[CP2K-user] [CP2K:11875] Restarting cell/geo opt project

[Patrick Gono]

Dear Kazimieras,

The file printed in the &MOTION section, the one with the .restart
extension is, technically, just a regular CP2K input file. It is printed
for you to be able to restart a calculation that involves any technique
specified in the &MOTION section (be it MD, GEO_OPT, CELL_OPT, NEB, etc.)
. It is an input file with all the parameters set to values that have been
reached in the last simulation step.

On the other hand, you have the wave function restart file (extension
RESTART.wfn) , which is not a CP2K input but rather a binary file that
holds the information about the wave function. The printing of this file is
specified in the &FORCE_EVAL section. It allows you to restart the SCF loop
(the solution of the Kohn-Sham equations) without having to recalculate the
wave functions from scratch.

In your input file, in the &FORCE_EVAL section, you can specify the name of
the wave function restart file. If you run a, say, GEO_OPT calculation, and
it finished at step 69, and a .restart file is written, the same wave
function restart file will be specified as it was in your original input.
Hence, if you submit a CP2K calculation starting from the .restart file, it
will look for the wave function file in the chosen directory. If there is
one present, it will start the solution of the KS equations from it. This
can usually save you a few minutes in most scenarios.

I include a sample input file for cell optimization, and a .restart file
that was generated at some point during the optimization procedure. Look at
both files and verify that they specify the same simulation with the only
differences being that the .restart file also holds the position of the
atoms from the last step, the cell parameters from the last step, and that
the current step (STEP_START_VAL) in the .restart file is set 12 instead of
the default 1. If I wanted to start the whole cell optimization from
scratch, I would run a job with test.inp as the input file. If I wanted to
restart the cell optimization from the 12th step onward, I would run a job
with test-1.restart as the input file. In both input files, you will see
the same test-RESTART.wfn wave function restart file is specified. Both
calculation would therefore look for this file and, if present, use it as
an initial guess for the KS solution in the next CELL_OPT step.

Yours sincerely,
Patrick Gono

On Wed, 12 Jun 2019 at 11:57, kazimieras tamoliunas <k.tam... at gmail.com>
wrote:

> Thank you! That's certainly a big part of my problem, but now i have a few
> other questions:
> 1. How do I restart the calculation with both the .restart file and with
> restart.wfn file? Or do you say the former file ( the one including atom
> coordinates, velocities, cell parameters, etc ) will be sufficient and i
> will obtain the same wavefunction without an increase in computational cost?
> 2. I don't need to change initial cell parameters specified in the input
> file?
>
> Regards,
>  Kazimieras
>
> --
> You received this message because you are subscribed to the Google Groups
> "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to cp... at googlegroups.com.
> To post to this group, send email to cp... at googlegroups.com.
> Visit this group at https://groups.google.com/group/cp2k.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/cp2k/ed46da3e-cd65-4c49-b799-d1f3b7c68c5c%40googlegroups.com
> <https://groups.google.com/d/msgid/cp2k/ed46da3e-cd65-4c49-b799-d1f3b7c68c5c%40googlegroups.com?utm_medium=email&utm_source=footer>
> .
> For more options, visit https://groups.google.com/d/optout.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20190612/db29706a/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test.inp
Type: application/octet-stream
Size: 2060 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20190612/db29706a/attachment.obj>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test-1.restart
Type: application/octet-stream
Size: 5052 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20190612/db29706a/attachment-0001.obj>