[Pw_forum] Trying to find reasonable parallelization parameters

Lorenzo Paulatto paulatto at sissa.it
Fri Jan 23 09:45:39 CET 2009


On Thu, 22 Jan 2009 19:37:00 +0100, Axel Kohlmeyer <akohlmey at cmm.chem.upenn.edu> wrote:
> very simple. -npools can be maximally the actual number of k-points that
> you are using, and that should be exploited to the maximum.

True, but the maximum may not be the number of kpoints; using the parallel davidson algorithm (default) the hamiltonian matrix will be distributed and diagonalized on the larger perfect square number of processor *inside* each pool. If you are low on memory you may need to spread the hamiltonian over several processors in order to fit in. I.e. for a 3000 electrons job I'm trying these days, I have to use at least 100 procs per pool (and set diago_david_ndim=2) to stay (well) under ~2GB.

cheers

-- 
Lorenzo Paulatto
SISSA  &  DEMOCRITOS (Trieste)
phone: +39 040 3787 511
skype: paulatz
www:   http://people.sissa.it/~paulatto/

     *** save italian brains ***
 http://saveitalianbrains.wordpress.com/



More information about the users mailing list