From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 31791 invoked by alias); 12 Nov 2003 12:53:59 -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 31783 invoked from network); 12 Nov 2003 12:53:58 -0000 Received: from unknown (HELO mail.ukpost.com) (217.158.120.130) by sources.redhat.com with SMTP; 12 Nov 2003 12:53:58 -0000 Received: from mail.ukfsn.org (unknown [217.158.120.143]) by mail.ukpost.com (Postfix) with ESMTP id 5BADB1F8061; Wed, 12 Nov 2003 12:53:14 +0000 (GMT) Received: from localhost (dsl-80-46-233-244.access.uk.tiscali.com [80.46.233.244]) by mail.ukfsn.org (Postfix) with ESMTP id CE170E6A8A; Wed, 12 Nov 2003 12:52:11 +0000 (GMT) Received: by localhost with local (Exim 3.35 #1 (Debian)) id 1AJuVQ-0000hi-00; Wed, 12 Nov 2003 12:53:56 +0000 From: Brian Gough MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-ID: <16306.11620.309100.991173@debian.local> Date: Wed, 12 Nov 2003 12:53:00 -0000 To: Marc Baaden Cc: help-gsl@gnu.org, gsl-discuss@sources.redhat.com Subject: Re: [Help-gsl] questions about gsl_multimin_f*_minimizers (efficiency, drawbacks) In-Reply-To: <200311081040.LAA15830@apex.ibpc.fr> References: <200311081040.LAA15830@apex.ibpc.fr> X-SW-Source: 2003-q4/txt/msg00065.txt.bz2 Marc Baaden writes: > I have some questions wrt the efficiency of the routines in gsl_multimin. > I have replaced a routine in an existing Fortran code (originally using > a Quasi-Newton minimizer (Harwell VA13A)) by a call to the gsl_multimin > routines, with the choice of either conjugate_pr, conjugate_fr, > steepest_descent, vector_bfgs or nm_simplex. > > The original Harwell VA13A algorithm should be quite similar to vector_bfgs. > So I was rather surprised that there is quite a noticeable "performance" > difference, the Harwell code being roughly a factor of 1.5-2 faster. Hi, How does the choice of line-minimisation tolerance affect the comparison? In addition to the number of function evaluations it would be of interest to compare the number of direction vectors used, since this is (to some extent) independent of the line minimisation tolerance. -- Brian Gough