An open API service indexing awesome lists of open source software.

https://github.com/aaronjackson/slurm-utils

:hammer_and_pick: Slurm utilities used on the CVL GPU cluster
https://github.com/aaronjackson/slurm-utils

Last synced: 7 months ago
JSON representation

:hammer_and_pick: Slurm utilities used on the CVL GPU cluster

Awesome Lists containing this project

README

          

* CVL GPU Cluster Slurm Utilities

This is a collection of scripts used by the Computer Vision Laboratory
at the University of Nottingham on our Slurm GPU Cluster.

** Utilities

*** clusterfree

This script displays GPU availability information, for example,
whether how many single GPUs, pairs, triplets or quads are
available. It also displays the QoS which are allowed to submit jobs
to those partitions.

#+BEGIN_SRC
$ clusterfree
+------------+---------+---------+--------------------------+------------------------+
| | # GPU | Avail | Pairs Triplets Quads | Allowed QoS |
+------------+---------+---------+--------------------------+------------------------+
| all | 49 | 21 | 10 3 3 | web |
| general | 31 | 11 | 5 1 1 | normal busybee short |
| turing | 8 | 8 | 4 2 2 | normal busybee short |
| undergrad | 2 | 2 | 1 0 0 | undergrad short |
+------------+---------+---------+--------------------------+------------------------+
#+END_SRC

*** smem

Finding out the MaxRSS of your Slurm jobs usually requires the use of
sstat with several arguments. This script enumerates the user's
running jobs and reports memory consumption for each one.

#+BEGIN_SRC
$ smem
User JobID Req Max Used
------------------------------------
pszaj 246317 954M 167M
pszaj 246318 954M 168M
pszaj 255305 30G 21G
pszaj 255415 12G 8.2G
#+END_SRC

** Contributions

If you have found a bug and would like to submit a fix, please do so
with a pull request or by emailing your git-formatted patch to
asj@cs.nott.ac.uk. Thanks!