mpich on scyld problems

henken henken at seas.upenn.edu
Wed Oct 31 10:18:21 PST 2001


Hello--
	I am tryign to test our scyld cluster with the cpi program. The
machinefile, "mfile" contains:
3:2
4:2

The command I am trying to run is:
mpirun -map 3:4:3:4 -v -machinefile mfile cpi

The output I get is:
running /home/henken/e2k/jobs/cpi on 4 LINUX ch_p4 processors
Created /home/henken/e2k/jobs/PI26817
Process 0 on eio5
pi is approximately 3.1416009869231254, Error is 0.0000083333333323
wall clock time = 0.026838

also, if I add a -t to the mpirun command, I get:
running /home/henken/e2k/jobs/cpi on 4 LINUX ch_p4 processors
Procgroup file:
3 1 /home/henken/e2k/jobs/cpi
4 2 /home/henken/e2k/jobs/cpi
rcp: /home/henken/e2k/jobs/PI26716: No such file or directory
/usr/bin/bpsh -n 3 /home/henken/e2k/jobs/cpi -p4pg
/home/henken/e2k/jobs/PI26716 -p4wd /home/henken/e2k/jobs

Does anyone know why the program is failing to run multiple times on the
SMP nodes? Or why it doesnt run on the bproc node4(eio6)?

Nicholas Henke




More information about the Beowulf mailing list