home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.programming
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!cs.utexas.edu!torn!watserv1!watmath!thinkage!atbowler
- From: atbowler@thinkage.on.ca (Alan Bowler)
- Subject: Re: Calculating a cube root
- Message-ID: <1992Jul22.213049.3589@thinkage.on.ca>
- Organization: /etc/organization
- References: <1992Jul20.123722.27958@cs.ruu.nl> <1992Jul20.160255.17562@organpipe.uug.arizona.edu> <1992Jul21.085023.11890@cs.ruu.nl>
- Date: Wed, 22 Jul 1992 21:30:49 GMT
- Lines: 10
-
- The easy way to go is clearly to use the exp + log library routines.
- If that is not quite accurate enough for you, run the answer through
- a single Newton iteration. If this is not fast enough, then you
- will have to resort to a customized routine. You may find something
- published on cube roots explicitly, or you can certainly look up how to
- do a Chebyshev fit of a rational function ( P(x)/Q(x) where P and Q are
- polynomials (Remez'z algorithm?)). In any case the speed will vary
- with the machine. A number of the IEEE standard implementations have
- hardware support for exp and log, and it may be impossible to
- outperform that.
-