Opened 8 years ago

Closed 8 years ago

#1115 closed error (fixed)

Problem with the build stage of the FCM tutorial

Reported by: s1251469 Owned by: ros
Component: FCM Tutorial Keywords:
Cc: Platform: HECToR
UM Version: 8.2


I am at the Running Jobs section of the FCM tutorial. I have completed the Setting Up section.

I have copied the job, made the changes, saved, processed, submitted and can see the job begin on Hector.

However, the build fails very early on. With the output errors available in:

I think it might be related to the compiler given an example error shown below. Some errors are not related to the compiler though.

USE Field_Types

ftn-1777 crayftn: ERROR UM_PARPARAMS, File = /home/n02/n02/s1251469/um/xiuha/umatmos/ppsrc/UM/control/mpp/um_parparams.f90, Line = 13, Column = 7

File "/home/n02/n02/s1251469/um/xiuha/umatmos/inc/FIELD_TYPES.mod" contains modules and/or submodules. The compiler being used is older than the compiler that created this file. The file was created with version 86 from release Unknown.

Do I need to add anything to my .profile on Hector?

Many thanks,

Change History (4)

comment:1 follow-up: Changed 8 years ago by ros

  • Owner changed from um_support to ros
  • Status changed from new to accepted

Hi Declan,

The problem you see is because yesterday when you submitted the job and it started to compile, you hadn't set up your .profile correctly and the latest version of the FORTRAN compiler was used. Since then you have corrected your .profile to set up the required environment for the UM which uses an older version of the compiler. The UMUI job by default is set to do incremental builds - i.e. after the first compilation it only compiles those files which have been modified since the last time. So you are getting an incompatibility with compiler versions.

Please go to the UMUI window FCM configuration —> FCM extract directories & Options and select Force full build. This will force a complete recompilation.


comment:2 in reply to: ↑ 1 Changed 8 years ago by s1251469

Thankyou Ros,

That problem is fixed. I assume that now it has built properly it makes sense to turn off the "force full build" option.

The build now runs properly for some time but I can't find the executable once the build has stopped. I believe it should be in a bin subdirectory along with the other subdirectories such as umatmos and umscripts. There appears to be no such subdirectory.

I took a look at the outputted files. There is now three when there was only one before which I think is good. I can't see any problem with them.

A colleague found that the compilation ran out of time so I tried in two ways to account for that but to no avail.

1) user information… > job submission … > qsub
job time limit to 3600

2) compilation and run options > … for atmosphere …
time limit for compilation to 10800

Can you see anything wrong?

comment:3 Changed 8 years ago by ros

Hi Declan,

The job has compiled successfully and you should find the exec in $DATADIR/um/xiuha/bin.

The xiuha000.xiuha.d13226.t130639.rcf.leave file contains the output from running the reconfiguration. Which also ran successfully. You will see near the top of the file "Job terminated normally" and you should have a new start dump in your $DATADIR/um/xiuha directory.

Both xiuha000.xiuha.d13226.t125800.leave and xiuha000.xiuha.d13226.t130639.leave indicate again that the model exec ran successfully. I haven't checked all the output but I can't see a file that says it ran out of time.

For the future:

To set the time for the compilation you change the time as you say in (2) above.
To set the time for the model run you do as in (1) above.

Could you also please change the permissions on your /work directory so that we can see your files and be able to help with any further problems.

chmod -r g+rX /work/n02/n02/s1251469


comment:4 Changed 8 years ago by ros

  • Resolution set to fixed
  • Status changed from accepted to closed

This query is now being closed.

Note: See TracTickets for help on using tickets.