Opened 7 years ago

Closed 7 years ago

#1070 closed help (fixed)

running HadCEM at Bristol

Reported by: ggjhtw Owned by: jeff
Component: UM Model Keywords:
Cc: Platform: Other
UM Version: 4.5

Description

Hi Jeff

My name is Jonny Williams and I am a postdoc at Bristol University working with Paul Valdes.

I am currently trying to get the high resolution ocean model HadCEM running (my latest job on the PUMA UMUI is TDLLL) but I am getting an error at run time (after successful compilation and reconfiguration) of…

*

Model completed with the following :

Error Code : 3392
Message : INIT_A2O: Coupling field not enabled - taux atmos

*

i.e. basically my job is not recognising the STASH code 3392, (X-COMP OF MEAN SEA SURF STRESS N/M2) even though this diagnostic is 'available' in STASH. I suppose it'd be good to define this variable in the 'initialisation of user prognostics' window since it is defined in my user STASHmaster file but there is no option to do so, even though it appears in the STASHmaster file?!

I understand that you have some experience with running HadCEM and so I thought that you would be a good person to chat to?

Do let me know if you require any further info from me and all the best.

Many thanks

Jonny

Change History (2)

comment:1 Changed 7 years ago by ggjhtw

Hi there

Since I sent the original ticket (above) I have managed to fix the issue above with some informal help from Robin Smith (editing the run's STASHC file on PUMA).

My job is now running in the atmosphere GCM as far as the first coupling period when it crashes out with a segmentation fault of myseterious origin! As far as I can tell the ocean model is not running at all and the failure is therefore likely in the initialisation of the ocean physics/stash.

To try and get past this I am using some compile overrides (including "-g" and "-traceback" in our Bristol FORTRAN compiler) to try and give me the routine and line numbers where the run is crashing but this is not working. I am simply getting "unknown" for the routine and line numbers where the segmentation fault is occurring.

If you have any further suggestions for how to get past this problem then I'#d be very grateful.

Many thanks!

Jonny Williams

comment:2 Changed 7 years ago by jeff

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