From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 27121 invoked by alias); 9 Feb 2010 08:47:23 -0000 Received: (qmail 27111 invoked by uid 22791); 9 Feb 2010 08:47:22 -0000 X-SWARE-Spam-Status: No, hits=-0.0 required=5.0 tests=AWL,BAYES_50,DATE_IN_PAST_12_24,SPF_PASS X-Spam-Check-By: sourceware.org Received: from mail.network-theory.co.uk (HELO mail.network-theory.co.uk) (66.199.228.187) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 09 Feb 2010 08:47:19 +0000 Date: Tue, 09 Feb 2010 08:47:00 -0000 Message-ID: <873a1bk6nu.wl%bjg@network-theory.co.uk> From: Brian Gough To: Rhys Ulerich Cc: gsl-discuss@sourceware.org Subject: Re: Where a generalized Richardson extrapolation routine would fit in GSL? In-Reply-To: <4a00655d1002060725q21794722n12f69160c26e381a@mail.gmail.com> References: <4a00655d0908201247g7d7bd9a1t466f4a66f08df4@mail.gmail.com> <4a00655d0911291538y9f29830v984d1a796fdd5d1@mail.gmail.com> <4a00655d0912131549w19638273nb51d723e9ddd9273@mail.gmail.com> <87d42ggnsv.wl%bjg@network-theory.co.uk> <4a00655d0912150931q4b34fd24p94594ed08857254f@mail.gmail.com> <87zl5ig7i0.wl%bjg@network-theory.co.uk> <4a00655d0912161643k63868ac4i2577f618f6e2196@mail.gmail.com> <87zl5hnrkh.wl%bjg@network-theory.co.uk> <87ljh0nqp4.wl%bjg@network-theory.co.uk> <4a00655d1002060725q21794722n12f69160c26e381a@mail.gmail.com> User-Agent: Wanderlust/2.14.0 (Africa) Emacs/22.1 Mule/5.0 (SAKAKI) MIME-Version: 1.0 (generated by SEMI 1.14.6 - "Maruoka") Content-Type: text/plain; charset=US-ASCII X-Message-Mac: e7432021530fbe552220df7d51a083be Mailing-List: contact gsl-discuss-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gsl-discuss-owner@sourceware.org X-SW-Source: 2010-q1/txt/msg00017.txt.bz2 At Sat, 6 Feb 2010 09:25:47 -0600, Rhys Ulerich wrote: > I'm good on moving the interface away from gsl_matrix and gsl_vector. > No problem there. But, from looking a bit at the levin sum > documentation, I'm not sure how to apply the abserr concept to > Richardson extrapolation. Any guidance would be much appreciated. The error term in the Levin sum is calculated by simple error propagation assuming the individual terms are accurate to double precision. Therefore it is just a measure of the cumulative cancellation error. Since there are a lot of subtractions, it's worth tracking this. -- Brian Gough