<html><body>
<p>Dear PWSCF users and developers.<br>
<br>
I'm now using PWSCF in order to calculate the electric state of the <br>
<br>
magnetic insulator. The material is ferri magnetic compounds and <br>
<br>
has 64 atom in the unit cell.<br>
<br>
We set LDA+U calculation. The results are not converged. <br>
<br>
I have also experience that the convergence is not achieved <br>
<br>
in anti-ferromagnetic materials with impurity.<br>
<br>
How we should treat such a difficult convergence materials, and <br>
<br>
in general what parameters should we corrected ? <br>
<br>
A parts of my input is as follow.<br>
<br>
&system<br>
ibrav=4<br>
celldm(1)=20.0<br>
celldm(3)=3.93844842285<br>
nat=64<br>
ntyp=4<br>
nbnd=320<br>
ecutwfc =50.0<br>
ecutrho = 500.0<br>
nspin = 2<br>
starting_magnetization(1) = 0.0<br>
starting_magnetization(2) = 3.0<br>
starting_magnetization(3) = -3.0<br>
starting_magnetization(4) = 0.0<br>
degauss=0.01<br>
lda_plus_u = .TRUE.<br>
occupations='smearing'<br>
smearing='gauss'<br>
degauss=0.05<br>
Hubbard_U(2) = 5.0 <br>
Hubbard_U(3) = 5.0<br>
/<br>
&electrons<br>
conv_thr = 1e-7,<br>
mixing_mode = 'plain'<br>
mixing_beta=0.1,<br>
/<br>
<br>
Sincerely,<br>
<br>
Yukihiro Okuno.</body></html>