Any explanation as to why one simulation runs slower/faster could depend on many additional factors than the algorithm used. If you post the models, we can take a look at them and see if there are additional factors affecting the speed of solution.
Generally speaking, the inverse dynamics problem is less computationally expensive. In forward dynamics, you have more states to calculate (position, velocity, and acceleration). For inverse dynamics fewer calculations are needed to calculate the forces as more of the states have been defined as inputs from the user.
--Steve
