05.08.2014 Views

Lecture Notes for Computer Architecture II - St. Cloud State University

Lecture Notes for Computer Architecture II - St. Cloud State University

Lecture Notes for Computer Architecture II - St. Cloud State University

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

<strong>Lecture</strong> 17 Per<strong>for</strong>mance Enhancement with added Functionality - Floating Point<br />

Adder<br />

In Computations<br />

We Need<br />

Speed<br />

Accuracy<br />

And range<br />

Page | 106<br />

Integers does not provide the accuracy needed in many computations<br />

and the range is limited<br />

The use of Floating point numbers provide the accuracy and range increase<br />

Floating numbers<br />

3.656914 10 2<br />

Sign Digit Fraction Exponent<br />

We can use the following <strong>for</strong>mat to represent the floating point numbers<br />

Sign Exponent<br />

Fraction<br />

There are two different <strong>for</strong>mats to increase the precision of computations.<br />

64 bit – Double precision<br />

32 bit – Single precision<br />

To avoid conflicts between different manufacturers IEEE proposed the standard in 1979<br />

32 bit <strong>for</strong>mat<br />

Sign Exponent Fraction<br />

1 8 23<br />

64 bit<br />

Sign Exponent Fraction<br />

1 11 52

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!