The freecycle partition is made up of lab owned nodes whose free CPU cycles are generously being shared by the owners. Priority is always given to jobs submitted by the owning Lab. This means that if those nodes were to get low on resources while a freecycle job was running, new jobs submitted by the owning labs would pre-empt and cancel enough running jobs to free up resources for the Lab job. For this reason, the runtime is restricted to 24 hours on the freecycle partition.
Intended use case
You have a short job you want to run but you notice that the job queues on the common (or your lab) partition. In order to run the job faster you could submit it to the freecycle partition with the understanding that your job could be cancelled.
The freecycle partition
The freecycle partition is made up of lab owned nodes whose free CPU cycles are generously being shared by the owners. Priority is always given to jobs submitted by the owning Lab. This means that if those nodes were to get low on resources while a freecycle job was running, new jobs submitted by the owning labs would pre-empt and cancel enough running jobs to free up resources for the Lab job. For this reason, the runtime is restricted to 24 hours on the freecycle partition.
Intended use case
You have a short job you want to run but you notice that the job queues on the common (or your lab) partition. In order to run the job faster you could submit it to the freecycle partition with the understanding that your job could be cancelled.