The ABINIT Users Mailing List ( CLOSED )

[marc.torrent@cea.fr]

Dear Bill,

In principle, the parallel version of PAW should work.
But because of the constant developpment of the PAW part of Abinit some 
errors can (unfortunately) appear.

If you want some help (if possible) you have to give us more details:
- Version of Abinit used ?
- Exact error message displayed ?
- Input file ?


You have to notice that PAW is still in beta-test (so you always have to use 
the last version of Abinit) and that only total energy is correctly computed 
(no forces or stresses).


Marc Torrent
CEA-Bruyeres-le-Chatel
France


>---------------------------------------------------------------------------
# From: sheltonwajr@ornl.gov
# To: forum@abinit.org
# Subject: PAW and parallel computations
# Date: Tue, 6 Apr 2004 22:16:34 +0200


I keep getting MPI errors when running PAW in parallel. Does PAW portion of
the code work correctly in parallel mode ? Note, the norm-conserving
pseudopotentials work fine in parallel mode.

Thanks
Bill Shelton