[Pw_forum] problem with scf in ruthenium - textbook example

Kucukbenli Emine emine.kucukbenli at epfl.ch
Wed Apr 17 13:08:48 CEST 2013


Dear Alexander,

On top of Gabriele's points I would like to add that 
in my experience, this is a rather hard pseudo.
In the past, for a very tight convergence, I used a rather zealous 800Ry for density expansion.
and had everything working. I did not optimize the ecutwfc. 

I am not saying one should go up to that value to make it converge,
especially not for a learning exercise,
but perhaps playing around with cutoffs is what you need,
on top of Gabriele's advice.
Below is an input that works for me.
ciao
emine kucukbenli, postdoc @ theos, epfl, switzerland

---------------------------------------------
 &system
    ibrav= 14, ntyp=1,
    occupations='smearing', degauss=0.01
smearing= 'marzari-vanderbilt' 
A= 2.76249
B= 2.76249
C= 4.35709
COSBC= 6.12323e-17
COSAC= 6.12323e-17
COSAB= 0.5
nat=2
  ecutwfc= 200
 /
 &electrons
    startingwfc="atomic"
    diagonalization='david'
    mixing_mode = 'plain'
    mixing_beta = 0.7
    conv_thr =  1.0d-8
 /
ATOMIC_SPECIES
Ru 1.00 Ru.pbe-spn-kjpaw_psl.0.3.0.UPF
ATOMIC_POSITIONS crystal
Ru 0 0 0
Ru 0.33333 0.33333 0.50000
K_POINTS AUTOMATIC
18 18 12 1 1 1



More information about the users mailing list