...
| No Format |
|---|
cp $M3HOME/scripts/cctm/run.cctm . |
Then run that script. It will use the binaries located in the distribution but with whatever model changes were set in the script. Of course, computations should be run on compute nodes so the module loading, setting MM3DATA, and running the CMAQ script would all be entered into an SGE job script.
Since there is no write access to the CMAQ distribution, if you need to recompile some components, you can copy the parts of the distribution, or the entire distribution, and compile with the build scripts. In this case, you will need to pay attention to the value of $M3HOME as that will be different from what is set in the environment module. You will also need to set the value of M3HOME in the config.cmaq script to point to where you copied the source to. Using this method you will still be using the patched scripts to build with the same tools as the central distribution. Note that even if you copy the distribution and compile you will still want to load the CMAQ environment module to set up all of the dependencies. You will want to pay attention to the M3HOME and EXEC_ID variables but those should both be reset The M3HOME variable in the config.cmaq file script will be used for building but you will need to set its value appropriately for your local copy.
To reiterate, if you do not need to recompile code, you can just edit run scripts to change settings for computations. Settings can either point to data in the shared installation or local data. All of the parameter and profile data that is part of CMAQ is in the CMAQ installation tree.