basic use of the cluster

[cjayb]

How do I:

• submit an arbitrary command-line (bash) command to be processed on the cluster
  • maxfilter
  • freesurfer
  • others?
• find out what is going on with my job
  • is it running/finished
  • why is it taking so long (is the queue full/busy)?
• avoid being yelled at by the admins because I have
  • "submitted a threaded/MPI/MKL/multicore process" (and what does that even mean?!?)
  • submitted to the wrong queue (how many queues are there anyway, and why?!?)

[cjayb]

See also work-in-progress regarding submitting via command-line https://github.com/meeg-cfin/configurations/issues/7

[MadsJensen]

Should also be mentioned how to set up a bashrc.

[cjayb]

Is this too well hidden? http://wiki.pet.auh.dk/wiki/Scientific_computing_basics#Configuration_and_customisation

[MadsJensen]

No, but there is very little info on e.g. commands you wrote for isis (e.g. set_mindlabproj) etc.

[cjayb]

http://wiki.pet.auh.dk/wiki/Utility_functions_for_CFIN_servers :) Perhaps my "conversational" style on the landing page (http://wiki.pet.auh.dk/wiki/Category:MEG) means no-one will find these...

[MadsJensen]

ups, I had not seen that :/

[MadsJensen]

It should link to How to use the Grid Engine Linux cluster.

[sarathykousik]

To keep persisting on my own view, we must probably collapse both to a single page?

[cjayb]

Persist all you like :) You mean you'd like to see "Scientific computing" and "Utility functions" on the same page? I'm OK with that, rather too many pages than too few! And good landing/redirection/summary pages, not to mention with descriptive names that make them searchable (looking at you, MediaWiki...)

[sarathykousik]

Hmm...yeah... topics are searchable as well..