Dear Maite (CC. Steven)
It seems that already today some sites are starting to fix their MPI
problems :)
We also got a few reactions wondering about this sudden urge to fix MPI site
problems now.
It would certainly help if the the ROCs receive an explanatory e-mail about
the MPI Task Force mission, containing also the link with to the official
documentation to MPI Support in EGEE, meaning this one:
https://twiki.cern.ch/twiki/bin/view/EGEE/MpiTools
that each ROC should distribute it to their sites.
Many people is concerned because they have followed some documentation which
is also online coming from SEE Grid, and particular to certain cluster in
Budapest.
There are reasons today for being optimistic, because people are fixing the
issues, and mpi-start continues to work without any problem in the CREAM CE
(see
http://indico.ifca.es/indico/getFile.py/access?contribId=10&sessionId=1&
amp;resId=1&materialId=slides&confId=249
)
However, in the timelife of EGEE we can probably only fix the current sites,
and arrange properly the documentation. Any other thing like new features of
the middleware will have to waitt for future developments. See here for
status of mpi-start:
http://indico.ifca.es/indico/getFile.py/access?contribId=2&sessionId=0&a
mp;resId=0&materialId=slides&confId=249
cheers,
Isabel
More information about the MPI knowledge DB:
http://wiki.ifca.es/e-ciencia/index.php/MPI_Errors