[Pw_forum] phonons of bcc iron bulk in parallel

[Andrea Floris]

Dear Espresso community,

I am running a phonon calculation for iron bulk (bcc structure),
using nspin=2, i.e. a collinear spin polarized calculation.

I am running the version 3.2 *with applied 3.2.1 and 3.2.2. patches*.

If I run in serial, I can reproduce the results
obtained by Dal Corso and de Gironcoli [PRB 62, 273 (2000)].
But if I run in parallel, with the same input except for the number of
processors,
the code runs but the phonons, e.g. at H point, are a disaster (i.e. -940
cm-1,
nonsense, they should be around +295 cm-1).
I attach the relavent inputs/outputs.

It is important to note that if I run an UNpolarized calculation, the
serial and parallel
runs agree very well. This doesn't seem a problem of unlucky combination
of compiler/mpi/machine. Anyhow, I am using:

espresso 3.2 with 3.2.1 and 3.2.2. patches
compiler:  pgi 6.1
mpi     :  /opt/pgi/mpich2/bin/mpif90
machine :  opteron amachine dual or four core


Does anyone have a clue of what is happening?
I would appreciate very much any suggestions,
ciao

Andrea

---------------
 Dr. Andrea Floris
 Institut f"ur Theoretische Physik
 Freie Universit"at Berlin
 Arnimallee 14
 D-14195 Berlin
 Germany
 Tel +49-30-838-53029
 Fax +49-30-838-55258
 e-mail: afloris at physik.fu-berlin.de
-------------- next part --------------
A non-text attachment was scrubbed...
Name: phonon_fe.tgz
Type: application/x-gtar
Size: 12938 bytes
Desc: 
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20070614/a6c33ae9/attachment.gtar>