[Pw_forum] Parallel efficiency problem

Paolo Giannozzi giannozz at democritos.it
Tue Jul 29 15:52:51 CEST 2008


lihm wrote:

> However, I found that the walltime is too long for parallel calculations:
> 
>      PWSCF        :  1m 4.86s CPU time,    32m53.67s wall time 

unless you have really slow communications between CPUs,
the only reason that can lead to such a "speedown" is heavy
I/O via the network (e.g. NFS-mounted disks)

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


More information about the Pw_forum mailing list