Urgency Policy
-- deadline -- wait time -- job size
Example:
-- The urgency policy is to give higher priority to larger jobs --- I have just submitted two test jobs, identical, both to R815.q, one of 192 slots, the other of 128 slots: qstat -pri job-ID prior nurg npprior ntckts ppri name user state submit/start at queue slots ja-task-ID ---------------------------------------------------------------------------------------------------------------------------------------------------- 77917 0.10500 1.00000 0.50000 0.50000 0 r815-tp-19 simonh qw 03/02/2012 15:01:04 192 77916 0.07149 0.66492 0.50000 0.50000 0 r815-tp-12 simonh qw 03/02/2012 15:01:00 128 -- You can see the result --- the priority, from urgency, is bigger for the bigger job. -- Now here is the interesting bit --- I submit a bigger job, 256 slots: qstat -pri job-ID prior nurg npprior ntckts ppri name user state submit/start at queue slots ja-task-ID ---------------------------------------------------------------------------------------------------------------------------------------------------- 77919 0.10500 1.00000 0.50000 0.50000 0 r815-tp-25 simonh qw 03/02/2012 15:05:01 256 77917 0.07990 0.74902 0.50000 0.50000 0 r815-tp-19 simonh qw 03/02/2012 15:01:04 192 77916 0.05480 0.49804 0.50000 0.50000 0 r815-tp-12 simonh qw 03/02/2012 15:01:00 128 It looks like the urgency priority contrib for jobs is normalised against the biggest, which makes sense I guess.