Mentions légales du service

Skip to content

Introduce slurm-mpirun scheduler

SCHOULER Marc requested to merge slurm-openmpi into develop

This MR was motivated by a recent issue encountered by T. Terraz on the MUSE cluster of Meso@LR. The cluster based on slurm does not support heterogeneous job submission with srun. To circumvent this issue, the idea is to use mpirun and its MPMD syntax but still in an indirect fashion (i.e. submission with sbatch scripts).

The adopted solution is to implement a new scheduler slurm-openmpi which inherits from the slurm one and which submits jobs by mixing openmpi and slurm schedulers.

To keep things as flexible as possible without doing any assumption about the resource usage, the configuration must comply with the following:

  • #SBATCH level options correspond to those passed through scheduler_arg_client,
  • mpirun level options correspond to those passed through scheduler_command_options.

WARNING: full flexibility means that the total number of tasks wont be derived automatically. Making sure group_size, #SBATCH --ntasks and mpirun -n are consistent is the user's responsibility.

NOTE: using slurm-openmpi fixes the virtual cluster unit-group limitation!

Edited by SCHOULER Marc

Merge request reports