Recommendation to update LHFCALC docs

Queries about input and output files, running specific calculations, etc.


Moderators: Global Moderator, Moderator

Post Reply
Message
Author
asrosen
Newbie
Newbie
Posts: 27
Joined: Wed Oct 18, 2023 4:51 pm

Recommendation to update LHFCALC docs

#1 Post by asrosen » Thu Oct 16, 2025 10:56 pm

Currently, the LHFCALC docs (https://www.vasp.at/wiki/LHFCALC) state the following:

It is recommended to select a "direct optimization" algorithm for HF/DFT hybrid functonal type calculations, i.e., ALGO=Damped (IALGO=53) or ALGO=All (IALGO=58) in the INCAR file.

The blocked-Davidson algorithm ALGO=Normal is, with certain caveat, also supported, whereas calculations for the other algorithms (ALGO=Fast) are not properly supported (note: no warning is printed). The blocked-Davidson algorithm (ALGO=Normal) is generally rather slow, and in many cases the Pulay mixer will be unable to determine the proper ground-state.

However, as of VASP 6, LFOCKACE (https://www.vasp.at/wiki/index.php/LFOCKACE) enables a 3x speedup when ALGO = Normal. As such, the recommendation to users to use ALGO = Damped or ALGO=All for hybrid calculations seems not ideal. Since few readers probably find the LFOCKACE docs, they may not be taking advantage of the significant speedup enabled by using ALGO = Normal.


martin.schlipf
Global Moderator
Global Moderator
Posts: 639
Joined: Fri Nov 08, 2019 7:18 am

Re: Recommendation to update LHFCALC docs

#2 Post by martin.schlipf » Fri Oct 17, 2025 6:32 am

Thank you for the recommendation. I updated the section to mention the availability of LFOCKACE. I believe the direct optimizers are still more robust than the Davidson one but if ALGO=Normal works it is a good choice too.

Martin Schlipf
VASP developer


Post Reply