Unformatted text preview: e question which then arises is: what about
;1? It turns out to be convenient to have representations for ;1 as well
as 1 and ;0 as well as 0. We will give more details later, but note for now
that ;0 and 0 are two di erent representations for the same value zero, while
;1 and 1 represent two very di erent numbers. Another special number
is NaN, which stands for \Not a Number" and is consequently not really a
number at all, but an error pattern. This too will be discussed further later.
All of these special numbers, as well as some other special numbers called
subnormal numbers, are represented through the use of a special bit pattern
in the exponent eld. This slightly reduces the exponent range, but this is
quite acceptable since the range is so large.
There are three standard types in IEEE oating point arithmetic: single
precision, double precision and extended precision. Single precision numbers
require a 32bit word and their representations are summarized in Table 1.
Let us discuss Table 1 in some detai...
View
Full Document
 Spring '09
 LIE
 oating point

Click to edit the document details