The ABINIT Users Mailing List ( CLOSED )

[Josef Zwanziger <jzwanzig@gmail.com>]

Dear Colleagues,
I thought I understood the use of mrgddb but maybe not. I am computing the
phonon dispersion curves of a primitive hexagonal structure, using as input

ngkpt 8 8 5
shiftk 0.0 0.0 0.5

I computed the GS WFK with kptopt 1, the DDK 1WF with kptopt 2, the Gamma 
point
responses with kptopt 2, and the other q perturbations with kptopt 3. This
leads to nkpt 162 for the Gamma point and nkpt 320 for the others. Then, 
mrgddb
fails with the error message that the nkpts differ between the different
blocks. But this must be normal, mustn't it? I've computed numerous other
dispersion curves in this way and the gamma point always has fewer kpts than
the rest, because of the use of kptopt 2 rather than 3. Can anyone give me any
suggestions as to what is going wrong?

thanks,
Joe