-
Notifications
You must be signed in to change notification settings - Fork 4
Open
Description
The idea is that for a large long-running ensemble of simulations, at the end you will begin to not have enough work to properly pack the nodes and use your compute time most efficiently. To help with this, give job runners priority and have certain nodes that stay more packed towards the end allowing the less packed nodes to finish their simulations and then relinquish the nodes thereby saving on compute time.
Reactions are currently unavailable
Metadata
Metadata
Assignees
Labels
No labels