On most computer systems, sizeof(float) returns 4 and sizeof(double) returns 8. Suppose we wanted to implement a two-byte float. One scheme is to partition the 16 bits of two bytes as follows:
sign
exponent
fraction (n/210)
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
In this scheme, bit 15 is the sign bit, the five bits #14 through #10 encode an exponent, and the low ten bits #9 through #0 encode a fraction. We can implement the following three cases, depending on the value of the exponent: