[Pw_forum] sufficient grid for phonon dispersion

[Miguel Martínez Canales]

Hi

Ezad Shojaee wrote:
> hi
> sorry if it is a silly question
> if we used a grid for calculating  phonon dispersion and we find that it 
> is'nt sufficient, what should we do? defining another grid ( with more 
> points ) and performing for all of it again?!

My probably naive way of doing it would be scf run -> nscf run -> phonon 
calculation at the desired q. Of course, with all the parameters being the 
same with the exception of the k-point sampling. After that, I would 
substitute the file containing the omegas and the eigenvectors with the new 
one.

As you see, pretty naïve. But faster than calculating every single point.