There are many possible formats for floating point numbers. Consider the following 10-format

where the sign ield is one bit (1: negative, 0: positive), the exponent ield is 4 bits with an excess 8 number for the bias, and the fraction ield is a 5-bit normalized number (i.e., for all numbers except 0, the number is stored as 0.1xxxxx).

(a) What are the largest and smallest positive nonzero quantities for normalized numbers?

(b) Using this loating point format, what is the binary representation of decimal -5.675?

Transcribed Image Text:

Sign Exponent Fraction