when is a relaxation complete?

[pranjal222]

Dear All,

In some calculations, the calculations stop by stating "reached required accuracy - stopping structural energy minimisation" and in some other calculations it stops by mentioning "aborting loop because EDIFF is reached".

I am confused as does it mean that in both the cases, the relaxation is complete? If not, which one is correct and in case I must get the message reached required.... " but I keep getting the "aborting loop because EDIFF is reached", what should I do to get the "reached required accuracy - stopping structural energy minimisation" status.

I am highly confused in the meaning of the two phrases and want to know which is one is the right one to be aimed for a relaxed calculation. My INCAR file is something like this.

ISTART = 1
ICHARG = 2
INIWAV = 1
ENCUT = 500.00 eV
PREC = Normal
LREAL = Auto
ISMEAR = 0
SIGMA = 0.01
EDIFF = 1.0E-5
NELM = 200
GGA = PE
NCORE = 04
NBANDS = 260
ADDGRID = .TRUE.

##Relaxation
NSW = 200
##EDIFFG = -1.0E-4
IBRION = 1
POTIM = 0.5
ISIF = 7
ADDGRID = .TRUE.

##LOPTICS =.TRUE.
##CSHIFT = 0.1

Looking forward to your guidance.

Thank you.
Pranjal

[manuel_engel1]

The message

Code: Select all

aborting loop because EDIFF is reached

is emitted when the convergence criterion specified by the EDIFF INCAR tag is reached for the electronic minimization.

On the other hand, the message

Code: Select all

reached required accuracy - stopping structural energy minimisation

is emitted when the convergence criterion specified by the EDIFFG INCAR tag is reached for the ionic minimization (in your case, during IBRION=1). Usually during this process, multiple electronic minimization steps will be performed. You can follow this process in the OUTCAR file.

For the definition of both kinds of criteria, I refer you to the respective VASP-wiki pages.
Keep in mind that whether or not a calculation is properly converged still depends on the situation (for example, the default EDIFF might be too large for accurate force/phonon calculations).