Main menu

Navigation

error with create_newcase

13 posts / 0 new
Last post
xgao304@...
error with create_newcase

I am able to port the previous cesm version 1.2.2 on our machine. However, when I tried to port cesm2 based on the following link

http://esmci.github.io/cime/users_guide/porting-cime.html), I got some error messages with creating new case.

Here is the relevant configuration for our machine:

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

config_compilers.xml:

<compiler MACH="svante">
  <NETCDF_PATH> $(NETCDF)</NETCDF_PATH>
  <ADD_SLIBS> $(shell $(NETCDF_PATH)/bin/nc-config --flibs) -llapack -lblas</ADD_SLIBS>
  <MPI_LIB_NAME>openmpi</MPI_LIB_NAME>
  <MPI_PATH> $(INC_MPI)/..</MPI_PATH>
</compiler>

config_machines.xml:

  <machine MACH="svante">
    <DESC>MIT Svante-Login, batch system is SLURM</DESC>
    <OS>LINUX</OS>
    <COMPILERS>pgi</COMPILERS>
    <MPILIBS>openmpi</MPILIBS>

...........

  </machine>

config_batch.xml

  <!-- svante is SLURM -->
  <batch_system MACH="svante" type="slurm">
    <batch_submit>sbatch</batch_submit>
    <directives>
      <directive default="/bin/tcsh" > -S {{ shell }}  </directive>
    </directives>
  </batch_system>

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

Here is the steps I have taken:

1. when I run "xmllint", the results indicate that config_machines.xml validates;

2. When I run "scripts_regression_tests.py", I have the following error message:

results:  Ran 129 tests in 100.785s    FAILED (failures=32, errors=64, skipped=9)

Some error messages are like "Could not find a matching MPI for attributes: {'mpilib': 'openmpi', 'threaded': False,'compiler': 'pgi'}"

3. When I run "create_newcase" as follows, I am getting the simiar error as I run "scripts_regression_tests.py"

./create_newcase --case ../../../cases/test2000 --compset I2000Clm50SpGs --res hcru_hcru --machine svante --run-unsupported

Compset longname is 2000_DATM%GSWP3v1_CLM50%SP_SICE_SOCN_MOSART_SGLC_SWAV
Compset specification file is /net/fs05/d1/xgao/cesm/cesm2/cime/../components/clm//cime_config/config_compsets.xml
Compset forcing is 1972-2004
ATM component is Data driven ATM GSWP3v1 data set
LND component is clm5.0:Satellite phenology:
ICE component is Stub ice component
OCN component is Stub ocn component
ROF component is MOSART: MOdel for Scale Adaptive River Transport
GLC component is Stub glacier (land ice) component
WAV component is Stub wave component
ESP component is
Pes specification file is /net/fs05/d1/xgao/cesm/cesm2/cime/../components/clm//cime_config/config_pes.xml
Could not find machine match for 'c062.svante.mit.edu' or 'c062.svante.mit.edu'
Machine is svante
Pes setting: grid is a%360x720cru_l%360x720cru_oi%null_r%r05_g%null_w%null_m%null
Pes setting: compset is 2000_DATM%GSWP3v1_CLM50%SP_SICE_SOCN_MOSART_SGLC_SWAV
Pes setting: tasks is {'NTASKS_ATM': -1, 'NTASKS_ICE': -1, 'NTASKS_CPL': -1, 'NTASKS_LND': -1, 'NTASKS_WAV': -1, 'NTASKS_ROF': -1, 'NTASKS_OCN': -1, 'NTASKS_GLC': -1}
Pes setting: threads is {'NTHRDS_ICE': 1, 'NTHRDS_ATM': 1, 'NTHRDS_ROF': 1, 'NTHRDS_LND': 1, 'NTHRDS_WAV': 1, 'NTHRDS_OCN': 1, 'NTHRDS_CPL': 1, 'NTHRDS_GLC': 1}
Pes setting: rootpe is {'ROOTPE_OCN': 0, 'ROOTPE_LND': 0, 'ROOTPE_ATM': 0, 'ROOTPE_ICE': 0, 'ROOTPE_WAV': 0, 'ROOTPE_CPL': 0, 'ROOTPE_ROF': 0, 'ROOTPE_GLC': 0}
Pes setting: pstrid is {}
Pes other settings: {}
Pes comments: none
Compset is: 2000_DATM%GSWP3v1_CLM50%SP_SICE_SOCN_MOSART_SGLC_SWAV
Grid is: a%360x720cru_l%360x720cru_oi%null_r%r05_g%null_w%null_m%null
Components in compset are: ['datm', 'clm', 'sice', 'socn', 'mosart', 'sglc', 'swav', 'sesp', 'drv', 'dart']
No project info available
No charge_account info available, using value from PROJECT
No project info available
cesm model version found: cesm2.0.0
starting mpirun_nodes
BAR BAR BAR
('here is an attribute: ', 'mpilib')
('here is an attribute: ', 'openmpi')
('here is an attribute: ', 'threaded')
('here is an attribute: ', False)
('here is an attribute: ', 'compiler')
('here is an attribute: ', 'pgi')
ERROR: Could not find a matching MPI for attributes: {'mpilib': 'openmpi', 'threaded': False, 'compiler': 'pgi'}

--------

Any information is appreciated about how to solve the problem.

Thanks.
jedwards

Could you run the create_newcase command with the --debug flag and send the resulting create_newcase.log file?   

xgao304@...

Please see the attached file for log file. The command will hang in there as follows:

> /net/fs05/d1/xgao/cesm/cesm2/cime/scripts/lib/CIME/utils.py(126)expect()
-> try:
(Pdb)

 

Thanks,

Xiang

jedwards

No attachement included, use CNTRL-C to exit when it hangs and attach the file. 

xgao304@...

I did attach the file, but not sure why it did not show up. I did use ctl-C to end the command, but just want to provide that information to you.

jedwards

Funny your attachment didn't contain any errors.   What python version are you using?

xgao304@...

I think it is python 2.7.3

jedwards

It's easy to confirm:

python --version

 

xgao304@...

python 2.7.13

 

jedwards

Could you create a git fork of cime and a branch with your changes?   I haven't figured out the issue and I want to try to reproduce the problem.    

xgao304@...

I am not very familiar with git, so not sure how to do that. Literally I only changed three files in cime/config/cesm/machines:

config_batch.xml, config_compilers.xml, config_machines.xml

I attached these three files. Please let me know if that helps.

 

 

jedwards

Okay the problem is that your machine definition is incomplete.

You need to have the <mpirun> section.

You need to have a <module_system> even if type='none'

You need to correct the case in <BATCH_SYSTEM> it should be 'slurm' not 'SLURM'

 

xgao304@...

Thanks. I will add the relevant machine definition and then give it a try.

Log in or register to post comments

Who's new

  • zweina@...
  • yuan.liang@...
  • lian.xue@...
  • 353482168@...
  • 76414461@...