[QE-developers] ?==?utf-8?q? Problem to gather phonons with multiple images

Antoine Jay ajay at laas.fr
Mon Jun 29 18:46:40 CEST 2020


Dear Paolo,

No it is not the same problem.
In my case the other representations are well finished without any crash.
Only the q points for which the representations have been splitted between several images are missing their  final dyn file.
Running ph.x with a restart solve the problem as it gathers the informations and create the missing dyn files in a few seconds.

Regards,

Antoine Jay


Le Lundi, Juin 29, 2020 17:19 CEST, Paolo Giannozzi <p.giannozzi at gmail.com> a écrit:
 Is it maybe the same problem mentioned here?   https://gitlab.com/QEF/q-e/-/issues/45Paolo On Sun, Jun 28, 2020 at 1:22 PM Antoine Jay <ajay at laas.fr> wrote:Dear developers,
I think there is a problem in the phonon code on several q points when it is used with multiple images:
sometimes, the last q point is separated between the images:
     Atomic displacements:
     There are  12 irreducible representations
     Representation     1      1 modes -A'  To be done
     Representation     2      1 modes -A'  To be done
     Representation     3      1 modes -A'  To be done
     Representation     4      1 modes -A'  Not done in this run
     Representation     5      1 modes -A'  Not done in this run
     Representation     6      1 modes -A'  Not done in this run
     Representation     7      1 modes -A''  Not done in this run
     Representation     8      1 modes -A''  Not done in this run
     Representation     9      1 modes -A''  Not done in this run
     Representation    10      1 modes -A''  Not done in this run
     Representation    11      1 modes -A''  Not done in this run
     Representation    12      1 modes -A''  Not done in this run

The programm well stop when the 12 representations are done, but the informations are not gathered into the final dyn file. All the dyn files are present exept the one corresponding to this q point.

Bellow a script that generate this error.

Regards,


Antoine Jay
LAAS-CNRS


# how to run executables
PH_COMMAND="$PARA_PREFIX $BIN_DIR/ph.x -ni 2 -nk 6"
PW_COMMAND="$PARA_PREFIX $BIN_DIR/pw.x -nk 6"
name=GaN

# self-consistent calculation
    cat > $name.scf.in << EOF
 &control
    calculation     = 'scf'
    prefix             = '$name',
    pseudo_dir     = '$PSEUDO_DIR/',
    outdir             = '$TMP_DIR/'
 /
 &system
    ibrav              = 4,
    celldm(1)       = 6.09308327
    celldm(3)       = 1.62987734
    nat                 = 4,
    ntyp               = 2,
    ecutwfc          = 80.0,
    nbnd              = 25
 /
 &electrons
    diagonalization = 'david'
    mixing_mode    = 'plain'
    mixing_beta      = 0.7
    conv_thr            = 1.0d-11
 /
ATOMIC_SPECIES
N 14.0067 N.pbe-n-kjpaw_psl.1.0.0.UPF
Ga 69.723 Ga.pbe-dn-kjpaw_psl.1.0.0.UPF
ATOMIC_POSITIONS (crystal)
Ga            0.0000000000        0.0000000000        0.0000000000    0   0   0
Ga            0.6666666667        0.3333333333        0.5000000000    0   0   0
N             0.0000000000        0.0000000000        0.3762844446    0   0   1
N             0.6666666667        0.3333333333        0.8762844446    0   0   1
K_POINTS {automatic}
12 12 12 0 0 0
EOF
$PW_COMMAND < GaN.scf.in > GaN.scf.out

# phonon calculation on a (666) uniform grid of q-points
cat > $name.ph.in << EOF
phonons of $name
 &inputph
  tr2_ph      = 1.0d-14,
  prefix       = '$name',
  ldisp         = .true.,
  nq1          = 6,
  nq2          = 6,
  nq3          = 6,
  amass(1) = 14.0067,
  amass(2) = 69.723,
  outdir      = '$TMP_DIR/',
  fildyn       = '$name.dyn',
 /
EOF
$PH_COMMAND < $name.ph.in > $name.ph.out

  _______________________________________________
developers mailing list
developers at lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/developers

--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/20200629/2ef96bd2/attachment-0001.html>


More information about the developers mailing list