[CP2K-user] [CP2K:15370] Re: Problem converge Cu bulk calculations using kpoints

[Lucas Bandeira]

Thank you, Nicholas and Lucas. It worked now!

It worked when I only swapped the WAVEFUNCTION flag from REAL to COMPLEX.
But the calculations were faster when I changed both the EXTRAPOLATION and
EXTRAPOLATION_ORDER flags to the default value in cp2k, too.

Thank you both very much.


On Fri, May 14, 2021 at 11:27 PM Lucas Lodeiro <eluni... at gmail.com>
wrote:

> Hi !
>
> Nicholas' advice is very important. Your output shows an
> oscillating behavior, and one reason could be the extrapolation. Also,
> there could be more reasons.
> One is related with the metallic character of Cu, a smearing of 300K could
> be so low, I would try with 2000K.
> Another reason could be the REAL option on the kpoints wavefunctions... I
> would use the COMPLEX option.
> If these advices do not work, you could explore the settings on the BFGS
> optimizer... for example decreasing the TRUST_RADIUS.
>
> Regards - Lucas Lodeiro
>
> El vie, 14 may 2021 a las 23:02, Nicholas Winner (<nwi... at berkeley.edu>)
> escribió:
>
>> Try removing the lines
>>
>> EXTRAPOLATION USE_GUESS
>> EXTRAPOLATION_ORDER 1
>>
>> Use guess will generally not be a good choice. The defaults should work
>> well. Also consider trying the DZVP basis set first, which should be easier
>> to converge, and moving to TZV2P only if necessary.
>>
>> -Nick
>>
>> On Friday, May 14, 2021 at 3:02:21 PM UTC-7 ban... at gmail.com wrote:
>>
>>> Dear cp2k community,
>>>
>>> I am performing some calculations for Cu bulk using kpoints. However,
>>> any of the calculations that I have performed converged. It seems rather
>>> that the results diverge. In addition, the cell vectors printed at the
>>> beginning of every optimization step diverge from the experimental value
>>> that is 3.64Angs. Before the optimization, I have performed convergence
>>> tests for kpoints, and based on these, I chose the kpoint grid 8x8x8. I
>>> also tested different optimizers: LBFGS, BFGS, and conjugate gradient, but
>>> for all them I have the same problem. Could somebody help me with this?
>>>
>>> Attached are the input file used for the calculations with BFGS, and the
>>> output file obtained.
>>>
>>> Yours sincerely,
>>>
>>> Lucas Bandeira
>>>
>> --
>> 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 view this discussion on the web visit
>> https://groups.google.com/d/msgid/cp2k/9affdc05-6a56-4bdb-9bcd-0631c312ce10n%40googlegroups.com
>> <https://groups.google.com/d/msgid/cp2k/9affdc05-6a56-4bdb-9bcd-0631c312ce10n%40googlegroups.com?utm_medium=email&utm_source=footer>
>> .
>>
> --
> 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 view this discussion on the web visit
> https://groups.google.com/d/msgid/cp2k/CAOFT4P%2B3OmGbybFNhEY0JsT3OWJvEGcCkzY2%3DOc9-8W8okva3w%40mail.gmail.com
> <https://groups.google.com/d/msgid/cp2k/CAOFT4P%2B3OmGbybFNhEY0JsT3OWJvEGcCkzY2%3DOc9-8W8okva3w%40mail.gmail.com?utm_medium=email&utm_source=footer>
> .
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210515/7aa028f0/attachment.htm>