This MR fixes two tests which fail in the pipeline.
test_version_string because it was just broken and
test_cumulant which failed due to timeout. This case was fixed by replacing a symbolic comparison between collision rules with a numeric comparison.
Furthermore, some long-running cases are parallelised now which should decrease the overall time a little bit and increases the readability when a test case fails since py-test will show the exact parameter setup now on failure.