[Pw_forum] How to estimate the memory requirement of ph.x?

Paolo Giannozzi giannozz at democritos.it
Fri Dec 10 12:33:20 CET 2010


lfhuang wrote:

> it is understandable that ph.x requires more memory than pw.x.
>  But how  to estimate it to be qualitatively precise?

sum the size of large arrays. I do not know any other way. If the
code runs, a call to "memstat" returns the size of dynamically
allocated memory (on operating systems that support it). Not sure
how much of the total memory is reported, though. Apparently the
question "how much memory does a process use" has become difficult
to answer in modern operating systems.

> How can a 3GB processor breaks down when doing a ph.x job, when the 
> requirement of it pw.x job is just 500MB?

the phonon code uses more arrays and distributes less of them across
processors.

> By the way, is it better to assign the N_pr to be m**2 (m=2,3,4...) than 
> other values (e.g. 6, 10, 11, 14...) if I want to divide the R-&G-space 
> in parallelization?

I think it depends upon the specific communication hardware and
MPI implementation. Not sure though.

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



More information about the users mailing list