High Performance Computing Service: Difference between revisions
No edit summary |
No edit summary |
||
Line 1: | Line 1: | ||
== Disclaimer == | |||
This page is a work in progress! | This page is a work in progress! | ||
Line 4: | Line 5: | ||
Kirstie x | Kirstie x | ||
== Introduction == | |||
If you have many jobs that are very similar and, importantly, '''independent''' then you can use the High Performance Computing Service [[http://www.hpc.cam.ac.uk/]]. The website is helpful, but here are a few commands that I think supplement their FAQs: | If you have many jobs that are very similar and, importantly, '''independent''' then you can use the High Performance Computing Service [[http://www.hpc.cam.ac.uk/]]. The website is helpful, but here are a few commands that I think supplement their FAQs: | ||
== If you want to see all the jobs you've submitted == | == If you want to see all the jobs you've submitted == | ||
As you may guess, qstat gives you some statistics on the queues. To look for just '''your''' jobs type: | As you may guess, qstat gives you some statistics on the queues. To look for just '''your''' jobs type: | ||
Revision as of 11:45, 18 May 2013
Disclaimer
This page is a work in progress!
I'm just using it as a repository for helpful hints that I've found useful! Kirstie x
Introduction
If you have many jobs that are very similar and, importantly, independent then you can use the High Performance Computing Service [[1]]. The website is helpful, but here are a few commands that I think supplement their FAQs:
If you want to see all the jobs you've submitted
As you may guess, qstat gives you some statistics on the queues. To look for just your jobs type:
qstat -u <user_id> eg: qstat -u kw401
If you want to delete all the jobs you've submitted
To just get the job IDs you need to use qselect and pipe these to qdel (which deletes the job, obviously) using xargs.
qselect -u <user_id> | xargs qdel eg: qselect -u kw401 | xargs qdel
Type man xargs if you want to know more about this terribly useful command which pipes information from standard input into the next command.