ABINIT Discussion Forums

Hello Users/Developers:

There is a strange problem came into my sight. When merging different DDB files, the nkpt should be the same to all DDB files, except for Gamma.

However, on the other hand, when merging with qpt=0.000000000 0.000000000 0.000000000, it cannot be merged, the log file of mrgddb says

Comparing integers for variable nkpt.
Comparing integers for variable nkpt.
Value from input DDB is 34 and
Value from input DDB is 34 and
from transfer DDB is 64.
from transfer DDB is 64.
Action : check your DDBs.


This means DDB for Gamma using kptopt=2 (T-reversal symmetry) has less kpts (kptopt=3) at other finite in-equivalent q-points.

If I delete the DDB at q=0 obtained through kptopt=2, then the merging is successful (but lacking Gamma the phonon cannot be calculated)..

Any hint for this problem? Many thanks.

Hi

It would be good if you take the same value of the input variable "kptopt" for gamma and other q points i.e kptopt = 3
Though the abinit documentation states that the merging can happen successfully even on application of time reversal symmetry at gamma point i.e. using kptopt = 2, but in practice this doesn't happen.

Hope the merging of all DDBs will go fine then.

Good Luck,
Meha

Hi

Even thought I used the same kptopt for all the data sets, the problem does happen, and the DDB can not be merged. This does not always happen, i.e., for some systems it does, some it does not. Any hint, your any help is appreciated. Please request more information if needed.

Huan

Meha wrote:Hi

It would be good if you take the same value of the input variable "kptopt" for gamma and other q points i.e kptopt = 3
Though the abinit documentation states that the merging can happen successfully even on application of time reversal symmetry at gamma point i.e. using kptopt = 2, but in practice this doesn't happen.

Hope the merging of all DDBs will go fine then.

Good Luck,
Meha

Thanks a lot!! There sounds like a good way to do it. Actually, even without Gamma point, the rest points should be OK; and Gamma-phonon can be plotted separately, which in practice does not affect the final spectra.

Thank you very much again!