On Jul 23, 2008, at 8:24 AM, Gabriele Fatigati wrote:
>You could always effect your own parallel IO (e.g., use MPI sends
and receives to coordinate parallel reads and writes), but >why?
It's already done in the MPI-IO implementation.
Just a moment: you're saying that i can do fwrite without any lock?
OpenMPI does this?
No. I'm saying that you can either effect your own locks with MPI
sends and receives (i.e., more-or-less equivalents of flock, because,
IIRC, flock on NFS is not reliable -- but I'm not an expert in this
area, so you might want to double check that), or you could just use
MPI-IO.
And, what is ROMIO? Where can i find any informations?
Sorry, ROMIO is the implementation of the MPI-2 IO chapter that Open
MPI uses. ROMIO is embedded in Open MPI and is built by default. So
you can use all the MPI-2 IO stuff if you're using Open MPI.
--
Jeff Squyres
Cisco Systems
