Subject

    Computer Science

    Topic

    Computer Hardware

    Which number format maintains a constant relative error, ideal for very small and very large values?

    ক)
    Two’s complement integers
    খ)
    IEEE-754 floating point
    গ)
    Fixed point
    ঘ)
    BCD

    Explanation

    Floating point represents numbers as mantissa×base^exponent, giving wide dynamic range. Rounding errors are proportional to magnitude (relative). Fixed point has uniform absolute steps, suitable for embedded control. IEEE-754 defines NaN, infinities, and rounding modes.

    Find Tutors

    View all Tutors