questionable total drift of charged system

[kutu]

Hello everyone:

I'm working with a charged slab. I have set Algo=Very_Fast, Prec=accurate, IDIPOL=3 and LDIPOL=.TRUE. . EDIFF and DIPOL are default. The result energy is reliable, BUT in OUTCAR I found the TOTAL-FORCE and total drift are questionable:

POSITION TOTAL-FORCE (eV/Angst)
-----------------------------------------------------------------------------------
0.00000 0.00000 0.00000 0.000000 0.000000 25.342048
4.17390 0.00000 0.00000 0.000000 0.000000 25.342048
0.00000 2.95139 0.00000 0.000000 0.000000 25.342048
................(I ommit some lines here)
4.17390 2.95139 5.82702 0.000000 0.000000 -24.235818
0.00000 5.90279 5.82702 0.000000 0.000000 -24.235818
4.17390 5.90279 5.82702 0.000000 0.000000 -24.235818
-----------------------------------------------------------------------------------
total drift: 0.000000 0.068893 15.879340

The force and drift are too large.
It seems that my situation is different from the old post . I use PAW GGA and ENCUT=400 for Au atoms.

Can anyone explain how to interpret the large force and drift?

Thank you in advance

KuTu