From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 3294 invoked by alias); 12 Nov 2003 15:25:54 -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 3256 invoked from network); 12 Nov 2003 15:25:53 -0000 Received: from unknown (HELO sterling.fpcc.net) (204.144.241.126) by sources.redhat.com with SMTP; 12 Nov 2003 15:25:53 -0000 Received: from sterling.fpcc.net (sterling.fpcc.net [127.0.0.1]) by sterling.fpcc.net (8.12.8/8.12.8) with ESMTP id hACFQP26028368; Wed, 12 Nov 2003 08:26:25 -0700 Received: (from apache@localhost) by sterling.fpcc.net (8.12.8/8.12.8/Submit) id hACFQP9f028366; Wed, 12 Nov 2003 08:26:25 -0700 X-Authentication-Warning: sterling.fpcc.net: apache set sender to rlee@fpcc.net using -f Received: from piko.cr.usgs.gov ([136.177.22.105]) (SquirrelMail authenticated user rlee) by webmail.fpcc.net with HTTP; Wed, 12 Nov 2003 08:26:25 -0700 (MST) Message-ID: <52642.136.177.22.105.1068650785.squirrel@webmail.fpcc.net> Date: Wed, 12 Nov 2003 15:25:00 -0000 Subject: Re: Here is some GSL code for a rank-revealing QRPT From: To: X-ISP: fpcc.net X-Abuse-Contact: abuse@fpcc.net X-UserAgent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20030225 In-Reply-To: <16306.13683.266366.946535@debian.local> References: <35906.64.24.109.89.1068572699.squirrel@webmail.fpcc.net> <16306.13683.266366.946535@debian.local> X-Priority: 3 Importance: Normal Cc: MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: 8bit X-SW-Source: 2003-q4/txt/msg00071.txt.bz2 I'll do some work on digging up refs. Also, the code I provided is buggy with rank calculation. BTW, yesterday I sat down read the GSL design document in entirety. Everyone working on scientific libraries should read it. -R > rlee@fpcc.net writes: > > I posted "Help developing a "better" QRPT routine". I guess > sometimes we help those who help themselves ;-) > > > > Here GSL code for rank-revealing QRPT. It is just a minor tweak of > gsl_linalg_QRPT_decomp -- the key feature is a different pivoting > strategy. > > > > I would appreciate any suggestions on making this more "GSL like". > I've only tested it with some M>N matrices -- some rank deficient. > > > > So, does anyone think this could be a valuable addition to GSL? > > Hi, > > If there is an improved algorithm I would be in favor of using it. > > Is there a paper or other reference (maybe a Lapack Working Note) > which describes this algorithm? The first step is to find out what has > been written about it. > > -- > Brian Gough