One too which I used a lot when running Torque was the pbstop command. It is an insightful way to see cores on nodes and what is running on them. It is easy to see which users are gaming the system by using partial nodes and allows for scheduling changes based on observations.

Before we begin down the path of rolling our own version of slurmtop, it would be wonderful if anyone has already this type of command already.

Bill

Reply via email to