[CP2K-user] [CP2K:17619] Re: CP2K+SIRIUS with dispersion functionals

[Mathieu Taillefumier]

Hi,

The vdw functionaliity supported by SIRIUS can be activated with the 
option  VDW_FUNCTIONAL FUNC_VDWDF (or one the two others options) that you 
have to add in the parameters section (*c.f. the attached file*).  I  
picked a different pp file but it does not matter much. 
You will need to use the current development version of cp2k though. 

If you already use it, then a* git pull -r* and *make .... *will be needed. 
the VDW support in SIRIUS was not on by default previously but you 
mentioned you compiled SIRIUS with it. 

I would still consider this as experimental. The library seems robust but 
we never really tested deeply so any comment / problem reporting is 
welcome. 

I also remarked that you are doing some lattice relaxation. If you 
encounter some odd behavior could you report them here so I can have a 
look. I would really appreciate to get this information back.

Best

M.
Le lundi 5 septembre 2022 à 10:09:08 UTC+2, Dialogicus a écrit :

> Thank you for your advice. Unfortunately, it doesn't work. CP2K can't 
> identify FUNC_VDWDF as a keyword or subsection. The out file says
>
>
> Possible matches for unknown subsection 
>
> FUNC_VDWDF
>
>    enum BEEFVDW in section 
> %__ROOT__%FORCE_EVAL%DFT%KG_METHOD%XC%XC_FUNCTIONAL for keyword 
> _SECTION_PARAMETERS_ score:  103
>    subsection LDA_C_VWN in section 
> %__ROOT__%FORCE_EVAL%DFT%KG_METHOD%XC%XC_FUNCTIONAL score:  103
>    subsection MGGA_C_DLDF in section 
> %__ROOT__%FORCE_EVAL%DFT%KG_METHOD%XC%XC_FUNCTIONAL score:  103
>    subsection GGA_C_CCDF in section 
> %__ROOT__%FORCE_EVAL%DFT%KG_METHOD%XC%XC_FUNCTIONAL score:  103
>    enum BEEFVDW in section 
> %__ROOT__%FORCE_EVAL%DFT%ENERGY_CORRECTION%XC%XC_FUNCTIONAL for keyword 
> _SECTION_PARAMETERS_ score:  103
>
>
>  *******************************************************************************
>  *   ___                                                                   
>     *
>  *  /   \                                                                 
>      *
>  * [ABORT]                                                                 
>     *
>  *  \___/         unknown subsection FUNC_VDWDF of section XC_FUNCTIONAL   
>     *
>  *    |                                                                   
>      *
>  *  O/|                                                                   
>      *
>  * /| |                                                                   
>      *
>  * / \                                               
> input/input_parsing.F:187 *
>
>  *******************************************************************************
>
>
> As far as I understand, this error means invalid syntax for CP2K, but not 
> for SIRIUS. So, is there any way to send information about vdW functional 
> directly to SIRIUS run (e.g. SIRIUS includes this functional automatically 
> in each run)?
>
> --
> D.
>
>
> Hi,
>>
>> CP2K implementation of vdw is completely independent of libvdwxc which is 
>> used in SIRIUS. It means in practice that 
>>
>>    - there is no direct path in place in cp2k to parse the VDW_POTENTIAL 
>>    section and find the relevant functional supported in libvdwxc (and 
>>    SIRIUS).It should probably be added if possible.
>>    - the VDW_POTENTIAL section is simply ignored when SIRIUS is 
>>    initialized. It means that the calculations are correct but simply ignore 
>>    the vdw corrections.
>>    - CP2K documentation is misleading
>>
>> If you wish to try a vdwxc functional with SIRIUS, then have a look at 
>> the functional supported by libvdwxc (
>> https://libvdwxc.gitlab.io/libvdwxc/background.html#van-der-waals-forces) 
>> that fit your needs and then add it to the list of functional 
>> XC_FUNC_VDWDF, XC_FUNC_VDWDF2, or XC_FUNC_VDWDFCX to the libxc section of 
>> your input file. 
>>
>> something like that might actually work.
>>
>> &XC_FUNCTIONAL
>>         &GGA_X_PBE_SOL
>>         &END GGA_X_PBE_SOL
>>         &GGA_C_PBE_SOL
>>         &END GGA_C_PBE_SOL
>>         &FUNC_VDWDF
>>         &END FUNC_VDWDF
>> &END XC_FUNCTIONAL
>>
>> let me know if it works or not.
>>
>> Best
>>
>> M.
>>
>> Le vendredi 2 septembre 2022 à 11:02:29 UTC+2, Dialogicus a écrit :
>>
>>> Dear CP2K users, 
>>>
>>> As CP2K manual says it’s possible to use Van der Waals functionals in 
>>> the plane-wave calculations with SIRIUS.
>>>
>>> I tried to run Fm3m gold cell optimization using pseudopotential PW 
>>> method. VdW interactions were taken into account in VDW_POTENTIAL 
>>> subsection with NON_LOCAL type of dispersion functional. However, it seems 
>>> there’s no difference between calculations with and without dispersion 
>>> correction. I suspect the vdW functionals don’t work in my case.
>>>
>>> My CP2K and SIRIUS are both linked with libvdwxc library. 
>>>
>>> I don’t know what I’m doing wrong. Is any ideas?
>>>
>>> My input and output files (with and without vdW correction) are attached 
>>> below.
>>>
>>

-- 
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 cp2k+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/c6729c8c-5da6-4da5-876e-ae89a06a4835n%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20220906/8e028c90/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Au_Fm3m_vdW_opt.inp
Type: chemical/x-gamess-input
Size: 2325 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20220906/8e028c90/attachment.inp>