Answer 1

Depends on the architecture. If double is native, but float is simulated, double will be faster.

In any case, you should probably use double by default anyway. Only use float if you have a very good reason to do so.