[mpich-discuss] mpich2 in a cluster (nfs?)

[Zhu Sha Zhang]

Hello, i'm trying to use mpich2 in my sick cluster. Sick because the
ancient admin are installed umbumbumtum in there. Sucks.

I'm already configured nfs and ssh. Now, mpdboot -n 3 work perfectly.

If i use mpiexec -n 3 /home/cluster/program and this programa only write
in a nfs exported file system maybe i have only one process shared in
all 3 nodes. Really, the process was more fast. 1/3 of time.

But, my "clients" using a program that write in a non exported nfs
filesystem where in each node the files generated by
"/home/cluster/program"  are the same and writed in a /scratch/user PATH.

In this case, the time of execution are more slow (16 minutes to on
machine <-> 19 minutes using 3 machines), then i thought that the same
process are running in the 3 machines but not shared.

The problem is the 3 filesystem not shared? Cos, the first program i've
tested not ported to parallelization and worked well.

Of course i'll try write in nfs shared partition, but the problem is in
some cases 100Gb can be generated.

Some hint to me?

Thanks.