Hum, the goal of the test is to test side effect and if there is any major deviation from the expected value. Maybe testing with larger, more random and fewer step is good enough.
What do you mean by 'larger' in this case?
Hm well changing to:
step1 = 0.007;
step2 = 0.061;
yields significant performance increases while providing the same coverage. Seems good enough.
sounds like a good solution to me too.
I asked @vtorri about it, will probably merge this for now to improve speed on a known-functional component and then update at a later point after a deeper review has been done.