Remote job submission for treeThreader

The file tt_boinc is a script for remotely submitting batches of treeThreader jobs to a BOINC project that has been set up to accept them. It is designed to be used in creating a web portal for submitting treeThreader jobs.

tt_boinc must be edited to supply the URL of the project (the PROJECT constant defined on line 8).

The script provides two commands:

tt_boinc submit sequence_file authenticator

sequence_file: a sequence file, which will be run against a large group of template files.

authenticator: the authenticator of the user's account on the project.

The script's output is either an error message or

Batch ID: N

where N identifies the batch of jobs.

tt_boinc get_output batch_id authenticator

batch_id: a batch ID, returned by a previous submit.

authenticator: as above

Output: an error message or

URL: xxxx

where xxxx is the URL from which the jobs' output files can be downloaded.

Preparing the project

The project must have

  • the treeThreader application installed
  • a collection of targz'd template files staged for download, and
  • a file tree_threader_template_files in its root directory, containing a list of these files.

Scientist wanting to submit jobs must create an account on the project, and must have their account enabled for job submission by project admins. They must also get their authenticators from the admins.

Additional APIs =

In addition to these APIs, the APIs described here will be needed to create the job-submission portal. In particular, the boinc_query_batch(), boinc_abort_batch(), and boinc_retire_batch() APIs may be useful.

Also, the project's web site contains a "batch control" page that can be used by scientists to monitor and control their job batches. Hence the portal need not supply these functions.

Last modified 6 years ago Last modified on 04/17/12 15:18:04