Bureaucrats, cc_docs_admin, cc_staff
2,879
edits
Frequently Asked Questions: Difference between revisions
Jump to navigation
Jump to search
no edit summary
(Marked this version for translation) |
No edit summary |
||
| Line 137: | Line 137: | ||
== How accurate is START_TIME in <tt>squeue</tt> output? == <!--T:33--> | == How accurate is START_TIME in <tt>squeue</tt> output? == <!--T:33--> | ||
Start times shown by <tt>squeue</tt> are | Start times shown by <tt>squeue</tt> depend on rapidly-changing conditions, and are therefore not very useful. | ||
<!--T:34--> | <!--T:34--> | ||
[[Running jobs|Slurm]] | [[Running jobs|Slurm]] computes START_TIME for high-priority pending jobs. These expected start times are computed from currently-available information: | ||
* | * What resources will be freed by running jobs that complete; and | ||
* resources needed by other, higher-priority jobs waiting to run | * what resources will be needed by other, higher-priority jobs waiting to run. | ||
<!--T:35--> | <!--T:35--> | ||
Slurm invalidates these future plans: | |||
* jobs end early, changing which resources become available; | * if jobs end early, changing which resources become available; and | ||
* prioritization | * if prioritization changes, due to submission of higher-priority jobs or cancellation of queued jobs for example. | ||
<!--T:36--> | <!--T:36--> | ||
On Compute Canada general purpose clusters, new jobs are submitted about every five seconds, and 30-50% of jobs end early, | On Compute Canada general purpose clusters, new jobs are submitted about every five seconds, and 30-50% of jobs end early, | ||
so | so Slurm often discards and recomputes its future plans. | ||
<!--T:37--> | <!--T:37--> | ||