Unformatted text preview: be oating point numbers, let +,;, ,= denote the four standard arithmetic
operations, and let , , , denote the corresponding operations as they are
actually implemented on the computer. Thus, x + y may not be a oating
point number, but x y is the oating point number which the computer
computes as its approximation of x + y . The IEEE rule is then precisely: x y = round(x + y )
x y = round(x ; y )
x y = round(x y )
and x y = round(x=y ):
16 From the discussion of relative rounding errors given above, we see then that
the computed value x y satis es x y = (x + y )(1 + )
where
for all rounding modes and jj 1
2
in the case of round to nearest. The same result also holds for the other
operations , and .
Show that it follows from the IEEE rule for correctly
rounded arithmetic that oating point addition is commutative, i.e. Exercise 13 a b=b a
for any two oating point numbers a and b.
Show with a simple example that oating point addition is not associative, i.e. it may not be true that a (b c) = (a b) c
for some oating point numbers a, b and c. Now we ask the quest...
View
Full Document
 Spring '09
 LIE
 oating point

Click to edit the document details