large total drift
Posted: Wed Oct 01, 2014 2:10 pm
Hello,
while trying to relax a CePt3Si structure (so far nonmagnetic, nonrelativistic) I can get
almost zero forces, but fairly large 'total drift' wrt. the forces.
I read in the VASP manual that the drift should always be watched to become smaller
than the forces. Since I don't know the very meaning of 'total drift' (and it is afaik
not really explained in the manual) I want to ask what is going on.
For instance:
POSITION TOTAL-FORCE (eV/Angst)
-----------------------------------------------------------------------------------
1.99171 1.99171 3.99280 -0.000000 0.000000 -0.000028
0.00000 0.00000 3.22907 0.000000 0.000000 0.000000
1.99171 0.00000 1.32553 -0.000000 0.000000 0.000006
0.00000 1.99171 1.32553 -0.000000 -0.000000 0.000006
0.00000 0.00000 0.03755 -0.000000 -0.000000 0.000015
-----------------------------------------------------------------------------------
total drift: -0.000000 0.000000 0.000685
The question is twofold:
1) What does the 'large' drift mean?
2) Should I try to change anything to decrease it? What should I then look for?
(Many of my parameters seem to be rather stringent as of now already.)
Thanks!
Rudolf Sykora
while trying to relax a CePt3Si structure (so far nonmagnetic, nonrelativistic) I can get
almost zero forces, but fairly large 'total drift' wrt. the forces.
I read in the VASP manual that the drift should always be watched to become smaller
than the forces. Since I don't know the very meaning of 'total drift' (and it is afaik
not really explained in the manual) I want to ask what is going on.
For instance:
POSITION TOTAL-FORCE (eV/Angst)
-----------------------------------------------------------------------------------
1.99171 1.99171 3.99280 -0.000000 0.000000 -0.000028
0.00000 0.00000 3.22907 0.000000 0.000000 0.000000
1.99171 0.00000 1.32553 -0.000000 0.000000 0.000006
0.00000 1.99171 1.32553 -0.000000 -0.000000 0.000006
0.00000 0.00000 0.03755 -0.000000 -0.000000 0.000015
-----------------------------------------------------------------------------------
total drift: -0.000000 0.000000 0.000685
The question is twofold:
1) What does the 'large' drift mean?
2) Should I try to change anything to decrease it? What should I then look for?
(Many of my parameters seem to be rather stringent as of now already.)
Thanks!
Rudolf Sykora