Skip to content

Add ray support#8

Open
JamesScott-swe wants to merge 7 commits intomainfrom
jscott/add_ray_support
Open

Add ray support#8
JamesScott-swe wants to merge 7 commits intomainfrom
jscott/add_ray_support

Conversation

@JamesScott-swe
Copy link
Collaborator

Adds option for running with ray on slurm. Ray entrypoint script is stored as large string rather than actual .sh file so it can be dynamically added to tar archives without altering the repo.

Tested with toy python -c "print('hello')" command, happy to run something more substantial if you have a suggestion

Copy link
Contributor

@gjoliver gjoliver left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

for testing, can you grab this script from my andromeda home dir?
/mnt/main0/home/jgong/evolutionaryscale/scratch/test_ray.py
thanks.

help="If set, will tag the job with the specified tag",
)
@click.option(
"--ray",
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

maybe we can brain storm a bit together.
I was thinking we can introduce this capability as a separate command like:

slurm_run ray <python .../job.py>

and hopefully we can move most of the ray specific logics into a ray.py file.
I just think we have a good opportunity to clean up the slurm_run implementation, is this possible?

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

Successfully merging this pull request may close these issues.

2 participants