Failed CLM4 (with modified code) single-point run at lnd_comp_mct.F90

[meir02ster@gmail_com]

Dear all,

I was trying to run CLM4 with some code modification added by myself. I was able to build it but failed to run. Based on the log files, what I can tell so far is that the model crashes at "call aerdepini()" in the file lnd_comp_mct.F90, A further check shows that it stops in the subroutine aerdepini in aerdepMod.F90 while it was executing this line "call shr_sys_flush(iulog)". shr_sys_flush routine is said to flush the I/O buffer (which I do not understand well). All these clues are not within the region where I made the modifications.

My modifications include the following files: clmtype.F90, clm_atmlnd.F90, clmtypeInitMod.F90, clm_driver.F90, clm_comp.F90, mkarbinithMod.F90 initsurfAlbMod.F90.

I also tried to build it with debug option and it does not help to give me more information written in the log files for diagnosis. I will keep on trying, meanwhile I am wondering could someone in this forum offer some advice on this issue.

Thank you for any comment.

Sincerely,
Rui

 

[meir02ster@gmail_com]

Here is the ending part of the log files:

Atm log:
%%%%%%%%%%%%%%%%%%%%%%
(shr_dmodel_readLBUB) reading file: /home2/meir02/ccsm4_0/inputdata/atm/datm7/atm_forcing.datm7.Qian.T62.c080727/TmpPrsHumWnd3Hrly/clmforc.Qian.c2006.T62.TPQW.2003-01.nc 1
(datm_comp_run) max values = 298.6308514760269
0.000000000000000 0.000000000000000
(datm_comp_run) atm: model date 20000101 0s
(datm_comp_init) datm_comp_init done
(datm_comp_run) atm: model date 20000101 0s
(datm_comp_init) datm_comp_init done
%%%%%%%%%%%%%%%%%%%%%%%%

ccsm log
%%%%%%%%%%%%%%%%%%%%%%%%
1 pes participating in computation for CLM

-----------------------------------

NODE# NAME
( 0) water
Attempting to read annual vegetation data .....
[mpiexec@water] ONE OF THE PROCESSES TERMINATED BADLY: CLEANING UP
APPLICATION TERMINATED WITH THE EXIT STRING: Floating point exception (signal 8)
%%%%%%%%%%%%%%%%%%%%%%

lnd log
%%%%%%%%%%%%%%%%%%%%%%
(aerdepini) Read time axis from /home2/meir02/ccsm4_0/inputdata/lnd/clm2/snicardata/aerosoldep_monthly_2000_mean_1.9x2.5_c090421.nc
(GETFIL): attempting to find local file
aerosoldep_monthly_2000_mean_1.9x2.5_c090421.nc
(GETFIL): using
/home2/meir02/ccsm4_0/inputdata/lnd/clm2/snicardata/aerosoldep_monthly_2000_mean_1.9x2.5_c090421.nc
(aerdepini) number of time samples in file = 12
(aerdepini) nt = 12: assume old format without a CF-1.0 time axis
(aerdepini) first time series cal date = 10115.500
(aerdepini) last time series cal date = 11215.500
(aerdepini) first time axis elapsed days since 0000-01-01 = 380.500
(aerdepini) last time axis elapsed days since 0000-01-01 = 714.500
%%%%%%%%%%%%%%%%%%%%%%%%%%%

 

[meir02ster@gmail_com]

cpl log
%%%%%%%%%%%%%%%%%
(seq_frac_check) [lnd init] lfrin min/max = 1.00000000000000000 1.00000000000000000
(seq_frac_check) [atm init] afrac min/max = 1.00000000000000000 1.00000000000000000
(seq_frac_check) [atm init] lfrac min/max = 1.00000000000000000 1.00000000000000000
(seq_frac_check) [atm init] ofrac min/max = 0.00000000000000000 0.00000000000000000
(seq_frac_check) [atm init] ifrac min/max = 0.00000000000000000 0.00000000000000000
(seq_frac_check) [atm init] lfrin min/max = 1.00000000000000000 1.00000000000000000
(seq_frac_check) [atm init] sum min/max = 1.00000000000000000 1.00000000000000000
(seq_frac_check) [atm init] sum ncnt/maxerr = 0 0.00000000000000000
(seq_mct_drv) : Setting fractions

(seq_mct_drv) : Model initialization complete
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

 

[slevis]

Dear Rui,

I assume that you ran the model successfully before making changes to the code. If not, I recommend doing so to confirm that the model is not crashing for some other reason. If you did run successfully without your changes, then the model is crashing because of your changes, and your next task is to discover the bug in your code.

Sincerely,
Sam Levis

 

[meir02ster@gmail_com]

Dear Sam,

Yes, it can run in default and I have also figured out why it crashes with my modifed code. It's because I have made some inconsistency about the type of the variables I added. But the log file is very implicit. I found it is always easy to diagnose the problem if it fails while building since the log file is more explicit on the "real" problem. Thank you anyway.

Best regards,
Rui