[QE-users] convergence NOT achieved after 100 iterations for spin-orbit coupling
Kazume NISHIDATE
nisidate at iwate-u.ac.jp
Fri Mar 12 11:45:48 CET 2021
Hi, Ream:
You should use smal value for the mixed_beta.
Attached is the example of the input file of your system which reached
to the first scf state at the 50 iterations.
========
convergence has been achieved in 50 iterations
Entering Dynamics: iteration = 1
========
This calculation was made by the NEC's LS parallel computer at the
TOHOKU university.
Note that I omitted the nspin=4 flag. It is not the the default flag
in the noncollinear calculation.
Here is the QE manual page on the issue of 'nspin=4' setting:
> nspin = 4 : spin-polarized calculation, noncollinear
> (magnetization in generic direction)
> DO NOT specify nspin in this case; specify
> noncolin=.TRUE. instead
Also see the line 1 of the page.4 of the following PDF file.
It stated that:
> - nspin:- the noncolin and nspin are conflicting flags, both should
> not be present within the same calculation
I do not know the correctness of the above statement, but you should
try this first.
I show you only the first part of the input file since the
other is the same of your input file. I can not describe the modifications,
but these meaning may be clear following the manual.
========
&CONTROL
calculation = 'relax' ,
restart_mode = 'from_scratch' ,
pseudo_dir = '../pseudo' ,
outdir = './work',
prefix= 'gr-sc-n',
tstress = .true. ,
tprnfor = .true. ,
forc_conv_thr = 1.00e-03,
wf_collect=.true.,
nstep = 200,
/
&SYSTEM
ibrav = 0,
celldm(1) = 27.89615,
nat = 66,
ntyp = 3,
starting_magnetization(1)= 0.5,
noncolin= .true. ,
lspinorb= .true. ,
ecutwfc = 60.0 ,
ecutrho = 600.0 ,
nbnd = 280,
occupations = 'smearing' ,
degauss = 0.02,
smearing = 'gaussian' ,
/
&ELECTRONS
conv_thr = 5.0e-8 ,
mixing_mode = 'plain' ,
mixing_beta = 0.25,
/
==============
best regards
kazume NISHIDATE
敬具 西館数芽
nisidate at iwate-u.ac.jp
kazume.nishidate at gmail.com
More information about the users
mailing list