Path: csiph.com!v102.xanadu-bbs.net!xanadu-bbs.net!news.glorb.com!news-in-01.newsfeed.easynews.com!easynews!core-easynews-01!easynews.com!en-nntp-15.dc1.easynews.com.POSTED!not-for-mail
From: Robert Wessel <robertwessel2@yahoo.com>
Newsgroups: comp.programming
Subject: Re: About divisions on the Taylor series in the error function
Message-ID: <tcf6o7hml1brnlqiq2rr7h2p9ipdibshfk@4ax.com>
References: <jlur7p$42p$1@dont-email.me>
X-Newsreader: Forte Agent 6.00/32.1186
MIME-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
Lines: 16
X-Complaints-To: abuse@easynews.com
Organization: Forte Inc.  http://www.forteinc.com/apn/
X-Complaints-Info: Please be sure to forward a copy of ALL headers otherwise we will be unable to process your complaint properly.
Date: Mon, 09 Apr 2012 14:52:28 -0500
X-Received-Bytes: 1227
Xref: csiph.com comp.programming:1452

On Mon, 09 Apr 2012 16:20:34 +0200, hopcode <hopcode@invalid.de>
wrote:

>Hi,
>i am actually using the Marsaglia's method
>to calculate the error function from
>(snipped) this series
>x + x3/3 + x5/(3 · 5) + x7/(3 · 5 · 7)
>
>is there a way to avoid divisions in there?
>what about a method interpolating on a precalculated
>lookup table ?


Replace the divisions with the usual multiplications by the
reciprocal.