Unformatted text preview: we will say a general real number is \normalized" if its
modulus lies between the smallest and largest positive normalized oating
point numbers, with a corresponding use of the word \subnormal". In both
cases the representations we give for these numbers will parallel the oating
point number representations in that b0 = 1 for normalized numbers, and
b0 = 0 with E = ;126 for subnormal numbers.
For any number x which is not a oating point number, there are two
obvious choices for the oating point approximation to x: the closest oating
point number less than x, and the closest oating point number greater than
x. Let us denote these x; and x+ respectively. For example, consider the
toy oating point number system illustrated in Figures 1 and 2. If x = 1:7,
for example, then we have x; = 1:5 and x+ = 1:75, as shown in Figure 3. : : :: : : 0 1 x2 3 Figure 3: Rounding in the Toy System
Now let us assume that the oating point system we are using is IEEE single precision. Then if our genera...
View Full Document
- Spring '09
- oating point