Am 14.04.2014 um 19:27 schrieb Reuti: > Am 14.04.2014 um 17:59 schrieb Michael Coffman: > >> On Mon, Apr 14, 2014 at 9:52 AM, Reuti <[email protected]> wrote: >> Am 14.04.2014 um 17:48 schrieb Michael Coffman: >> >>> With 0.5 of course it shouldn't hit this value - or is 10 not the real core >>> count? >>> >>> In this case, 10 is not the real core count. >>> >>> Is there something I can set on my production grid to get it to print >>> similar scheduling info?
Without specifying a job number you can use: $ qstat -j But it will of course not show the reason for the not starting job, as it blocks itself. But I'm not sure under what circumstances it happens. It doesn't seem to be always the case. -- Reuti >>> >>> Also - how is np_load_avg calculated? >> >> It's outlined in load_parameters.asc too. >> >> Very good... >> >> How do I have the scheduler print the following if jobs can't run? >> >> np_load_avg=1.865000 (= 0.000 + 0.50 * 3.730000 with nproc=1) >= 1.75 >> >> >> >> np_load_avg : The same as load_avg but divided by the number of >> processors. This value allows to compare the load of >> single and multi headed hosts. >> >> I am a bit slow, sorry :( So the description above doesn't seem to map >> into the formula to me. What is the 3.73 number? > > Well, ok - not for this output, as it's adjusted. Sorry. I thought of > np_load_avg in general. > > The 3.73 is the number of cores (starting at 4 here), decreasing over the > specified time for load_adjustment_decay_time of 0:7:30 to zero. > > -- Reuti > _______________________________________________ > users mailing list > [email protected] > https://gridengine.org/mailman/listinfo/users _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
