NELM and algo=chi

Problems running VASP: crashes, internal errors, "wrong" results.

Moderators: Global Moderator, Moderator

Post Reply
Message
Author
oelerich
Newbie
Newbie
Posts: 1
Joined: Wed Feb 28, 2018 10:19 am
License Nr.: 5-2082

NELM and algo=chi

#1 Post by oelerich » Thu Apr 12, 2018 12:19 pm

Hi Vasp experts,

I am trying to calculate the frequency dependent dielectric response with VASP 5.4.4 including local fields. I basically did what is explained here, but I'm running into the issue that the response is calculated over and over again. The OUTCAR simply grows, while the results of the response don't seem to change.

I found, that the response is calculated NELM times, and NELM is by default 60. Is there any reason to recalculate the response several times, when the calculations was previously converged with algo=exact and loptics=True?

Is the workaround of setting NELM=1 reasonable?

Thank you,
Jan

Post Reply