facebookincubator/submitit

Issues

• Failure of submitit 1.5.1 with python 3.12 because of missing pkg_resources

  #1765 opened 20 days ago by GianlucaFicarelli
  0

• UnicodeDecodeError fails the job

  #1712 opened 2 years ago by phtu-cs
  2

• Can I use torchrun with submitit?

  #1764 opened 3 months ago by vasudev-sharma
  0

• Turn off Signal Handling

  #1760 opened 5 months ago by lukasbm
  2

• Using `RsyncSnapshot` with a editable package install

  #1763 opened 3 months ago by jc-audet
  2

• be tolerating about sacct error?

  #1720 opened a year ago by min-xu-ai
  2

• Submitit jobs die with no error on cluster with SLURM 19.05

  #1762 opened 4 months ago by mihdalal
  0

• Unexpected behavior of memory specification between `AutoExecutor` and `SlurmExecutor`

  #1761 opened 4 months ago by mshvartsman
  0

• Too many sacct requests for batched tasks

  #1759 opened 5 months ago by Fadelis98
  0

• Failed to launch: Invalid wckey specification

  #1758 opened 6 months ago by rskwesterman
  0

• When 'submitit' meet 'mpirun', there will be a very strange BUG.

  #1757 opened 6 months ago by yinkaaiwu
  0

• Improving performance with NVidia GPU affinity?

  #1756 opened 6 months ago by giorgos117
  0

• Consider supporting slurm rest api

  #1719 opened 2 years ago by zeronewb
  2

• No user code logging output is shown in logs

  #1721 opened a year ago by fleimgruber
  2

• Documentation of `executor.update_parameters` arguments

  #1735 opened a year ago by JoeZiminski
  1

• SLURM Job keeps running after Successful Job Completon (Hydra Submitit Plugin)

  #1731 opened a year ago by subho406
  2

• Enabling sbatch file re-use.

  #1739 opened a year ago by alexnwang
  2

• Submitit map_array is using 2 GPUs but only requested 1

  #1753 opened 7 months ago by eigenvivek
  1

• Does setting `folder` in `AutoExecutor` interfere with sattach?

  #1748 opened 8 months ago by fleimgruber
  1

• timeout_min=0 results in pending jobs when a Slurm partition timelimit is set

  #1742 opened 9 months ago by ddangu525
  0

• Support Slurm Heterogeneous Job

  #1741 opened 9 months ago by sunshine-syz
  2

• Submit Over SSH?

  #1711 opened 2 years ago by JRJacoby
  4

• Conda version out of date

  #1737 opened a year ago by Ubadub
  1

• AttributeError , AutoExecutor attribute not recognised by submitit

  #1738 opened a year ago by willianck
  1

• array_parallelism on local machine

  #1736 opened a year ago by sparisi
  1

• Submitit with SLURM sub-scheduling

  #1734 opened a year ago by giorgos117
  2

• duplicate tasks when using `SlurmExecutor.map_array`

  #1727 opened a year ago by eringrant
  3

• Requeueing on timeouts when launching jobs with CommandFunction

  #1732 opened a year ago by Niccolo-Ajroldi
  0

• SLURM Jobs keep running after successful job completion.

  #1730 opened a year ago by subho406
  0

• Add custom options to sbatch command in SLURM

  #1728 opened a year ago by nilskober
  4

• Submitit with sbatch

  #1726 opened a year ago by pfrwilson
  6

• Unwanted behavior after a slurm job time limit

  #1708 opened a year ago by ofir1080
  1

• Printing in Signal Handlers May Be Unsafe

  #1714 opened 2 years ago by Queuecumber
  1

• How to specify GPUs when executing locally?

  #1696 opened a year ago by j0ma
  5

• Submitit puts all tasks on a single GPU

  #1704 opened a year ago by Bai-YT
  3

• Should we submit job on login node?

  #1722 opened 2 years ago by surajmenon72
  1

• InfoWatch might get previous jobid info after slurm restart

  #1716 opened 2 years ago by Liangtaiwan
  1

• array_parallelism for LocalExecutor

  #1718 opened a year ago by se-ok
  2

• submitit.core.utils.FailedJobError: sbatch: error: Parameter --gres=gpu:1 no longer acceptable, please switch to --gpus=1

  #1710 opened a year ago by RoyAmoyal
  1

• Can submitit manage chain dependencies?

  #1723 opened a year ago by eserie
  1

• Compute Canada

  #1724 opened a year ago by kaijieshi7
  1

• NodeList Declaration

  #1692 opened a year ago by Bontempogianpaolo1
  2

• Remove `#SBATCH --nodes=1`

  #1725 opened a year ago by sgbaird
  3

• How to load the original code point when preempted and rescheduled if the code is changed before rescheduling?

  #1717 opened 2 years ago by dahyun-kang
  1

• Switching from USR1 Breaks Pytorch Lightning

  #1709 opened 2 years ago by Queuecumber
  4

• Recover jobs after kernel dies

  #1707 opened 2 years ago by SamuelGabriel
  1

• submit job array to multiple partitions

  #1698 opened 2 years ago by MinkyuHa
  0

• Latest versions' tags not on Github

  #1690 opened 2 years ago by tmct
  2

• Task does not wait for GPU memory resources

  #1689 opened 2 years ago by chirico85
  1

• filedescriptor out of range in select()

  #1687 opened 2 years ago by timlacroix
  1