On the supercomputer system, user’s program is executed as a job.
Job script file
You first need to prepare a job script file, in which you specify the amount of required resources (e.g. the number of nodes, maximum elapsed time. etc.). Then, you submit your job script file for executing your program.
Outline of a job script
An example of a job script is as follows. A job script file usually has the ".sh" extension.
#!/bin/sh #------ pjsub option --------# #PJM -L rscgrp=ea #PJM -L node=1 #PJM -L elapse=1:00:00 #PJM -g group1 #PJM -j #------- Program execution -------# ./a.out
#!/bin/sh: shell path
#PJM ***: job options
./out: execution of a program (i.e. a.out)
In the case of using the
module command in a job script written in sh excepting bash,
you need to specify
Basic job options are as follows.
|-g <groupname>||specify a group|
|-L rsgrp=<name>||specify a resource group (i.e. job queue)|
|-L node=<num>||specify the number of reuiqred nodes|
|--mpi proc=<num>||specify the total number of MPI processes (in the case of executing a MPI program)|
|-j||specify to output the standard error output to the standard output|
|-N <jobname>||specify the name of a job|
|-L elapse=<hh:mm:ss>||specify the maximum elapsed time of a job|
Maximum elapsed time
If an appropriate maximum elapsed time is specified,
Examples of job script
Please visit here.
Job operation commands
Major commands for job operation are as follows.
- Job submission:
- Job status check:
- Job cancel:
The working directory when executing a job is that the
pjsub command was issued.
Results of job execution
Standard output and standard error output are saved in files.
If the output file(s) was not specified in a job script file, the following files are generated in the directory where the
pjsub command was issued. ((XXXXX is the job ID.)
- (job name).oXXXXX：standard output
- (job name).eXXXXX：standard error output