Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!samsung!uakari.primate.wisc.edu!uflorida!ukma!usenet.ins.cwru.edu!ncoast!catfood
From: catfood@NCoast.ORG (Mark W. Schumann)
Newsgroups: comp.lang.c
Subject: Re: # to the nth power
Summary: Yet another algorithm...
Keywords: Powers Integers
Message-ID: <1990Nov2.182217.13958@NCoast.ORG>
Date: 2 Nov 90 18:22:17 GMT
References: <9750@helios.TAMU.EDU> <1990Nov1.232830.17131@NCoast.ORG> <522@ssp9.idca.tds.philips.nl>
Sender: catfood@ncoast.org (Mark W. "Catfood" Schumann)
Organization: Random
Lines: 45

In article <522@ssp9.idca.tds.philips.nl> dolf@idca.tds.philips.nl (Dolf Grunbauer) writes:

> [I had written]:
>int powi (int root, int exponent)
>{
>int i;
>int result = 1;
>  for (i = exponent; i > 1; i--) result *= root;

Dolf says:

>Oeps, I think this one ---^ should be a 0

Nope.  I intended to leave out the iteration where you'd only be multiplying
by one.

>A quicker version (throw in a few registers if still not fast enough):
>int powi (int root, int exponent)
>{
>   int result = 1;
>
>   while (exponent > 0)
>   {
>      while (!(exponent & 1))
>      {
>         root     *= root;
>         exponent /= 2;
>      }
>      result *= root;
>      exponent--;
>   }
>
>   return result;
>}

Spiffy, but it does depend on (exponent & 1) being the same as saying
"exponent is odd."  Most implementations support this, though.
Again, neither solution supports negative exponents.

-- 
============================================================
Mark W. Schumann  3111 Mapledale Avenue, Cleveland 44109 USA
Domain: catfood@ncoast.org
UUCP:   ...!mailrus!usenet.ins.cwru.edu!ncoast!catfood
============================================================