My Community

I set "Prec=High",but the wrap around error still exist from the titled WARNING.What is stranger to me is that,there is some ENCUT value,where the warning message could be avoided.But when changing this value larger or smaller,the Warning message exist again.I didn't know why and its effect on the total energy.Any advice would be appreciated.

I think the source for this warning is that not all integers are allowed for
the FFT grid numbers. Say, 17 is expected, but it's not allowed.
In this case, VASP will use 16 instead. That means a small "wrap around
error". So, no need to be over-worried on this issue.

the required precision depends somewhat on the problem you want to investigate (stress tensors, changes in cell shape and/or volume,... require very high precision to give reliable results)
the safest way is to set
PREC=Accurate
ENCUT=1.3*max(ENMAX) in your POTCAR or even higher
A good check is to have a look at the total drift in the forces written in OUTCAR:
it should not be larger that the forces acting on the individual ions themselves.

I know that PREC=High is not encouraged anymore, however the warning about wrap around errors can be avoided if it is used, because it uses a denser FFT grid than PREC=accurate. At the moment I am using ENCUT~1.8ENMAX, and PREC=accurate, and I still get the warning. The suggested NGX,Y,Z values in the OUTCAR file are close to the ones that PREC=accurate gives, but I am tempted to use PREC=High, to avoid them completely. The total drift as you suggest looks ok, but I am still uncomfortable when I see warnings. Am I wrong to be worried?