EFIELD bug in VASP 6.1.2

[nfchilton]

We are running optimisations in various electric fields, starting from a relaxed structure (all degrees of freedom, ISIF=3) with no electric field. We find that for any choice of magnitude, eg 10^-5 though 10^-9 ev/Å, that the optimised structures are all the same as one-another, but different from the zero field original relaxation. A test where we set EFIELD=0 (otherwise, all other other parameters are the same), shows that at the first optimisation step the total drift vector is 718 4187 -14011 ev/Å. Clearly, some enormous electric field is being applied.

Can anyone comment?

[merzuk.kaltak]

Hello,

please upload an error report, so that we can look into it.
This report should include POTCAR, POSCAR, KPOINTS, INCAR and at least OUTCAR.