[Q-e-developers] Fwd: [Pw_forum] gamma approximation does not work for hybrid

[Paolo Giannozzi]

---------- Forwarded message ----------
From: Paolo Giannozzi <p.giannozzi at gmail.com>
Date: Thu, Nov 17, 2016 at 2:52 PM
Subject: Re: [Pw_forum] gamma approximation does not work for hybrid
To: PWSCF Forum <pw_forum at pwscf.org>


Just to let everybody know: there is a bug in hybrids+Gamma
point+LSDA+k-point parallelization (with 2 pools, that is, spin-up and
spin-down bands). Anybody who is affected may download a hopefully fixed
version from here: http://www.qe-forge.org/snapshots/ . This version should
also fix a nasty bug in CP with nonlocal functionals.

Paolo

On Wed, Nov 9, 2016 at 1:17 PM, Tiana Davide <davide.tiana at epfl.ch> wrote:

> Dear all,
>
> I just want to report a problems I found running hybrid calculations
> (pbe0, hse, but also gaupbe).
> In practice if you run using KPOINTS{gamma} pw crashes with this error:
>
>      Error in routine electrons (1):
>      dexx is negative!   Check that exxdiv_treatment is appropriate for
> the system
>
> However, running at gamma point without approximation KPOINT(automatic) 1
> 1 1 0 0 0 everything goes smoothly and pw got slowly to convergence.
>
> Considering the speed-up one can get using the gamma approximation I was
> wondering if it can this be fixed or if somehow this is normal and simply
> hybrid functional cannot be run using gamma approx
>
> Cheers
> Davide
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>



-- 
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




-- 
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/developers/attachments/20161117/1369ce1a/attachment.html>