Search All of the Math Forum:

Views expressed in these public forums are not endorsed by Drexel University or The Math Forum.

Topic: Root
Replies: 21   Last Post: Jul 14, 1999 3:46 PM

 Messages: [ Previous | Next ]
 John Popelish Posts: 37 Registered: 12/8/04
Re: Root (EXACT Square Root Algorithm)
Posted: Jul 12, 1999 4:45 PM

Gareth McCaughan wrote:
>
> Thomas Chrapkiewicz wrote:
>

> > This method for square roots is EXACT for the number of digits that you have
> > the patience to use it to.

> [SNIP]
>
> It might be worth mentioning that this algorithm works especially
> nicely in binary.
>
> If you can do division efficiently, another possibility is to use
> Newton's method (though I think this particular application of it
> goes back to before Newton): start with any half-decent guess at
> the square root and iterate x -> (x+a/x)/2 where a is the thing
> whose square root you want.

AsI recall, I once coded a fast Newton's method square root algorithm
that minimized the number of iterations required by referring to a 256
entry lookup table of guesses, and used the most significant 8 bits as
pointer. There was a little more to it than that, but a good guess can
shorten things considerably.

John Popelish

Date Subject Author
7/10/99 User
7/10/99 Prasanth A. Kumar
7/11/99 William L. Bahn
7/11/99 Kaimbridge
7/11/99 carel
7/11/99 chri0562@sable.ox.ac.uk
7/11/99 spamless@nil.nil
7/12/99 test@test.com
7/14/99 Chris J. Bennardo
7/12/99 John Popelish
7/12/99 philnovi
7/12/99 Ken Cox
7/12/99 randy
7/13/99 John Savard
7/13/99 Paul J. Gans
7/13/99 Jack Silver
7/13/99 Paul Schlyter
7/13/99 philnovi
7/12/99 Gareth McCaughan
7/12/99 randy