[Pw_forum] vc-relax + vdW-DF2. PWSCF v.6.1
Ronald Cohen
rcohen at carnegiescience.edu
Wed Nov 1 09:57:01 CET 2017
This problem seems to arise if the moments go to zero (low spin solution). Perhaps it is dividing by zero (or a small number) as the moments decrease. With a large moment (high spin solution)
the stresses seem to converge fine with -D__SPIN_BALANCED.
---
Ronald Cohen
Extreme Materials Initiative
Geophysical Laboratory
Carnegie Institution
5251 Broad Branch Rd., N.W.
Washington, D.C. 20015
rcohen at carnegiescience.edu <mailto:rcohen at carnegiescience.edu>
office: 202-478-8937
skype: ronaldcohen
> On 31. Oct 2017, at 21:27, Ronald Cohen <rcohen at carnegiescience.edu> wrote:
>
> In 6.2 compiled with -DSPIN_BALANCED rvv10 seems to work fine with DFT+U but DF2 with spin_balanced gives stresses that do not converge. The error is not large but finite:
> df2 dft+u
> total stress (Ry/bohr**3) (kbar) P= 24.12
> total stress (Ry/bohr**3) (kbar) P= 24.96
> total stress (Ry/bohr**3) (kbar) P= 25.25
> total stress (Ry/bohr**3) (kbar) P= 25.47
> total stress (Ry/bohr**3) (kbar) P= 25.80
> total stress (Ry/bohr**3) (kbar) P= 25.41
> total stress (Ry/bohr**3) (kbar) P= 25.43
> total stress (Ry/bohr**3) (kbar) P= 25.30
> total stress (Ry/bohr**3) (kbar) P= 25.28
> total stress (Ry/bohr**3) (kbar) P= 25.13
> total stress (Ry/bohr**3) (kbar) P= 25.33
> total stress (Ry/bohr**3) (kbar) P= 25.44
> total stress (Ry/bohr**3) (kbar) P= 25.49
> rvv10 (same system and setup) converges fine (still running):
> total stress (Ry/bohr**3) (kbar) P= 105.54
> total stress (Ry/bohr**3) (kbar) P= -36.11
> total stress (Ry/bohr**3) (kbar) P= -37.82
> total stress (Ry/bohr**3) (kbar) P= 5.90
> total stress (Ry/bohr**3) (kbar) P= -2.78
> total stress (Ry/bohr**3) (kbar) P= -8.11
> total stress (Ry/bohr**3) (kbar) P= -5.48
> total stress (Ry/bohr**3) (kbar) P= -1.83
>
> ---
> Ronald Cohen
> Extreme Materials Initiative
> Geophysical Laboratory
> Carnegie Institution
> 5251 Broad Branch Rd., N.W.
> Washington, D.C. 20015
> rcohen at carnegiescience.edu <mailto:rcohen at carnegiescience.edu>
> office: 202-478-8937
> skype: ronaldcohen
>
>> On 27 Oct 2017, at 15:10, Paolo Giannozzi <p.giannozzi at gmail.com <mailto:p.giannozzi at gmail.com>> wrote:
>>
>> Please delete Modules/xc_vdW_DF.o and PW/src/stres_nonloc_dft.o, recompile with -D__SPIN_BALANCED
>>
>> Paolo
>>
>> On Fri, Oct 27, 2017 at 1:34 PM, Ronald Cohen <rcohen at carnegiescience.edu <mailto:rcohen at carnegiescience.edu>> wrote:
>> I still get:
>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>> Error in routine stres_vdW_DF (1):
>> vdW stress not implemented for nspin > 1
>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>>
>> stopping ...
>> Performing spin-balanced Ecnl stress calculation!
>>
>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>> Error in routine stres_vdW_DF (1):
>> vdW stress not implemented for nspin > 1
>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>>
>> stopping ...
>> ---
>> Ronald Cohen
>> Extreme Materials Initiative
>> Geophysical Laboratory
>> Carnegie Institution
>> 5251 Broad Branch Rd., N.W. <https://maps.google.com/?q=5251+Broad+Branch+Rd.,+N.W.+Washington,+D.C.+20015&entry=gmail&source=g>
>> Washington, D.C. 20015 <https://maps.google.com/?q=5251+Broad+Branch+Rd.,+N.W.+Washington,+D.C.+20015&entry=gmail&source=g>
>> rcohen at carnegiescience.edu <mailto:rcohen at carnegiescience.edu>
>> office: 202-478-8937 <tel:(202)%20478-8937>
>> skype: ronaldcohen
>>
>>> On 27 Oct 2017, at 07:14, Prof. Dr. Ronald Cohen <r.cohen at lmu.de <mailto:r.cohen at lmu.de>> wrote:
>>>
>>> Hi! I compiled qe 6.2 with -D__SPIN_BALANCE but I still get:
>>> from stres_vdW_DF : error # 1
>>> vdW stress not implemented for nspin > 1
>>> Should I just take out that error in the code and let it try to run?
