Skip to content
This repository has been archived by the owner on May 19, 2023. It is now read-only.

hephaisto/psm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

19 Commits
 
 
 
 
 
 
 
 

Repository files navigation

psm

Python SLURM Manager

Installing

  • make sure psm.py is in your $PATH
  • copy psm_config.py.template to psm_config.py in your project directory
  • edit psm_config.py (especially your default partition)
  • create ~/.psm/outputs (or whatever path you specified in your psm_config.py)

Running

  • run psm in your project directory
  • If you use the default configuration, PSM will save job definitions in your project directory and your joblist globally in your user directory.

User Interface

Left Pane

The left Pane shows a list of all jobs.

Buttons

  • enable automatic refresh
  • delete current item
  • clear all finished/cancelled jobs
  • cancel current job

Middle Pane

The middle pane shows a list of all defined jobs.

Buttons

  • execute current job definition
  • create/edit/delete job (opens job definition window)

Right pane

The right pane shows the output of the currently (left pane) selected job.

Job definition window

This window has four fields (from top to bottom):

Job name

This name will be shown in lists in PSM and as title in slurm. Names must be unique.

Commands

List of commands to execute. Use python format-style placeholders for parameters, e.g.

sleep {0}
sleep {1}

Per-instance parameters

List of parameters given to each instance. These parameters are formatted into the command template. Separate multiple instances with linebreaks and multiple parameters with ,, e.g.

1,1
1,10
10,10

Use

RUN=SCRIPTNAME

to use a script name that outputs parameters to stdout.

Global parameters

Global parameters are passed to SLURM to request e.g. additional memory or a certain number of CPUs. Separate multiple options with linebreaks and key/value pairs with =, e.g.

mem-per-cpu=5000
num-tasks=2

For a list of possible options, refer to the SLURM manual.

License

This code is released under the MIT License (see LICENSE file for more information).

About

Python SLURM Manager

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages