Welcome, Guest
Username: Password: Remember me

Here we can describe more what should be posted here

TOPIC: Trying to reduce reserved wall time for gl_bd jobs

Trying to reduce reserved wall time for gl_bd jobs 9 months 1 week ago #2507

  • Oskar Landgren
  • Oskar Landgren's Avatar
  • Fresh Boarder
  • Using HCLIM for Arctic climate projections
  • Posts: 6
I'm trying to reduce the waiting time in the queue on ECMWF CCA by reducing the wall time requested for the gl_bd jobs. (The waiting time is often 1-3 hours.)
The default wall time requested for this job seems to be 24 hours, but it rarely uses more than a minute. I therefore tried to add
$job_list{$task}{'WALLTIME'}         = $submit_type.'-l walltime=0:30:00';
for the gl_bd job in Env_submit, but when I run qstat on CCA the jobs are now requesting 24:30:00:
                                                            Req'd  Req'd   Elap
Job ID          Username Queue    Jobname    SessID NDS TSK Memory Time  S Time
--------------- -------- -------- ---------- ------ --- --- ------ ----- - -----
8205294.ccapar  fa0o     nf       gl_bd         --    9   8   16gb 24:30 Q   --

I guess this means there is some more abstraction involved here and it's not as straightforward as I first assumed. So, how does the actual wall time request work? Is is possible to decrease the request in order to reduce the waiting time?
(I see that the log files also print Historic runtime average for each job type. I assume that this is just for our convenience and not used by the queueing system, but please correct me if I am wrong!)

Trying to reduce reserved wall time for gl_bd jobs 8 months 4 weeks ago #2518

  • Ulf Andrae
  • Ulf Andrae's Avatar
  • Administrator
  • Posts: 308
  • Thank you received: 35

There is no way to control the wall clock limit on cca:


It's all about how much resources you need. The default allocations in terms on memory and nodes are possibly a bit too gready for your setup?

Time to create page: 0.074 seconds