[Q-e-developers] Phonon matdyn problem1

[Antoine JAY]

Dear developers,

I had trouble performing phonon DOS on a nq=5^3 grid:
frc_blk','wrong total_weight'

This is due to the fact that the condition
ABS(total_weight-nr1*nr2*nr3)    PHonon/PH/matdyn.f90/subroutinefrc_bkl
is not filled.

In fact, nr1, nr2 and nr3 are well 5 in this subroutine as asked in the
input file ph.in through the read_ifc_param subroutine.
But the total_weight seems to be calculated through parameters defined in
the q_gen subroutine (matdyn.f90).
In this subroutine, nr1 n2 and nr3 are fixe:
INTEGER, PARAMETER:: nr1=4, nr2=4, nr3=4
which give in my example, a tota_weight that is 124.1, far from 125 (5^3)

When I changed in q_gen the 4^3 to 5^3 values, I obtained the good weight
of 125 and was able to run properly the phonon DOS.

I think the nr1, nr2 and nr3 that are use in q_gen must be the same as the
one ask by the user and not fixed parameters.

here are the input files and as attached the direct calcul on the 27
points (4 tar files <8Mo in 4 mails)

For scf:
cat > $name.$Pression.scf.in << EOF
 &control
    calculation = 'scf',
    prefix      = '$name.$Pression',
    tstress     = .true.,
    tprnfor     = .true.,
    pseudo_dir  = '$PSEUDO_DIR/',
    outdir      = '$TMP_DIR/',
    etot_conv_thr=1.0d-6,
    forc_conv_thr=1.0d-4,
    nstep=1400,
    wf_collect=.true.
/
&system
    ibrav = 8,
    ibrav =8
    celldm(1) = 8.7945600
    celldm(2) = 1.115149
    celldm(3) = 1.3518333
    occupations='smearing',
    smearing='mp',
    degauss=0.0015,
    nat=  28,
    ntyp= 5,
    ecutwfc =80,
/
 &electrons
    diagonalization='david'
    mixing_beta = 0.7
    conv_thr =  1.0d-13
    electron_maxstep=800
/
ATOMIC_SPECIES
B1   10.810  B.gga.pw91.UPF
B2   10.810  B.gga.pw91.UPF
B3   10.810  B.gga.pw91.UPF
B4   10.810  B.gga.pw91.UPF
B5   10.810  B.gga.pw91.UPF
ATOMIC_POSITIONS {crystal}
B1       0.171061682   0.468596962   0.000000000
B2       0.157692029   0.720137600   0.629558413
B3       0.345741206   0.910230223   0.783563420
B4       0.348540590   0.723689473   0.000000000
B5       0.163279542   0.990279732   0.000000000
B1       0.828938318   0.531403038   0.000000000
B1       0.328938378   0.968596756   0.500000000
B1       0.671061622   0.031403244   0.500000000
B2       0.842307971   0.279862400   0.629558413
B2       0.342308178   0.220137754   0.870441624
B2       0.657691822   0.779862246   0.870441624
B2       0.842307971   0.279862400   0.370441587
B2       0.157692029   0.720137600   0.370441587
B2       0.657691822   0.779862246   0.129558376
B2       0.342308178   0.220137754   0.129558376
B3       0.654258794   0.089769777   0.783563420
B3       0.154258705   0.410230070   0.716436519
B3       0.845741295   0.589769930   0.716436519
B3       0.654258794   0.089769777   0.216436580
B3       0.345741206   0.910230223   0.216436580
B3       0.845741295   0.589769930   0.283563481
B3       0.154258705   0.410230070   0.283563481
B4       0.651459410   0.276310527   0.000000000
B4       0.151459094   0.223689467   0.500000000
B4       0.848540906   0.776310533   0.500000000
B5       0.836720458   0.009720268   0.000000000
B5       0.336720334   0.490279317   0.500000000
B5       0.663279666   0.509720683   0.500000000
K_POINTS {automatic}
6 6 6 1 1 1
EOF
    $ECHO "  running the scf calculation for $name at $Pression kbar...\c"
    $PW_COMMAND < $name.$Pression.scf.in > $name.$Pression.scf.out
    $ECHO " done"

for direct phonons:
cat > $name.$Pression.ph.in << EOF
phonon of $name
 &inputph
  recover=.true.
  tr2_ph=1.0d-14,
  prefix='$name.$Pression',
  amass(1)=10.810,
  amass(2)=10.810,
  amass(3)=10.810,
  amass(4)=10.810,
  amass(5)=10.810,
  outdir='$TMP_DIR/'
  fildyn='$name.$Pression.dyn',
  fildrho='$name.$Pression.drho',
  ldisp=.true.,
  nq1=5, nq2=5, nq3=5,
 /
EOF

for q2R:
cat > q2r.in <<EOF
 &input
   fildyn='$name.$Pression.dyn',
   flfrc='$name.$Pression.fc'
   zasr='simple'
 /
EOF

for matdyn with N=30

cat > phdos.in <<EOF
 &input
  asr='crystal',
  dos=.true.,
  amass(1)=10.801,
  amass(2)=10.801,
  amass(3)=10.801,
  amass(4)=10.801,
  amass(5)=10.801,
  flfrc='$name.$Pression.fc',
  fldos='$name.$Pression.phdos.$N',
  nk1=$N,nk2=$N,nk3=$N,
 /
EOF



Thank you,

Antoine Jay



> Bonjour Antoine
>
> peux-tu stp envoyer un message a q-e-developer at qe-forge.org avec les
> donnees necessaires a resoudre (peut-etre) le probleme de matdyn.x que tu
> as rencontre'? notamment:
> - le fichier de constantes de force (si c'est grand, un link a un endroit
> ou' on puisse les retrouver, pas comme attachment)
> - les donnees de input de matdyn.x
> - une courte description du probleme
> Merci
>
> Paolo
> --
> 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 --------------
A non-text attachment was scrubbed...
Name: points1.tar
Type: application/x-tar
Size: 6225920 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/developers/attachments/20161118/d204f794/attachment.tar>