Job Results

Where do my results go?

After submitting your job to the DSH HPC Cluster, you can use the command qstat to view the status of all the jobs you have submitted. Once you can no longer see your job on the list, this means your job has completed. After a job is completed, you can see some technical statistics about it using the command qacct -j <job-ID>.

Output and error files

When writing your job script you can either tell it to start in the directory you submit it from (-cwd), from a particular directory (-wd <dir>), or from your home directory (the default). When your job runs, it will create files in this working directory for the job's output and errors:

File Name	Contents
myscript.sh	Your job script.
myscript.sh.o12345	Output from the job. (stdout)
myscript.sh.e12345	Errors, warnings, and other messages from the job that aren't mixed into the output. (stderr)

If you change the name of the job in the queue, using the -N option, your output and error files will use that as the filename stem instead. Most programs will also produce separate output files, in a way that is particular to that program. Often these will be in the same directory, but that depends on the program and how you ran it.