You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A user may not know until runtime which files need to be transmitted back to the manager. For example, that decision may depend on some function evaluation, such as when to make a workflow snapshot, or which function call optimized some metric. In the case of the topEFT, the design simplifies if we mark all of the accumulation task results as temps, and then we simply fetch the file for the last accumulation.
These retrievals currently can be done by constructing a task, e.g.:
# generating the temp fileproducer=Task("cmd > output.xyz")
my_tmp_output=declare_temp()
producer.add_output(my_tmp_output, "output.xyz")
m.submit(producer)
t=m.wait(...)
...
# fetching the temp filefetcher=Task("ln -s output.xyz output.to_fetch")
my_fetched_output=declare_file"my_final_output")
# using the yet to merge pr for fixed locations, as we don't want this file to be moved twice just for the transferfetcher.add_input(my_tmp_output, "output.xyz", fixed_location=True)
fetcher.add_output(my_fetched_output, "output.to_fetch")
fetcher.set_category("fetch_files")
fetcher.set_tag("my_final_output")
m.submit(fetcher)
t=m.wait(...)
iftandt.category=="fetch_files"andt.tag=="my_final_output":
print("the file I wanted is now at the manager")
Thus, the API call to fetch files may simply construct a task as above.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
-
A user may not know until runtime which files need to be transmitted back to the manager. For example, that decision may depend on some function evaluation, such as when to make a workflow snapshot, or which function call optimized some metric. In the case of the topEFT, the design simplifies if we mark all of the accumulation task results as temps, and then we simply fetch the file for the last accumulation.
These retrievals currently can be done by constructing a task, e.g.:
Thus, the API call to fetch files may simply construct a task as above.
Beta Was this translation helpful? Give feedback.
All reactions