B6 - Plant Hydraulics Performance Phase 1
AttachmentsThis page should describe Performance Assessment Tests performed for this stand alone feature and should provide links to all the result pages.
Summary
A two-year long simulation was performed with ELM+PHS. The test was performed on COMPY using compset ICLM45 and grid resolution ne30_oEC.
Simulation was completed with no numerical issues.
Performance Test 1
Performance Test 1: Two-year long simulation with ELM+PHS
Date last modified: Jun 27, 2019
Contributors: (add your name to this list if it does not appear)
Provenance: (Run provenance Link, Code Tag, etc:)
Test ran on COMPY
Run length = 2 years
Number of MPI Tasks = 128
Number of threads = 1
Hydrology model : Default ELM v0 model
Results: (link to results, data and plots)
Available on COMPY in file /compyfs/d3m045/global_run_clmphs_review/ICLM45/timing/e3sm_timing.ICLM45.1243.190627-160438
TOT Run Time | 3459.374 seconds | 4.739 seconds/mday | 49.95 myears/wday |
CPL Run Time | 446.780 seconds | 0.612 seconds/mday | 386.77 myears/wday |
ATM Run Time | 376.393 seconds | 0.516 seconds/mday | 459.09 myears/wday |
LND Run Time | 2079.490 seconds | 2.849 seconds/mday | 83.10 myears/wday |
ROF Run Time | 213.133 seconds | 0.292 seconds/mday | 810.76 myears/wday |
CPL COMM Time | 1370.002 seconds | 1.877 seconds/mday | 126.13 myears/wday |
:
Comments
Yilin Fang I am trying to look at the effect of the plant hydraulic model on overall land model throughput. I found that you used the ICLM45ED compset here. I wonder if you had some reasons to use FATES (ICLM45ED) instead of ELM (ICLM45). Based on your branch in github, it seemed that all code changes were in ELM. Thanks.
Min Xu (Unlicensed) That was a mistake in the summary. I've corrected it. You can see the directory name "/compyfs/d3m045/global_run_clmphs_review/ICLM45" that I used ICLM45.
Thanks a lot. I doubt it may be just a typo. I have cloned your repo and set up the ICLM45 case on Summit, but it seemed that the plant hydraulic model needs some additional parameters such as krmax that are not in the parameter file in default. Did you update these land model input files and upload to the e3sm input data web server? I am sorry that I have not had a Compy account yet, so I cannot check out your case directory on Compy. Could you point me any changes in namelist and input datasets that are needed for successfully running the model? Thanks!
I didn't update the land model input files to the web server as it will take a while to parameterize them for application. To trigger the plant hydraulics model, I modified the CLM namelist:
cat >> user_nl_clm <<EOF
paramfile = '~/projects/E3SM_scripts/cases/clm_params_n.nc'
use_hydrstress = .true.
EOF
File clm_params_n.nc is attached here.
Thanks a lot. I will use the file for performance testing.