21. Which of the following exception conditions are recognized and detected by the floating point unit (FPU)?

a. Numeric underflow

b. Numeric overflow

c. Divide by zero

d. Inexact precision

22. How many operands can a floating-point instruction have?

a. Must have at least one operand.

b. Can have zero operands.

c. Can have two operands.

d. Can have one immediate operand.

23. What effect does initializing the floating point unit (FPU) have on the contents of the FPU control word.

a. Nothing, the FPU control word is not changed at initialization.

b. All floating-point exceptions are masked.

c. Rounding is set to nearest even.

d. Calculation precision is set to 80 bits.

24. How are unsigned comparisons performed by the floating point unit (FPU)?

a. Twos-complement is used in the comparison.

b. A default value is used for the sign.

c. Never, floating-point values are implicitly signed.

d. Integer subtraction is used.

25. Which of the following statements accurately describes floating point comparisons?

a. They incur less runtime overhead than integer comparisons.

b. They incur more runtime overhead than integer comparisons.

c. Conditional jumps using the contents of the EFLAGS register are immediately available after a comparison.

d. The extreme accuracy of floating-point numbers makes comparing for equality much easier.