[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Condor-users] machine ranking issues



Hello,

I am trying to use MPI to read a file and to output with all processors.  I
am using a resource sharing framework entitled Condor to pass the executable
and data file to each machine.  I am using three machines with Condor 6.8
with MPICH 1.2.4 on Fedora 5.

My questions do not deal with MPI as much as how Condor would pass a file
and read from the directory.  Hence, my problem is with file reading and
output difficulties.

I have the following job submission file.

Universe = parallel
executable = jobfile.sh
arguments = 3 test
output  = test$(NODE).out
error   = test$(NODE).error
Log     = test.log
machine_count = 3
should_transfer_files = yes
when_to_transfer_output = on_exit
transfer_input_files = test, randomOut5.txt
queue

My questions are as follows.

1) If in my MPI code, the rank for the exe is 1, does this correspond to
test1.out on that machine?

2) Why does my machine ranked 0 in my code output to the data file on the
machine ranked 1 and 2?  Shouldn't machine 0's output have only messages for
the rank of 0, machine 1 in the file for rank of 1 and so on?  For example,
it would appear my that output from machine 1 is being output on the file
test2.out.

I would like to thank everyone for the assistance thus far and reading my
messages.

Sincerely,

Christopher Jon Jursa
Geoinformatics Laboratory
School of Information Sciences
University of Pittsburgh
web: http://gis.sis.pitt.edu
email: cjursa@xxxxxxxxxxxx
phone: 412-624-8858