Main menu

Navigation

[Porting] some validation tests fail at ./create_test

11 posts / 0 new
Last post
l.vankampenhout@...
[Porting] some validation tests fail at ./create_test

Hi all,

First time poster. As an experiment, I'm currently porting CESM v1.2 to our tiny local Bull cluster called "Hopf". After a few days of tinkering, I've got the build and job submission working. I now landed at page 54 of the user guide, "Step 3: Port validation" which still gives me a hard time.

What's the case, given the list of suggested validation tests:
  ERS_D.f19_g16.X
  ERS_D.T31_g37.A
  ERS_D.f19_g16.B1850CN
  ERI.ne30_g16.X
  ERI.T31_g37.A
  ERI.f19_g16.B1850CN
  ERS.ne30_ne30.F
  ERS.f19_g16.I
  ERS.T62_g16.C
  ERS.T62_g16.D
  ERT.ne30_g16.B1850CN

Only some of them appear to work correctly (e.g. no 1, 4) whereas others cannot even create themselves (no 2, 3).

Error for number 2:
ERROR: 2000_DATM%NYF_DLND%NULL_DICE%SSMI_DOCN%DOM_DROF%NYF_SGLC_SWAV
 is not supported for
 a%T31_l%T31_oi%gx3v7_r%r05_m%gx3v7_g%null_w%null

Error for number 3:
rtm.buildnml.csh could not find restart file for branch or hybrid start
ERROR: rtm.buildnml.csh failed

 

Full output attached. Please let me know how I could proceed. Am I missing something?

Cheers,

Leo van Kampenhout (IMAU / Universiteit Utrecht)

Attachment: 
jedwards

Hi Leo,


I'm not positive about this but I think that you are using the 1.2.0 release and we've fixed many (hopefully most) of these testing problems with the 1.2.1 update.    Can you download the 1.2.1 source tree and try again?

CESM Software Engineer

l.vankampenhout@...

Hi J,

thanks for the quick reply. I've checked the CESM version number it is 1.2.1. However, the SVN revision was slightly out of date: rev 57180. Just now I've updated to the latest 57595. Unfortunately, problems persist...

jedwards

I'm not sure then.  What case is it that you want to run, perhaps we should focus on that?

CESM Software Engineer

l.vankampenhout@...

OK. At the moment there is not particular case I want to run on this machine. Main reason behind this excercise is to gain experience with the CESM technicalities and if possible, make it available for future use. Regarding that last bit I would say that at a minimum, all the validation tests should run successfully.

There is another machine in Amsterdam where we do plan to do some real work on in the near future. It is also a Bull Linux system like this one, so I'm curious to know whether the same problem will manifestate there. Will keep you posted.

l.vankampenhout@...

Hi, discovered a possible cause of the problems with the f19_g16 grid (test 3, 6 and 8)

./create_test -compiler intel -mach cartesius -testid t02 -testname ERS_D.f19_g16.B1850CN
(...)
rtm.buildnml.csh could not find restart file for branch or hybrid start
ERROR: rtm.buildnml.csh failed

Looks like the restart files are not found (correct) but not automatically downloaded (faulty). Could this be related to the fact that the *.input_data_list files under /Buildconf/ are irregular? Namely they read:

clm.input_data_list:finidat = b40.1850.track1.2deg.003.clm2.r.0501-01-01-00000.nc
cam.input_data_list:ncdata = b40.1850.track1.2deg.003.cam.i.0501-01-01-00000.nc

which are exceptions to the (apparent) rule of prefixing the filename by $DIN_LOC_ROOT and the directory tree.

Could you confirm that the absence of the absolute path in theses files may be a problem?

Best regards,
Leo

 

l.vankampenhout@...

In test ERI.f19_g16.B1850CM --> Buildconf/cam.input_data_list

There is a reference to:

b40.1850.track1.2deg.003.cam.i.0501-01-01-00000.nc

Shouldn't this be:

b40.1850.track1.2deg.003.cam2.i.0501-01-01-00000.nc

This is at least the filename I see listed in SVN:

svn ls https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata/ccsm4_init/b40.1...

The full entry then becomes:

ncdata = <replace with $DIN_LOC_ROOT>/ccsm4_init/b40.1850.track1.2deg.003/0501-01-01/b40.1850.track1.2deg.003.cam2.i.0501-01-01-00000.nc

jedwards

We no longer use the 2 in the cam file name so it should be just cam.   For some of the older files in the repo that still have the cam2 name, there is supposed to be a link created in the

run directory.   

CESM Software Engineer

somanath999@...
Hi I am also facing same kind of error.please suggest something I am using f19_g16 and B1850CN rtm.buildnml.csh could not find restart file for branch or hybrid start ERROR: rtm.buildnml.csh failed ERROR: /home/somanath.amd/cesm1_2_1/scripts/test4/preview_namelists failed: 25344
mcus2rail@...

Hi Leo,

I am getting the same error for test 2:

"ERROR: 2000_DATM%NYF_DLND%NULL_DICE%SSMI_DOCN%DOM_DROF%NYF_SGLC_SWAV
 is not supported for
 a%T31_l%T31_oi%gx3v7_r%r05_m%gx3v7_g%null_w%null"

Have you had luck getting this to run?

 

Best,

Marcus Trail (Emory University)

jedwards

Test 2 should be ERS_D.T31_g37_rx1.A

CESM Software Engineer

Log in or register to post comments

Who's new

  • afroberts@...
  • richard.allard@...
  • tracy@...
  • 498749303@...
  • lijundesio@...