Provide reference results for ScalableTestSuite test runs

[Francesco Casella]

We now have several different ScalableTestSuite runs, using old/new FE and ode/dae mode. It would be good to also verify the results obtained with the newest configurations agains reference ones.

I would suggest to used the results obtained using the settings of ​this test run, which should be reliable, as reference values.

@sjoelund.se, can you please take care of this?

[Martin Sjölund]

We don't store any simulation results. They have to be constructed manually (and I suggest selecting which variables to store in the result-file manually as well; something I cannot do). Especially for the larger test-cases, the result-files are otherwise huge.

[Francesco Casella]

OK, I guess I'll have to build a script to generate those results, I'll try to pick some variables which are representative of the correctness of the results.

Are .mat files OK as a final product? How can I provide them to you?

[Martin Sjölund]

If the files are small and/or few (only verification models, etc), they fit in ScalableTestSuite.git somewhere (​https://openmodelica.org/doc/OpenModelicaUsersGuide/latest/scripting_api.html#filtersimulationresults can shrink files considerably by only including desired variables). Mat-files are fine.

If the files are large, we'll need to store them somewhere else (if they fit on GitHub, a new repository with the reference files?) or some sftp that we can provide for you.

[Francesco Casella]

I was thinking of using the regex filter provided by the variableFilter simflag, but that API function is also a nice option.

I'll see what the size of the outcome is and then we can decide on where to store the results. I guess it would be good to keep it public, so that other people can access it and use it.

[Francesco Casella]

The test cases are generated by the attached script and stored in the ​ScalableTestSuite-ref GitHub repo. They are correctly imported by the Hudson job and used for verification.