>>>
>>> Attached is my input and output. Thanks!
>>>
>>> Ron
>>>
>>> ---
>>> Professor Dr. Ronald Cohen
>>> Ludwig Maximilians Universität
>>> Theresienstrasse 41 Room <https://maps.google.com/?q=Theresienstrasse+41+Room%C2%A0207&entry=gmail&source=g> 207 <https://maps.google.com/?q=Theresienstrasse+41+Room%C2%A0207&entry=gmail&source=g>
>>> Department für Geo- und Umweltwissenschaften
>>> München
>>> 80333
>>> Deutschland
>>>
>>> office: +49 089 2180 4294 <tel:+49%2089%2021804294>
>>> ronald.cohen at min.uni-muenchen.de <mailto:ronald.cohen at min.uni-muenchen.de>
>>> skypename: ronaldcohen
>>>
>>>
>>>> On 24 Oct 2017, at 11:27, Bogdan Yavorskyy <Bogdan.Yavorskyy at lrz.uni-muenchen.de <mailto:Bogdan.Yavorskyy at lrz.uni-muenchen.de>> wrote:
>>>>
>>>>
>>>>
>>>>> Begin forwarded message:
>>>>>
>>>>> From: Paolo Giannozzi <p.giannozzi at gmail.com <mailto:p.giannozzi at gmail.com>>
>>>>> Subject: Re: [Pw_forum] vc-relax + vdW-DF2. PWSCF v.6.1
>>>>> Date: 24 October 2017 at 17:24:41 CEST
>>>>> To: PWSCF Forum <pw_forum at pwscf.org <mailto:pw_forum at pwscf.org>>
>>>>> Cc: bogdan.yavorskyy at lrz.uni-muenchen.de <mailto:bogdan.yavorskyy at lrz.uni-muenchen.de>
>>>>>
>>>>> In the 6.2 version, it is possible to use the "spin-balanced" approach, setting the -D__SPIN_BALANCED precompilation flag at compile time. This allows to compute the stress. Some explanation by Per Hyldgaard:
>>>>>
>>>>> "There is an Occam's razor argument for keeping the full spin vdW-DF formulation (from the PRL 115, 136402 (2015)) as standard: that the vdW-DF method be used and thus tested without options that lies outside the logic of the construction. In the present vdW-DF framework the aforementioned PRL give the one spin vdW-DF formulation that is compatible with spin scaling of exchange.
>>>>>
>>>>> On the other hand, one can certainly vc-relax in the spin-balanced approach and then do sanity checks on the resulting structure afterwards."
>>>>>
>>>>> On Wed, Oct 18, 2017 at 5:03 PM, Bogdan Yavorskyy <Bogdan.Yavorskyy at lrz.uni-muenchen.de <mailto:Bogdan.Yavorskyy at lrz.uni-muenchen.de>> wrote:
>>>>> Dear Developers,
>>>>>
>>>>> in the PWSCF code v.6.1 the possibility to relax cell parameters (vc-relax) together with van der Waals functional, input_dft=' vdW-DF2’
>>>>> for magnetic systems (nspin=2) is not implemented. While this option is highly desirable for my research, I would kindly ask You if You are going to include this possibility in the future version?
>>>>>
>>>>> Many thanks in advance,
>>>>> Bogdan Yavorskyy.
>>>>>
>>>>> _______________________________________________
>>>>> Pw_forum mailing list
>>>>> Pw_forum at pwscf.org <mailto:Pw_forum at pwscf.org>
>>>>> http://pwscf.org/mailman/listinfo/pw_forum <http://pwscf.org/mailman/listinfo/pw_forum>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
>>>>> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy <https://maps.google.com/?q=Udine,+via+delle+Scienze+208,+33100+Udine,+Italy&entry=gmail&source=g>
>>>>> Phone +39-0432-558216 <tel:+39%200432%20558216>, fax +39-0432-558222 <tel:+39%200432%20558222>
>>>>>
>>>>
>>> <F10C20H10Fe2.vcrelax.in <http://f10c20h10fe2.vcrelax.in/>>
>>> <F10C20H10Fe2.vcrelax.out>
>>
>>
>>
>>
>> --
>> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
>> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
>> Phone +39-0432-558216, fax +39-0432-558222
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20171101/f287ef27/attachment.html>
More information about the users
mailing list