positive dE and d eps values with a direct optimization algorithm

[IBRAHIM]

Dear VASP Admin,

In our calculations using ALGO=All for the hybrid functional and starting from WAVECAR with the symmetry turned off, we obtained positive dE and/or d eps values (changes in total energy and eigenvalues) during electronic iterations. However, the calculation is electronically converged.

Is it normal to observe positive dE and/or d eps values during electronic iterations when using direct optimization algorithms?

Many thanks in advance,
Ibrahim

[pedro_melo]

Dear Ibrahim,

From you OSZICAR files it seems that a positive dE or deps value occurred only once. This can happen during the convergence loop, as there is no guarantee that it is a monotonous sequence. So long as the total energy converges to a lower value, your calculation should be fine.

Kind regards,
Pedro