Floating-point maximum value (single precision) (Floating-point Maximum (Single))
This instruction extracts the maximum value from the single-precision floating-point format data in general-purpose registers reg1 and reg2, and stores it in general-purpose register reg3.
If one of the source operands is S-NaN, an IEEE754-defined invalid operation exception is detected.
If invalid operation exceptions are not enabled, Q-NaN is stored and no exception occurs.