-
Notifications
You must be signed in to change notification settings - Fork 10
Issues in Python parallel slurm example #58
Description
There is a broken link in the Python README section header, it's missing the /parallel/ portion of the path.
Also, the text in the paragraphs of that section and the parallel README switch between use of njobs, n_jobs and n_jobs. The variable in the Python script is n_jobs, but it could also be worth clarifying if the "jobs" here are "tasks" or "threads" as we refer to them in the CCR documentation.
Finally, both READMEs mention that the n_jobs variable is currently set to 8 and can be changed in line 23 of the script. That line of the serial script is currently a comment and there doesn't seem to be an example slurm script for the parallel implementation. Either a new script should be added or the wording changed in the READMEs to make it clear how a user would run the fibonacci_joblib.py file with an arbitrary number of jobs, tasks, or threads.