Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

Validation of CESM2 recently ported to new unsupported machine

NiCosentino

NiCosentino

Nicolas Cosentino
New Member
We have recently ported CESM2.1.1 to the NLHPC cluster in Chile, and we would like to validate fully coupled simulations in our port with simulations ran in supported machines. There is a series of scientifically validated configurations in https://csegweb.cgd.ucar.edu/experiments/public/. However, these are two long to be ran on our cluster (i.e., >100 yr) . We may run no more than 50 years for a fully coupled B1850 compset. Is it possible to run some of these scientifically validated simulations for a shorter period and compare with output for that time of span? In other words, are there history files for these simulations that integrate, say, 50 years? Any other recommendations for where to find output for B1850 simulations in supported machines that are <50 yr?
 
J

jedwards

CSEG and Liaisons
Staff member
First welcome to the CESM community. Please follow the port and testing procedures found in 6. Porting and validating CIME on a new platform — CIME ufs_release_v1.0 documentation
In lieu of reproducing a long fully coupled run we recommend running the ensemble verification tests as documented there. Once completed we invite you to submit your port as a pull request (PR) to the cime git site: ESMCI/cime (maint-5.6 branch) this will allow you to get updates to cime and the CESM component models without having to redo your port.
 
NiCosentino

NiCosentino

Nicolas Cosentino
New Member
jedwards said:
First welcome to the CESM community. Please follow the port and testing procedures found in 6. Porting and validating CIME on a new platform — CIME ufs_release_v1.0 documentation
In lieu of reproducing a long fully coupled run we recommend running the ensemble verification tests as documented there. Once completed we invite you to submit your port as a pull request (PR) to the cime git site: ESMCI/cime (maint-5.6 branch) this will allow you to get updates to cime and the CESM component models without having to redo your port.
Click to expand...

Thanks for the welcome! I am trying to perform the basic functionality test described in section 6.5 of the documentation you pointed me to. I perform the prealpha suite of tests, but all of them fail at the "sharedlib_build" phase. There are different error messages for different tests. The first test prints the following errors:

1) Building test for ERS in directory /home/cr2/projects/cesm/scratch/ERS_Ld3.f45_g37_rx1.A.leftraru_intel.20200605_151136_mbsko2
File not found: factorfn = "/home/cr2/projects//cesm/inputdata/atm/datm7/CORE2/COREv2.correction_factors.T62.121007.nc", will attempt to download in check_input_data phase.

2) ERROR: /home/cr2/projects/cesm/model/cime/src/build_scripts/buildlib.pio FAILED, cat /home/cr2/projects/cesm/scratch/ERS_Ld3.f45_g37_rx1.A.leftraru_intel.20200605_151136_mbsko2/bld/pio.bldlog.200605-151201.

I believe the first error should not make this ERS test fail (will try to download missing data later). I have looked at the log for this error (attached), but I don't know what I am looking for.
 

Attachments

  • log.txt
    27.4 KB · Views: 9
J

jedwards

CSEG and Liaisons
Staff member
The file not found is just a warning and not an error, the file will be downloaded in a later step.

The log file that you attached is a word document and not a txt file, it took me a bit to figure out how to open it.
Here is the error:
cmake: command not found

The tool cmake is a required prequisite for CESM, you can find it here. CMake
 
NiCosentino

NiCosentino

Nicolas Cosentino
New Member
Hello. We have installed CESM2.1.3 in the NLHPC supercomputer in Chile. I am now running scripts_regression_tests.py to check the correctness of our installtion. Of 361 tests, 11 resulted in failure. One of the error messages that appears a lot is:

ERRPUT: ERROR: Error gathering provenance information from manage_externals.

The log file suggests to fix this by either:

(1) Find and fix the problem: From /home/cr2/projects/cesm2.1.3, try to get this command to work:
manage_externals/checkout_externals --status --verbose --no-logging
(2) If you don't need provenance information, rebuild with --skip-provenance-check.

I have re-run the tests adding the option --skip-provenance-check to ./case.build in scripts_regression_tests.py. However, I do not understand what this error implies, and am in fear that I may be masking an important problem by doing what I am doing. I attach the log file (.pdf) for one of the tests in scripts_regression_tests.py that shows this error as an example. I would appreciate advise on this.
 

Attachments

  • log_scripts_regression_test.pdf
    49.1 KB · Views: 7
J

jedwards

CSEG and Liaisons
Staff member
You can run with --skip-provenance-check but it acts as if svn isn't working which may indicate that you don't have a complete copy of the code.
 
NiCosentino

NiCosentino

Nicolas Cosentino
New Member
Hi, we are still trying to port CESM2.1.3 to Chile's national supercomputing cluster: Leftraru/NLHPC. We are performing the prealpha tests, and most of them fail in relation to downloading required input data files. We tried to access de server manually to see if the inputdata is available (ftp://gridanon.cgd.ucar.edu:2811/cesm/inputdata/), but appearently that URL is down. We have tried using all the protocols available (wget, ftp, svn, git) but nothing seems to work. Hopefully you can help us.
 
You must log in or register to reply here.
Top