[Pw_forum] Problems with example05

Paolo Giannozzi giannozz at nest.sns.it
Fri Oct 12 23:16:41 CEST 2007


On Oct 12, 2007, at 21:19 , Jšrg Sa§mannshausen wrote:

> However, in parallel mode some examples are stalling

not all examples work in parallel execution (they should, but your case
demonstrate that they don't). The reason is that only the most important
codes: scf, phonon, MD calculations, are parallelized. Smaller codes,
auxiliary codes are mostly not parallelized, for obvious reasons. From
http://www.quantum-espresso.org/wiki/index.php/ 
Running_on_parallel_machines:
"Please note that all postprocessing codes not reading data Þles  
produced
  by pw.x Ñ that is, average.x, voronoy.x, dos.x Ñ the plotting codes  
plotrho.x,
  plotband.x, and all executables in pwtools/, should be executed on  
just one
  processor. Unpredictable results may follow if those codes are run  
on more
  than one processor."

In reality nothing unpredictable should happen, because all nonparallel
codes run serially on one processor while all others do nothing. It is
however likely that there are cases, notably in the pp.x code, where
something goes wrong. It is typically something really stupid, like a  
call
to a routine that performs some hidden communications and waits for an
answer from other processors that are instead idling. All such cases
should have been fixed in the CVS version, I think.

In any case: if the problem is limited to a few nonparallelized
post-processing codes stalling in parallel execution, you can
safely ignore it

Paolo
---
Paolo Giannozzi, Dept of Physics, University of Udine
via delle Scienze 208, 33100 Udine, Italy
Phone +39-0432-558216, fax +39-0432-558222






More information about the users mailing list