[Wien] parallel calc. without NFS

[Peter Blaha]

> What about files named *_1, *_2, *_3 etc. in the case directory? I assumed
> that they were generated by processes 1,2,3 etc. Is this correct? If so
> *_[1-4] will only exist on node 1, *_[5-8] only on node 2, etc...

Yes, *_1 files are generated by process 1,...

Nevertheless, all files are located in the "case" directory, except
*.vect* and *.help*, which may go into $SCRATCH.

*.vect* are written by   lapw1 lapw1_i.def     and used by   lapw2 lapw2_i.def,
and the latter also writes *help*.
Those are the "biggest" files and they can be considered as "temporary" files
during scf, thus they can go to $SCRATCH (where we assume that process 1 of
lapw1 AND 2 runs on the same machine (only valid in "clean" distributions, see
discussions some time ago).

*.energy*, *scf1_*, (written by lapw1) need to be present in the common case
directory.

lapw2 with "FERMI" reads those *.energy files and creates   *.weigh*, which
must be available to the respective    lapw2 lapw2_i.def jobs.

The latter create *.scf2_* and *.clmval*, which must also go to a common
case directory, since they are summed up by a single
sumpara sumpara.def i     job.

Hope, this helps
                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------