Compiling WRF-Chem on Derecho #326
Conversation
mrinalbiswas
requested review from
benkirk,
dphow,
mtrahan41,
negin513,
roryck,
rvalent and
vanderwb
| git clone --recurse-submodule --branch v4.7.1 https://github.com/wrf-model/WRF.git | ||
| ``` | ||
| This will create a local copy of the repository that you can edit and manage using Git. | ||
|
|
There was a problem hiding this comment.
A minor suggestion, but I suggest removing this sentence and covering all WPS under the WPS portion of this page, since WPS is already covered at the bottom. What are your thoughts on this?
"After downloading the WRF model, the next step is to download and extract the WRF Preprocessing System (WPS). "
There was a problem hiding this comment.
@negin513 I was also torn in deciding whether to include WPS or not. So I did refer to the other WRF documentation for more information. I will look into this again. Thanks.
docs/tutorials/wrfchem_on_derecho.md
Outdated
| ```bash | ||
| ls main/*.exe | ||
| ``` | ||
| If the compilations fail, follow the [WRF Users Guide](https://www2.mmm.ucar.edu/wrf/users/wrf_users_guide/build/html/compiling.html) for directions to debug. |
There was a problem hiding this comment.
Here we can add a sentence on "or please contact [email protected] for additional help compiling the code"?
docs/tutorials/wrfchem_on_derecho.md
Outdated
| Once WRF is successfully compiled, the next step is to compile the WRF Preprocessing System (WPS), which is required to process input data for WRF. Please follow the directions to [compile WPS](https://ncar-hpc-docs.readthedocs.io/en/latest/tutorials/wrf_on_derecho/#compiling-the-wrf-preprocessing-system-wps) on Derecho. | ||
|
|
||
| ## Submitting jobs | ||
| WPS jobs can be run in an interactive mode. However, WRF jobs are more memory-intensive and should be submitted using a PBS batch script. |
There was a problem hiding this comment.
I'd suggest removing this sentence. "WPS jobs can be run in an interactive mode. However, WRF jobs are more memory-intensive and should be submitted using a PBS batch script. "
negin513
left a comment
negin513
left a comment
There was a problem hiding this comment.
This looks great to me! Thanks so much @mrinalbiswas for drafting this! I added a few minor suggestions.
| ```bash | ||
| qsub script_name | ||
| ``` | ||
| To know more about starting and managing jobs, please refer to the [documentation](https://ncar-hpc-docs.readthedocs.io/en/latest/pbs/). |
There was a problem hiding this comment.
One general point worth highlighting is that WRF-Chem runs typically use significantly more memory than standard WRF runs, as they include additional chemistry species. This often increases the memory footprint per core required for running the case and sometimes users should be aware of this.
vanderwb
left a comment
vanderwb
left a comment
There was a problem hiding this comment.
I started a review but got pulled into other work. Negin is doing a more thorough review I believe, but I didn't want these pending comments to disappear into the ether. 😃
docs/tutorials/wrfchem_on_derecho.md
Outdated
| @@ -0,0 +1,253 @@ | |||
| # How to compile and conduct a basic WRF-Chem on Derecho | |||
docs/tutorials/wrfchem_on_derecho.md
Outdated
| export FLEX_LIB_DIR=/usr/lib64 | ||
| ``` | ||
| ## Configuring the Build | ||
| Once the environment is set up, you can configure the code for compilation. It is strongly recommended to use the CMake-based configuration option provided with recent versions of WRF, as it simplifies and standardizes the build process. |
There was a problem hiding this comment.
I like the CMake process, but given that the WRF devs seem to be recommending the traditional system still (https://www2.mmm.ucar.edu/wrf/users/wrf_users_guide/build/html/compiling.html#building-the-wrf-code), should we be strongly recommending something else? Does that recommendation have the support of the devs?
docs/tutorials/wrfchem_on_derecho.md
Outdated
| Next step is to compile the code using: | ||
|
|
||
| ``` | ||
| ./compile_new >& compile.log |
There was a problem hiding this comment.
Maybe ./compile_new |& tee compile.log instead? I suspect users prefer to understand whether something is happening since it's a long compile.
| ``` | ||
| Upon successful compilation, the executables wrf, real, ndown, and tc will be generated and located in the _build/main/ directory. Compilation progress and details are recorded in the compile.log file. If the compilation fails, refer to compile.log for diagnostic messages and error information to assist with troubleshooting. | ||
|
|
||
| ## Compiling WRF-Chem with traditional method on Derecho |
There was a problem hiding this comment.
Following on from my prior comment - is there a reason to document both methods here? Unless each method only provides certain functionalities, it seems unnecessary to me.
docs/tutorials/wrfchem_on_derecho.md
Outdated
|
|
||
| - Locate the line that defines FCDEBUG. It may be commented out. | ||
|
|
||
| - Uncomment and modify it to the following: |
There was a problem hiding this comment.
Is there a reason this is preferred over using ./configure -d? That's what I've always done.
|
I echo @negin513's compliments! I put "request changes", but feel free to not change anything I've pointed out if you disagree - just ask me for a re-review and I'll approve. |
vanderwb
left a comment
vanderwb
left a comment
There was a problem hiding this comment.
Thanks for the fixes, @mrinalbiswas. This LGTM now and should be a great resource!
4 participants
This pull request updates the instructional webpage to provide clear and current guidance for downloading and compiling WRF-Chem on NCAR’s Derecho system. For WPS: Links to the HPC Documentation are provided to compile WPS to avoid duplication and minimize updates in multiple files.
Included changes:
Clarified instructions for users compiling WRF-Chem (Cmake and traditional).
Updated guidance on choosing appropriate GLADE storage locations based on quotas and purging policies