On Jul 9, 2008, at 15:00 , ??? wrote: > Is the phonon code (ph.x) really implemented in parallel version?
it is. Parallelization on plane waves and on k-points is implemented. Other kinds of parallelization aren't. > Furthermore, the speed of computation shows no difference when > I used 32 CPUs from that when I only 8 CPUs. the efficiency of parallelization for each phonon calculation is similar to that for the self-consistent solution. The optimal number of cpus depend upon a number of factors: cell size and type, kind of parallelization, speed of communications in your machine. This is more or less explained in the documentation Paolo --- Paolo Giannozzi, Dept of Physics, University of Udine via delle Scienze 208, 33100 Udine, Italy Phone +39-0432-558216, fax +39-0432-558222
