Review Questions for Floating Point Representation
- Convert a decimal number to binary.
- Convert a binary number to decimal.
- What are the differences between floating point and fixed point representation?
- Given a real number, how would you store it as a machine number?
- Given a real number, what is the rounding error involved in storing it as a machine number? What is the relative error?
- Explain the different parts of a floating-point number: sign, significand, and exponent.
- How is the exponent of a machine number actually stored?
- What is machine epsilon?
- What is underflow (UFL)?
- What is overflow?
- Why is underflow sometimes not a problem?
- Given a toy floating-point system, determine machine epsilon and UFL for that system.
- How do you store zero as a machine number?
- What are subnormal numbers?
- How are subnormal numbers represented in a machine?
- Why are subnormal numbers sometimes helpful?
- What are some drawbacks to using subnormal numbers?