Adaptive SDESolver yields "Warning: dt <= dtmin. Aborting", unsure why, large non-adaptive dt works fine

[ChrisRackauckas]

@Torkel

In the non-adaptive time-stepper, there’s a fixed dt , so it never actually checks the error, and this is never a problem. If I set my fixed dt too high, I get a

that’s different, that’s the solution diverging.

The error is measured in strong error as well, so it’s not necessarily the error of the average.