My Community

Not a bug (but error message could be improved)

Also says " WARNING: dimensions on CHGCAR file are different "

It took me a while to find the source of this issue, and so I wanted to document it somewhere for others to find. When doing a band structure calculation, the precision and the energy cutoff must be the same as in the self-consistent run that generated the CHGCAR you are going to use in for band structure run.

Several other common tags can inadvertently cause this error as well.

1) As the above post suggests, make sure PREC is the same for both calculations, but don't just rely on automatic values for bands and cutoff energy. Instead, explicitly set ENCUT and NBANDS to the value contained in your SC calculation's OUTCAR file. Try using ISMEAR = 0.

2) Make sure your cell dimensions have not changed (e.g. if you are using ISIF=3). If they have, do a static (IBRION=-1, NSW=0) SC calculation to get a good CHGCAR.

3) Make sure your mesh hasn't changed (NGX, NGY, NGZ, NGXF, NGYF, NGZF). This part is important if you ever use tags like ADDGRID or ENAUG.

4) If all else fails, do a static self-consistent calculation and use the resulting CHGCAR in your NSC calculation.