% qsub -l nodes=4 -I qsub: waiting for job 8854.node1 to start qsub: job 8854.node1 ready node3:~> mpirun -np 4 -machinefile $PBS_NODEFILE /home/psmith/mpich/cpi Process 2 on node11 Process 0 on node3 Process 3 on node12 Process 1 on node3 pi is approximately 3.1416009869231245, Error is 0.0000083333333314 wall clock time = 0.001397 node3:~> logout qsub: job 8854.node1 completed
This will also let you run your jobs on compute nodes and make it possible to use debuggers, or other processes that don't work so well in batch.