Floating-Point Evaluations

In general, if your arithmetic evaluation has either of the characteristics listed below, it will be evaluated by the compiler in floating-point arithmetic: