[Pw_forum] Phonon calculation crash in orthorhombic cell.
Miriam Marques
mmarques at staffmail.ed.ac.uk
Fri Aug 17 13:46:50 CEST 2007
Dear PWSCF users,
I am trying to perform a phonon calculation on a uniform grid of
q-points for an orthorhombic metallic structure.
First, I did the scf calculation for a 16x16x16 Monkhorst-Pack grid
and tried to calculate phonons on a uniform (4x4x4) grid via ldisp,
but the program phonon stopped with the error "end of self consistent
calculation. No convergence has been achieved" in the representation 4
(mode 4) of the first q point different from gamma. (I am using
tr2_ph=1.0d-12)
Surprisingly, when I did the scf calculation for a 13x13x13
Monkhorst-Pack grid and then the phonon calculation via ldisp on a
4x4x2 grid, the program ran (it crashed again when I tried the 4X4x8
grid).
I attached the input for the first case.
# self-consistent calculation
cat > pcmcm.scf.in << EOF
&control
calculation='scf'
restart_mode='from_scratch',
pseudo_dir = '$PSEUDO_DIR/',
outdir='$TMP_DIR/',
prefix='pcmcm'
tprnfor = .true.
tstress = .true.
/
&system
ibrav= 9, celldm(1) =6.089228715122, celldm(2) =0.86757351,
celldm(3)= 2.5838341, nat= 4, ntyp= 1, nbnd= 20, ecutwfc = 30.0,
ecutrho=300, occupations='smearing', smearing='methfessel-paxton',
degauss=0.02
/
&electrons
diagonalization='david'
mixing_beta = 0.7
conv_thr = 1.0D-10
/
ATOMIC_SPECIES
P 30.973761 P.pw91-n-van.UPF
ATOMIC_POSITIONS crystal
P 0.0000000000 0.0000000000 0.00000000000
P 0.0000000000 0.0000000000 0.50000000000
P 0.0000000000 0.0000000000 0.25000000000
P 0.0000000000 0.0000000000 0.75000000000
K_POINTS {automatic}
16 16 16 0 0 0
EOF
$ECHO " running the scf calculation for P...\c"
$PW_COMMAND < pcmcm.scf.in > pcmcm.scf.out
$ECHO " done"
# phonon calculation on a (4 4 4) uniform grid of q-points
cat > pcmcm.ph.in << EOF
phonons of cmcm P
&inputph
tr2_ph=1.0d-12,
prefix='pcmcm',
ldisp=.true.,
nq1=4, nq2=4, nq3=4
amass(1)=30.973761,
outdir='$TMP_DIR/',
fildyn='pcmcm.dyn',
/
EOF
$ECHO " running the phonon calculation at Gamma...\c"
$PH_COMMAND < pcmcm.ph.in > pcmcm.ph.out
Thank you very much for your help,
Miriam.
----------------------
Miriam Marques
School of Physics
University of Edinburgh
----------------------
More information about the users
mailing list