(pronounced dub-flow), is a suite of scripts which facilitates using a single cluster as a home base, written by S. Kirklin. Many aspects of its operation are shamelessly stolen from Bryce's ClusterCommander utility, but with expanded options, functionality and documentation.
- Default behavior is to find the cluster with the shortest queue and submit jobs to that cluster
- Any time while the job is running, the std out can be checked from the home base
- Once a job is completed, the completed results are transfered to the home base
- The inputs used to generate the job are backed up in an 'input' directory within the submitted directory.
- Must have passwordless ssh set up between your home base and all clusters you want to use
- Must be willing to tell me what functionality you want, so I can keep making it better and more useful
How to get it?
Just ask me and I will get you started!
I am working on an "installer" to get everything auto-configured, but it isn't quite ready yet.