[Pw_forum] Segmentation fault in exx_stress routine

[Paolo Giannozzi]

Are you using different dimensions for the  "smooth" and "dense" FFT
grids (ecutrho > 4*ecutwfc) together with hybrid functionals? if so,
that part of code is invalid: nls(ig) should be replaced with nl(ig).
Not sure however whether this fixes all problems.

Background: until recently, calculations with hybrid functionals were
restricted to norm-conserving PPs, for which the "smooth" and "dense"
FFT grids coincide. Recently, hybrid  functionals have been implemented
as well for Ultrasoft PPs, for which the two grids usually differ; I do
not think however that the stress term has been extended to that case.

Paolo

On Thu, 2014-09-25 at 16:12 +0900, Hosik Lee wrote:
> Dear pw_forum members,
> 
> 
> I found sementation fault error during calculation of function
> exx_stress of QE ver 5.1. ZnO system was under calculation and the
> error does not occur with Si system. I tried to find the origin of
> segmentation error and found that rhoc(nls(ig)) in the following line
> of function exx_stress is the origin. The other variables like
> fac_tens, fac_stress, delta, fac do not show the segmentation error. 
> 
>                       DO ig = 1, ngm
>                           vc(:,:) = vc(:,:) +
> rhoc(nls(ig))*CONJG(rhoc(nls(ig))) * &
> 
> (fac_tens(:,:,ig)*fac_stress(ig)/2.d0 - delta(:,:)*fac(ig))
>                       ENDDO
> 
> 
> Since the stress is not required in my work, I tried to coment out the
> above lines and then works go well. Anyway, I just report the bug in
> rhoc array in exx_stress.  
> 
> 
> The followings are of input file for the calculation.
> 
> --  &control
>     calculation = 'relax'
>     tstress=.true.,
>     pseudo_dir='/home/p012lhs/src/pwscf/pseudo',
>     outdir='./'
>     tprnfor=.TRUE.,
>     nstep =300,
>     verbosity = 'high',
>     disk_io='low'
>  /
>  &system
>     ibrav=4, celldm(1)=6.12781734 ,
>     celldm(3)= 1.601998,
>     nat=4, ntyp=2,
>     ecutwfc = 30,
>     ecutrho = 240.,
>     nbnd = 40,
>     degauss = 0.0003,
>     occupations='smearing',
>     nspin=1,
>     input_dft=HSE,
>     exx_fraction=0.10,
>     nqx1 = 1, nqx2 = 1, nqx3 = 1,
>  /
>  &electrons
>   conv_thr=1.D+2,
>   mixing_beta=0.1,
>   diagonalization = 'david',
>  /
>     &IONS
>     trust_radius_max=0.2,
>     trust_radius_ini=0.2,
>     /
>    &CELL
>    cell_dynamics='damp-w' ,
>    wmass= 0.0005,
>    cell_factor=2.0,
>    /
> ATOMIC_SPECIES
>     Zn   65.38    zn_pbe_v1.uspp.F.UPF
>      O   15.999   O.pbe-rrkjus.UPF
> ATOMIC_POSITIONS {angstrom}
> Zn    0.00000    1.87217    0.00000
> Zn    1.62135    0.93609    2.59740
>  O    0.00000    1.87217    1.94805
>  O    1.62135    0.93609    4.54545
> K_POINTS {automatic}
> 2  2  2   0    0   0
> 
> 
> Sincerely,
> 
> Hosik
> 
> 
> ===================================
> Dr. Hosik Lee
> School of Mechanical and Neclear Engineering
> Ulsan National Institute of Sci. & Tech. (UNIST)
> Unist-gil 100 Eonyang-eup , Ulsan
> 689-798 South Korea
> contact: hosiklee at gmail.com
> ====================================
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum

-- 
 Paolo Giannozzi, Dept. Chemistry&Physics&Environment, 
 Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
 Phone +39-0432-558216, fax +39-0432-558222