[Pw_forum] parallel run of phonon calculation (ph.x)

Paolo Giannozzi giannozz at nest.sns.it
Mon Apr 7 15:00:32 CEST 2008


Huiqun Zhou wrote:

> Is my understanding of q point parallelization wrong?

it is. You can have parallelization on plane waves,
parallelization on the k-points (those used in the
calculation of the self-consistent charge density)
but there are no other parallelization levels in the
phonon code. In particular, there is no parallelization
on either the node patterns or the q-points (those used
to calculate the phonon dispersions). It is relatively
easy to implement them and sooner or later it will be done,
since it is very useful for execution on the so-called
"grid" (large clusters of loosely coupled machines that
appear as a single machine to the user)

Paolo
-- 
Paolo Giannozzi, Democritos and University of Udine, Italy



More information about the users mailing list