Opened 12 years ago

Closed 12 years ago

#145 closed help (fixed)

Hector run producing spurious output on a few lines

Reported by: fuensd@… Owned by: jeff
Component: UM Model Keywords:
Cc: Platform:
UM Version:

Description

Hi,

Two of us here are trying to get a global UM run at 40 km resolution to work on Hector.

My job is xddka under user id fuensd. The job has been copied from a working simulation on hpcx (xdaha) with a few changes to the compilation and model modifications listed in UMUI.

The start dump was copied across from hpcx.

The output we get for the u field at t=0 in the pp1 file is attached. It has 7 latitude lines where the values look out of place. View data shows that it is just one line of spurious data at each latitude.

The printed out for the job is at :
/home/n02/n02/fuensd/um/umui_out/xddka000.xddka.d08177.t100127.leave

The job runs for all timesteps but by the next output at t+3 the values on some of the lines has become massive (+-1057). Other data looks fine.

These odd lines only seem to affect the u and v fields. The w field output on the pp3 file wont display in xconv - it gives a segmentation fault. Other fields in pp6 and pp9 files display fine for all time steps. The output lateral bc file also looks fine.

Nick Dixon

Change History (5)

comment:1 Changed 12 years ago by umdoc

My e-mail is fuensd@…

Couldn't attach the image as its too big

Nick

comment:2 Changed 12 years ago by jeff

  • Cc fuensd@… added
  • Owner changed from um_support to jeff
  • Status changed from new to assigned

Hi

I don't have permission to read your files, can you give me access, this command will fix it

chmod -R g+rX $HOME

also it might be useful to see the output in your work directory

chmod -R g+rX $WORKDIR

Jeff.

comment:3 Changed 12 years ago by umdoc

Jeff

I've changed the permissions on my files

Nick

comment:4 Changed 12 years ago by jeff

  • Cc fuensd@… removed
  • Reporter changed from umdoc to fuensd@…

Hi Nick

I've found out why you get spurious lines on your output data. In umui window

Sub-Model Independent → Post Processing → Initialization and processing of mean & standard PP files

you have used "Packing profile" 1 in the PP Files part, these packing profiles don't seem to work correctly on hector, you should use packing profile 0 instead. I've done a test of your run with this option and the output looked normal.

Jeff.

comment:5 Changed 12 years ago by jeff

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