Check this thread, part (but only part) of explicit solver algorithm is not done in parallel:
When it comes to rigid body, it should work in explicit dynamics procedure as well. Apparently, you came across some bug. This issue would require further research. Can you share more details about this?