Improve handling of long workgraph tests for use case parameters

[agoscinski]

In PR #74 we reduced the test size as much as possible without affecting coverage but it still takes considerable amount of time to run the test (on my machine 70s). This can be still improved by doing one of the following things:

• We could split up the test case into smaller cases to not affect coverage
• We could run it concurrently using submit instead of run (even though I think this will not help so much with the CI because it only has two processes running)
• The overall test time could be reduced by executing them in parallel with pytest-xdist or pytest-run-parallel (again I am not sure how much this will help with CI)
• Mark it as long test so it can be turned of easily when developing