My Community

hi, I can find "magnetization" for every atom at the bottom of the outcar file when I used VASP4.6. But using VASP5.2, I can not find it. The input files are the same. Could you give me some clues? Thanks


The incar file
-------------------------
SYSTEM = Pnma

Electronic Relaxation 1
NELM = 50
EDIFF = 1E-05 stopping-criterion for ELM

Magnetic stuff
ISPIN = 2
MAGMOM = 2*3 2*0
RWIGS = 4*1.35

Ionic Relaxation
NSW = 40 number of steps for ionic relaxation
IBRION = 2 quasi-Newtonian
ISIF = 4 forces, relax ions, calc stress
EDIFFG = 1E-04 convergence criterion for ir

If you run in parallel you will have to set NPAR = 1 in order for the magnetization to be written. There is a pronounced warning written at the end of the OUTCAR if you run in parallel and with NPAR > 1.

Regarding your RWIGS tag. Please notice that you only should supply one value per atomic type not one per atom. So unless you have four different atom types in your POTCAR file you should modify this line.

If not NPAR = 1 helps, then also try to see if setting LORBIT = 1 or 2 make any difference. However, I get a proper output with only NPAR = 1 and the RWIGS tag specified, so it should work.

Cheers,
/Dan
<span class='smallblacktext'>[ Edited Sun May 23 2010, 08:37PM ]</span>

to obtain the local magnetic moments you have to set LORBIT, the complete magnetization is written in OUTCAR if ISPIN = 2

forsdan
Thank you for your reply, helped me to solve my problem here in Uberlandia/Brasil [:d],
I was getting crazy because i was shure the compilation was right, fully with mkl, because i thought it would be a intels problem i downloaded all the open libs (blas, blacs, lapack, scalapack, fftw3) i did it all but the error persisted, when in fact, every thing we should have done was add NPAR=1 to de INCAR file. It seems to be the difference between 4.6 and 5.2 INCAR'S file. Thanks a lot