Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Test dynamic AWS system #470

Open
pearce8 opened this issue Dec 3, 2024 · 0 comments
Open

Test dynamic AWS system #470

pearce8 opened this issue Dec 3, 2024 · 0 comments
Assignees

Comments

@pearce8
Copy link
Collaborator

pearce8 commented Dec 3, 2024

@wdhawkins The AWS system definition has been updated (systems/aws-pcluster), please test and let us know if you think it does what it should. Right now, this produces experiment scripts like:

#SBATCH --mpi=pmix --export=ALL,FI_EFA_USE_DEVICE_RDMA=1,FI_PROVIDER="efa",OMPI_MCA_mtl_base_verbose=100
#SBATCH -n 8
#SBATCH -N 1
...
srun --mpi=pmix --export=ALL,FI_EFA_USE_DEVICE_RDMA=1,FI_PROVIDER="efa",OMPI_MCA_mtl_base_verbose=100 -n 8 -N 1 saxpy -n 512 >> ...

(i.e. these are generated as sbatch opts and command opts) - this is decided by the allocation modifier (specifically for the SLURM scheduler). I think it's worth checking if these should be batch opts

  • If so, this can be rewritten to generate a separate batch opt for --mpi and --export
  • If not, I can change the allocation modifier

I don't think this decision need stop this PR, but the generated experiment scripts will have weird sbatch directives until that is decided.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants