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

Running iCESM1.2 on Cheyenne?

M

michael_p_erb@gmail_com

New Member
Hi. I would like to run some iCESM experiments on Cheyenne using the code here: GitHub - NCAR/iCESM1.2: Isotope-enabled CESM1.2. After setting up a test run and executing the *.run file, I get the following error in my logfile:

MPT ERROR: bad process count
(HPE MPT 2.19 02/23/19 05:31:12)

The logfile says nothing else. I searched for that error on these forums and found a thread (Running F compset of CESM1_2_2 on cheyenne: MPT ERROR: bad process count) which says that CESM1.2 is no longer supported. So, is it possible to run new iCESM simulations? If so, are there new instructions somewhere? Thanks!
 
D

dbailey

CSEG and Liaisons
Staff member
Unfortunately, iCESM has not been developed for CESM2 yet. However, I do think that version 1.2 should run. I am moving this to the paleo forum.
 
M

michael_p_erb@gmail_com

New Member
Thanks for the replies. I feel like I'm missing something, though. I cloned the "GitHub - NCAR/iCESM1.2: Isotope-enabled CESM1.2" directory and then followed the instructions here: iCESM1.2/water_isotope_instructions.txt at master · NCAR/iCESM1.2. @jiangzhu (or anyone else), could you look at my commands below and see if I'm missing something?


Create a new case with these commands
cd /glade/work/merb/iCESM1.2/scripts
./create_newcase -case ~/cesm/iso_test1 -res f19_g16 -compset B1850C5CN -mach cheyenne

Go to ~/cesm/iso_test1/ and do steps 2-4 in the instructions above.

For step 5, copy the restart files into the run directory
cp /glade/work/merb/data_iCESM1.2/restart_files/preind/* /glade/scratch/merb/iso_test1/run/

Make sure the new run uses the files copied above, with the following commands:
./xmlchange RUN_REFCASE=b.ie12.B1850C5CN.f19_g16.LME.002
./xmlchange RUN_REFDATE=1850-01-01
./xmlchange GET_REFCASE=FALSE
./xmlchange RUN_STARTDATE=1850-01-01

Make more changes:
In user_nl_cice, add the line: “tr_iso = .true.”
In user_nl_rtm, add the line: “wiso_runoff = .true.”
In iso_test1.run, add the line toward the beginning: “#PBS -A my_account_number”

Run the model:
./iso_test3.build
./iso_test3.submit


Am I missing something? These commands resulted in the "MPT ERROR: bad process count" error that I mentioned above. Are there any changes that I omitted? Thanks!
 
M

michael_p_erb@gmail_com

New Member
Alright, I've tried a few more options and I'm still running into problems. Any tips on how to resolve this?

Attempt 1: After realizing that the command "./iso_test1.run" command wasn't working, I tried the "./iso_test1.submit" command, and it got farther. It ended up giving me an error which says:

1: Opened existing file
1: b.ie12.B1850C5CN.f19_g16.LME.002.clm2.r.1850-01-01-00000.nc 1572864
1: check_dim ERROR: mismatch of input dimension 8012 with expected value
1: 8010 for variable landunit

I'd been using the restart files mentioned in the instructions (iCESM1.2 restart file from "The Connected Isotopic Water Cycle in the Community Earth System Model Version 1"), so this is an odd problem, but it seems like the solution is to use interpinic to regrid this restart file. However, to regrid this file, I need a reference file with the right grid, so I decided to try a "startup" run.

Attempt 2 (startup run): To run a startup simulation, I made the following changes. I want to run a 1-month simulation to use as reference for interpinic.

./xmlchange RUN_TYPE=startup
./xmlchange RUN_STARTDATE=1849-12-01
./xmlchange DATM_CLMNCEP_YR_ALIGN=1849
./xmlchange CLM_FORCE_COLDSTART=on
./xmlchange STOP_OPTION=nmonths
./xmlchange STOP_N=1
./xmlchange RESUBMIT=0

This simulation ran for a while, but ultimately failed with a "walltime exceeded" error. I'd given the simulation 12 hours, which seems like plenty of time to finish a 1-month simulation, so I'm not sure why it failed.

Attempt 3: I repeated the startup experiment above with the pe_layout that @jiangzhu mentioned, and it seems like it didn't get as far, ending with the error:

1: Opened existing file
1: /glade/p/cesmdata/cseg/inputdata/atm/cam/topo/USGS-gtopo30_1.9x2.5_remap_c05060
1: 2.nc 131072
1: NetCDF: Invalid dimension ID or name
3:(shr_sys_abort) ERROR: SPMD_DYN_SET : incorrect yz domain decomposition - aborting
3:(shr_sys_abort) WARNING: calling shr_mpi_abort() and stopping

Any suggestions on what I should try next? Have I set things up wrong? Thanks!
 
J

jiangzhu

Member
Hi Michael,

Usually, it's best to provide water isotope initial conditions, otherwise you need a long time to spin up the isotope values (e.g. in soil water and subsurface ocean). So, your Attempt 1 makes more sense.

The error message you got with Attempt 1 was caused by mismatch between clm surface data and restart files.
You could try to use the clm surface data included with the restart files, through namelist setting in user_nl_clm: fsurdat = '/YOUR_DIRECTORY/surfdata_1.9x2.5_simyr1850_c140303.nc'
 
M

michael_p_erb@gmail_com

New Member
Thanks @jiangzhu

I redid Attempt #1 above after prescribing that clm surface data file and everything seems to have worked. Now that things are running I just need to figure out the right settings for my experiments. Is there good checklist somewhere for setting up a new experiment? I've run the CESM before, but it's been years. I don't want to overlook something important.
 
L

LisaCickyham

LisaCickyham
New Member
michael_p_erb@gmail_com said:
Hi. I would like to run some iCESM experiments on Cheyenne using the code here: GitHub - NCAR/iCESM1.2: Isotope-enabled CESM1.2. After setting up a test run and executing the *.run file, I get the following error in my logfile:

MPT ERROR: bad process count
(HPE MPT 2.19 02/23/19 05:31:12)

The logfile says nothing else. I searched for that error on these forums and found a thread (Running F compset of CESM1_2_2 on cheyenne: MPT ERROR: bad process count) which says that CESM1.2 is no longer supported. So, is it possible to run new iCESM simulations? If so, are there new instructions somewhere? Thanks!
Download hot entertainment gaming apps at Techzapk, the leading download site for free entertainment apps.
Click to expand...
Based on the information you provided, it appears that you are encountering an error related to the "MPT ERROR: bad process count" while attempting to run iCESM experiments on Cheyenne using the iCESM1.2 code from the NCAR GitHub repository. The thread you mentioned from the forums indicates that CESM1.2 is no longer supported, which suggests that running new iCESM simulations with that specific version may not be recommended. To determine if there are new instructions or updated versions available for running iCESM simulations, it would be best to refer to the official documentation and resources provided by the iCESM project or NCAR. They will have the most up-to-date information on supported versions, instructions, and any specific requirements for running iCESM experiments. It's possible that newer versions of iCESM or alternative frameworks may be available for running isotope-enabled CESM experiments. Checking the project's official website, GitHub repository, or contacting the project maintainers or support channels can provide you with the necessary guidance and information for running iCESM simulations using the latest available resources. Keep in mind that it's important to follow the recommended and supported versions to ensure the best compatibility and functionality of the iCESM simulations you intend to run. I encourage you to explore the official documentation and reach out to the iCESM project or NCAR for specific instructions and support related to running iCESM experiments on Cheyenne or other platforms. Best of luck with your iCESM simulations, and if you have any further questions, feel free to ask!
 
You must log in or register to reply here.
Top