WPS/WRF is being added to E4S (E4S - Home) so we are creating a 'sanity check' for our testsuite to make sure the spack deployed versions are working on the various platforms E4S supports. I am trying to track down a quick example for both wps and wrf that I can script.
This tutorial: ARW Online Tutorial requires downloading a 30gb geographic dataset which exceeds the resources we want our testsuite to consume (though for our wps test I am tentatively using that procedure up through ungribbing the mathew dataset). A run on an idealized test as described here Running WRF — WRF Users Guide documentation would be good for us (it's nice to have the test data provided by the install) but our spack-installed wrf doesn't include ideal.exe for some reason. real.exe is present. The real.exe example on that page doesn't go into detail on the wps procedure and I suspect that would still require a large geographic dataset.
Are there any good options for quick, small-scale tests we could run, just to confirm a valid installation and maybe check for obvious performance regressions between deployments? Alternatively, if anyone knows why ideal.exe might be absent from our install we could try to fix the spack package so it shows up and could then run with the ideal example. Thanks!
This tutorial: ARW Online Tutorial requires downloading a 30gb geographic dataset which exceeds the resources we want our testsuite to consume (though for our wps test I am tentatively using that procedure up through ungribbing the mathew dataset). A run on an idealized test as described here Running WRF — WRF Users Guide documentation would be good for us (it's nice to have the test data provided by the install) but our spack-installed wrf doesn't include ideal.exe for some reason. real.exe is present. The real.exe example on that page doesn't go into detail on the wps procedure and I suspect that would still require a large geographic dataset.
Are there any good options for quick, small-scale tests we could run, just to confirm a valid installation and maybe check for obvious performance regressions between deployments? Alternatively, if anyone knows why ideal.exe might be absent from our install we could try to fix the spack package so it shows up and could then run with the ideal example. Thanks!