Standardized Time Limits for Queues
Date Received:
February 16, 2007

Centers: I'm only familiar at this point with two, ERDC and NAVO, but it should apply to all centers.
Comments: The debug queue has a time limit of 1 hour on machine Sapphire at ERDC, but only 0.5 hour on machine Babbage at NAVO. These values should be standardized. I cannot comment on the time limits for other queues due to lack of familiarity. Obviously, the processor limit for each platform is a machine-dependent value. It would be nice if there were a minimum, however. For example, the debug queue everywhere should support a minimum of 32 processors.

BC Team Feedback
Reply Date:
April 17, 2007

Your request for standardized time limits for queues falls within the scope of the DoD HPCMP User Advocacy Group (UAG). The UAG has had a number of discussions in the past on common queue names and common queue limits (based on both "time" and "number of CPUs"). As the UAG meeting minutes indicate, the UAG has not been in favor of efforts to standardize queue time limits in the past. However, if you wish to pursue this topic further, we kindly ask you to contact the UAG at hpc-uag-members@hpcmo.hpc.mil.

Related to queue time limits, the BC Team has already received a request for easy access to queue time limits. The BC Team has plans to incorporate this capability either in an existing or a future BC policy soon. For more details on this subject, please click on the link "Ease of Use" on the BC Web site's User Input/Input Received page.