[Pw_forum] Problem with "nmix_ph" in ph.x

[Vic Bermudez]

Hello All,

	I'm running ph.x in QE vers. 5.0 and having a bit of a convergence problem.
Reducing the value of alpha_mix(1) from the default value of 0.7 helps, but
if I try to change "nmix_ph" the program crashes immediately with an error
message saying "wrong nmix used". Is this a bug in vers. 5.0, or am I doing
something wrong ? Here's my input file, with the nmix_ph entry commented
out. In case it's significant, the input structure for this calculation was
relaxed using PBE/PAW, and the force and scf convergence criteria were set
at 1.d-5 and 1.d-10 respectively.

title_line='Phonon calc. '
&INPUTPH
outdir='/lustre/cmf/scratch/b/bermudez/XX-input',
iverbosity=0,
epsil=.TRUE.,
trans=.TRUE.,
tr2_ph=1e-14,
alpha_mix(1)=0.3,
! nmix_ph=6,
max_seconds=8.63900d5,
recover=.false.
/
0.0 0.0 0.0

	Thanks in advance for your help.


Victor M. Bermudez
E-mail: bermudez at alum.mit.edu