[slurm-users] SLURM heterogeneous jobs, a little help needed plz

[Frava]

Hi all,

I'm struggling to get an heterogeneous job to run...
The SLURM version installed on the cluster is 17.11.12

Here are the SBATCH file parameters of the job :
================================================================================
#!/bin/bash
#SBATCH --threads-per-core=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=1
#SBATCH --mem-per-cpu=8G
#
#SBATCH packjob
#
#SBATCH --threads-per-core=1
#SBATCH --cpus-per-task=4
#SBATCH --mem-per-cpu=2G
#SBATCH --nodes=2
#SBATCH --ntasks-per-node=2
#SBATCH --gres=gpu:2
#SBATCH --gres-flags=enforce-binding
================================================================================

Now the SRUN commands that I tried in the SBATCH file:

1) srun --mpi=pmix myapp
=> The app gets allocated only 1 MPI rank

2) srun --mpi=pmix --pack-group=0 --ntasks=1 : --pack-group=1 --ntasks=4
myapp
=> srun: fatal: Job steps that span multiple components of a heterogeneous
job are not currently supported

3) srun --mpi=pmix --pack-group=0 --ntasks=1 myapp : --pack-group=1
--ntasks=4 myapp
=> srun: fatal: Job steps that span multiple components of a heterogeneous
job are not currently supported

So my question is: How do we run an heterogeneous job ?

Thanks for your tips,
Rafael N.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20190319/b64587fd/attachment-0001.html>