[Wien] AFM NiO wrong magnetic moment

[Peter Blaha]

From:
Marcus Ekholm <marekh at ifm.liu.se>
Date:
12/09/2013 02:57 PM
To:
wien at zeus.theochem.tuwien.ac.at

Hello again,

Thank you for the struct file.
However I still suspect something is wrong.
Using LSDA I reproduce the equilibrium volume, but the magnetic moment 
is somewhat too low, 1.16, compared to 1.21 bohr magnetons as in your 
paper (PRB74:155108), and VASP calculations.

I use your struct file (same sphere sizes as in the paper) and having 
set up-down in the case.inst file, I initialise with:

init -b -sp -vxc 5 -numk 10000 -rkmax 9.0

and run it with

runsp_lapw -p -NI -in1new 7 -ec 0.000001

This is higher precision than what you used, and I've even tried with 
your settings, so I'm obviously doing something wrong.

I've attached my .scf file as well.
-------------------------------------------------

You are using:

NiO_AFII
R   LATTICE,NONEQUIV.ATOMS:  3 (166_R-3m )
MODE OF CALC=RELA unit=ang
   5.438490  5.43849026.6430489 90.000000 90.000000 120.000000

which are slightly different lattice parameters than what is in 
(PRB74:155108).

PS: Your scf file shows:
        MPI-parallel calculation using     2 processors
         Scalapack processors array (row,col):   2   1

        K=   0.02381   0.02381   0.02381            1
:RKM  : MATRIX SIZE  586LOs:  34  RKM= 9.00  WEIGHT= 2.00  PGR:


mpi.parallelization with just 2 cores is VERY ineffective and slower 
than the sequential run.
If you want to speed up things, use k-parallelism.
mpi-parallelism needs larger matrix-sizes (you have 500, but at 5000 it 
starts to get interesting to use mpi on a quadratic number of cores (4, 
16,64,...)




-- 

                                       P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-165300             FAX: +43-1-58801-165982
Email: blaha at theochem.tuwien.ac.at    WWW: 
http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------