From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 32194 invoked by alias); 8 Nov 2003 12:51:51 -0000 Mailing-List: contact gsl-discuss-help@sources.redhat.com; run by ezmlm Precedence: bulk List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gsl-discuss-owner@sources.redhat.com Received: (qmail 24327 invoked from network); 8 Nov 2003 07:00:13 -0000 Subject: Re: GSL 1.4: BUG #8 [rk8pd.c] From: Luc Maisonobe To: gsl-discuss@sources.redhat.com Content-Type: text/plain Message-Id: <1068274810.1218.19.camel@lehrin.local.spaceroots.org> Mime-Version: 1.0 Date: Sat, 08 Nov 2003 12:51:00 -0000 Content-Transfer-Encoding: 7bit X-SW-Source: 2003-q4/txt/msg00052.txt.bz2 Hello, The remaining problem is still the same as one year ago: the comments say the method is an 8(9) method (i.e. extrapolation beeing of order 8 and error estimation of order 9), but it is really an 8(7) method. Despite I think nobody is interested anymore, I promised to check the coefficients of the method, so here are my results. The maximal error in rk8pd.c is about 3.92e-17. The maximal error in rksuite.f is about 4.49e-29. These values were computed by picking two coefficients from rksuite (c(9) and a(8,4)) as the remaining free parameters of the model, transforming them into rational numbers by continued fractions method and using them to solve exactly all the equations of conditions and all the additional equations from the article. The exact result is available in http://www.spaceroots.org/software/rkcheck/DormandPrinceRKsuite.xml Luc