Listing 6: Warewulf – Part 4

Listing 6 for Warewulf Part 4

Listing 6: Torque Job Script

[laytonjb@test1 TEST]$ more pbs-test_001
1  #!/bin/bash
2  ###
3  ### Sample script for running MPI example for computing PI (Fortran 90 code)
4  ###
5  ### Jeff Layton
6  ### 8/5/2012
7
8  ### Set the job name
9  #PBS -N mpi_pi_fortran90
10
11 ### Run in the queue named “batch”
12 #PBS -q batch
13
14 ### Specify the number of cpus for your job.  This example will allocate 4 cores
15 ### using 3 processors on 1 node.
16 #PBS -l nodes=1:ppn=3
17
18 ### Tell PBS the anticipated run-time for your job, where walltime=HH:MM:SS
19 #PBS -l walltime=0:10:00
20
21 ### Load needed modules here
22 . /etc/profile.d/modules.sh
23 module load compilers/open64/5.0
24 module load mpi/mpich2/1.5b1-open64-5.0
25
26 ### Switch to the working directory; by default TORQUE launches processes
27 ### from your home directory.
28 cd $PBS_O_WORKDIR
29 echo Working directory is $PBS_O_WORKDIR
30
31 # Calculate the number of processors allocated to this run.
32 NPROCS=`wc -l < $PBS_NODEFILE`
33
34 # Calculate the number of nodes allocated.
35 NNODES=`uniq $PBS_NODEFILE | wc -l`
36
37 ### Display the job context
38 echo “Running on host `hostname` “
39 echo “Start Time is `date` “
40 echo “Directory is `pwd` “
41 echo “Using ${NPROCS} processors across ${NNODES} nodes “
42
43 mpirun -np 3 ./mpi_pi < file1 > output.mpi_pi
44
45 echo “End time is `date` “