@uekerman Do you think there are other tests that we would need here to make sure we don't break the classical subcycling? The only thing I can imagine is data initialization with subcycling. But this should already be covered by other tests.

We could also shift some tests from other places over here. I just looked into SerialTests.cpp and ParallelTests.cpp for subcycling and found the following test.

If the main purpose of this test is not subcycling, then we should leave it as it is (in this PR).

What's the difference between DoNothingWithSubcycling and ReadWriteScalarDataWIthSubcycling? Why does the one only exist for serial coupling? Do we need both? If yes, maybe we could find better names.

Removed this test in 381c7de. I was hesitant about removing an existing test, but if you also don't really see a purpose for this test anymore, we can remove it. I think the test ReadWriteScalarDataWIthSubcycling tests the functionality of DoNothingWithSubcycling and even more.

Flite and Calculix testcase

That's still one from Bernhard. Not sure whether this one is about subcycling mainly. I guess you have a better overview.

Opened an issue. I think we should delete it, because I don't really see why we need it. #1214

Maybe testing a CompositionalCouplingScheme case? Not much more comes to mind.

Added a very simple for-sake-of-completeness test in 29c6ae2. Doing something more exciting would require rewriting most of the test and we are testing compositional coupling schemes somewhere else. I think to make sure that subcycling does not do something strange this should be enough (e.g. counting iterations twice or not at all, if a compositional scheme is used).