Noisy potential with LTRUNCATE
Posted: Wed May 13, 2026 1:49 pm
We have noticed a bug in the latest release for the 2D Coulomb kernel truncation. You can find details about this bug in our known issues list. The code produces a noisy potential in VASP 6.6, which is different from the expected behaviour, which is present in VASP 6.5. We are sorry for the inconvenience.
Here is an example between VASP 6.6 (orange) and VASP 6.5 (blue). Thank you to Harini S. at IIT Bombay for reporting this to us and producing the image.
This is due to line 914 in src/pot_electrostat.F:
Code: Select all
call setunb(potential,gridc)should be replaced with:
Code: Select all
call setunb_compat(potential,gridc)We will add a patch here in the near future.