[Pw_forum] grid phonon calculation at non-gamma point cannot reach convergence

[yuan liang]

Dear QE developers and users,

Recently I'm trying to calculate phonon dos via ph.x, however, although I
can get fine convergence result at gamma point, ph.x cannot reach
convergence within 100 steps and stop automaticly at the 1st non gamma
point on my mesh.

The material is YVO3, orthorhombic, with 4 Y atom, 4 V atom and 12 O atom
per unit cell.

The QE version is 5.4.0

I tried both GGA PBEsol USPP and LDA USPP(from THEOS SSSP)

Are there any methods to solve this problem? Any suggestion will be welcome.

Here are my input files:
*scf calculation*
&control
   calculation= 'scf',
   restart_mode= 'from_scratch',
   prefix='YVO3_exc',
   pseudo_dir = '...',
   outdir= '...',
   wf_collect=.true.,
   tprnfor=.true., tstress=.true.
/
&system
 ibrav = 8,
 celldm(1) = 9.9893,
 celldm(2) = 1.0581,
 celldm(3) = 1.4296,
 nat = 20,
 ntyp = 3,
 ecutwfc = 60,
 ecutrho = 480,
 ecfixed = 60,
 qcutz = 20,
 q2sigma = 5,
 occupations= 'smearing', smearing='gauss', degauss=0.01,
/
&electrons
 mixing_beta = 0.4,
 mixing_mode = 'plain',
 conv_thr = 1.0d-10,
/
ATOMIC_SPECIES
 Y 88.906 y_lda_v1.4.uspp.F.UPF
 V 50.942 v_lda_v1.4.uspp.F.UPF
 O 15.999 o_lda_v1.2.uspp.F.UPF
ATOMIC_POSITIONS crystal
 Y 0.97830 0.07150 0.25000
 ...
 O 0.18720 0.20040 0.55663
K_POINTS automatic
 6 6 4 0 0 0
*phonon calculation on grid*
&inputph
 prefix='YVO3_exc',
 outdir='...',
 recover=.true.,
 *ldisp=.true*.,
 nq1=6,
 nq2=6,
 nq3=4,
 tr2_ph=1.0d-14,
 fildyn='YVO3_dyn.G',
 amass(1)=80.906,
 amass(2)=50.942,
 amass(3)=15.999,
/
0.000000 0.000000 0.000000

Best regards,

Yuan Liang
Physics Department,
University of Virginia
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20171002/34357723/attachment.html>