Opened 7 years ago

Closed 7 years ago

#1221 closed help (fixed)

job on archer runs too slow

Reported by: till Owned by: um_support
Component: UM Model Keywords:
Cc: Platform: ARCHER
UM Version: 6.1

Description

Hi there,
trying to run HiGEM on archer, the problem I have got now is that it runs too slow. The HiGEM jobs usually take about 3 hrs on Hector and they have nominal wallclock time of 6 hrs. But now my HiGEM job stopped because it took longer that 6hrs, while I was expecting it to run faster than on Hector. Could you please check? The .leave file is
/home/n02/n02/till/um/umui_out/xgvwt000.xgvwt.d14049.t090415.leave
Thank you very much!

Change History (5)

comment:1 Changed 7 years ago by grenville

Till

Please try running a short test on fewer processors - Jeff was running with 8x15.

Grenville

comment:2 Changed 7 years ago by till

Ok, I'm trying to run with 8x15 processors now, but the job is stuck in the queue:

Not Running: Host set host=archer_2901 has too few free resources

I'll wait overnight to see what happens.

comment:3 Changed 7 years ago by till

OK this seems to work now, thank you.

Next question is why it shows up in the queue with 24 hrs requested time, while I specified 6hrs (21600 sec) in the UMUI? It's not a big deal, but it would be nice to know …

Thank you!
Till

comment:4 Changed 7 years ago by ros

Hi Till,

I think this is because your job has automatic resubmission switched on but doesn't specify a QSUB time in this window for the CRUNs (Sub-model indept → Job submission and resources Click Next to get the Auto-resubmission window). Hence in the umuisubmit_run QSUB script we get walltime= and ARCHER then defaults this to 24hours.

Cheers,
Ros.

comment:5 Changed 7 years ago by ros

  • Resolution set to fixed
  • Status changed from new to closed
Note: See TracTickets for help on using tickets.