test.
Is it the case that periodic Hartree Fock calculation must use truncated 
potential?

Thank you for your time.

Regard,
Yiyang


On Tuesday, October 25, 2016 at 12:49:09 PM UTC-7, Matt W wrote:
>
> Hi,
>
> you need to set the FIRST_SNAPSHOT variable to 1 (or higher). 
>
> BTW, your settings for exact exchange in the B3LYP functional are not good 
> - you truncate the coulomb interaction very hard, but don't include a long 
> range correction.
>
> Matt
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20161025/3def8705/attachment.htm>