[Pw_forum] Fwd: Error in Hybrid Exc calculations

Giuseppe Mattioli giuseppe.mattioli at ism.cnr.it
Tue Sep 13 23:03:34 CEST 2016


Dear user (please sign always the posts with name and scientific affiliation)

There is generally no strict reason in using an US pseudopotential for  
Ge, and in this case US is detrimental because 1) the EXX calculation  
is still less stable for US and 2) ecutfock is forced to assume the  
ecutrho value and you don't gain any speedup. I suggest you to try  
with a NC pseudopotential.
HTH
Giuseppe

Quoting Debnath Talukdar <debnath.talukdar at students.iiserpune.ac.in>:

> ---------- Forwarded message ----------
> From: Debnath Talukdar <debnath.talukdar at students.iiserpune.ac.in>
> Date: Tue, Sep 13, 2016 at 4:25 PM
> Subject: Error in Hybrid Exc calculations
> To: PWSCF Forum <pw_forum at pwscf.org>
>
>
> Hi,
>    I was trying to do hybrid calculations in quantum espresso. The
> calculation stopped by printing the following message
>
>     Error in routine electrons (1):
>      dexx is negative!   Check that exxdiv_treatment is appropriate for the
> system
>
> As I know that in hybrid Exc calculation to solve the poisson equation in
> reciprocal space their is
> fourier transform of density from real to q space. That colomb potential in
> poisson equation is diverge at q=0. To remove this divergencies in coulomb
> potential in q space is adds a small G to
> denomenetor(q+G); known as   gygi-baldereschi method, but it is
> appropriate  for cubic and quasi cubic supercell (Gygi-Baldereschi PRB 34,
> 4405 (1986)).
>
> At first I have used  gygi-baldereschi method for my system that is
> hexagonal metallic but I have found above error then I have changed this
> method from 'gygi-baldereschi' to 'vcut_ws' that is appropriate for
> anisotropic supercell but I still got that same error message.
>
> After that I have changed Hybrid functional from HSE to gau-PBE which used
> exxdiv_treatment='none' and x_gamma_extrapolation=.false. but I still got
> that same error
>  message again.
>
>
> Can anybody give any suggestion that how can I solve this problem ?
>
> I have also doubt about following...
>
> 1)As we all know that hybrid calculation is very expansive so how can I
> reduce different   parameter in Hybrid calculation as compared to normal
> scf calculations.
>
>    Such that in my normal scf with DFT Exc I have used k -mesh 21x21x1, scf
> threshold 10-8Ry,
>    then how can I reduce these parameters in Hybrid calculations; is it
> arbitrary?
>
> 2)How can I reduce ecutfock to speed up the calculation. Currently I have
> used ecutfock=ecutrho (for my system ecutrho=10*ecutwfc), is their is any
> way to reduce this kinetic energy cutoff for fock (exchange) operator at
> the cost of loss in accuracy in my calculations.
>
> 3) In our pwscf input file we give explicitly effect of
> exchange-correlation functional through card
> input_dft, I am doubt that how can I gave both vdW(rev-vdw-b86r) functional
> with hybrid calculations(where input_dft='HSE').
>
> I have read papers but I am unable to solve these issues can anyone will
> give any other papers that how can I solve this.
>
> Below I have attached input, output files of my calculations.
> 1)PRB 34, 4405 (1986)
> 2)PHYSICAL REVIEW B 80, 115205 ͑2009͒
> 3)PHYSICAL REVIEW B 74, 121102͑R͒ ͑2006͒
> 4)JCP 118, 8207 (2003)
> 5)THE JOURNAL OF CHEMICAL PHYSICS 134, 024116 (2011)
>
>
>
>
> Thanks


-- 
********************************************************
- Article premier - Les hommes naissent et demeurent
libres et égaux en droits. Les distinctions sociales
ne peuvent être fondées que sur l'utilité commune
- Article 2 - Le but de toute association politique
est la conservation des droits naturels et
imprescriptibles de l'homme. Ces droits sont la liberté,
la propriété, la sûreté et la résistance à l'oppression.
********************************************************

    Giuseppe Mattioli
    CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
    v. Salaria Km 29,300 - C.P. 10
    I 00015 - Monterotondo Stazione (RM)
    Tel + 39 06 90672836 - Fax +39 06 90672316
    E-mail: <giuseppe.mattioli at ism.cnr.it>
    http://www.ism.cnr.it/en/staff/giuseppe-mattioli/
    ResearcherID: F-6308-2012





More information about the users mailing list