From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 4859 invoked by alias); 28 Feb 2012 16:36:35 -0000 Received: (qmail 4569 invoked by uid 22791); 28 Feb 2012 16:36:32 -0000 X-SWARE-Spam-Status: No, hits=-2.8 required=5.0 tests=ALL_TRUSTED,AWL,BAYES_00 X-Spam-Check-By: sourceware.org Received: from localhost (HELO sourceware.org) (127.0.0.1) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 28 Feb 2012 16:36:18 +0000 From: "vincent-srcware at vinc17 dot net" To: glibc-bugs@sources.redhat.com Subject: [Bug math/706] pow() produces inaccurate results for base ~ 1.0, and large integer exponent on 32-bit x86 Date: Tue, 28 Feb 2012 16:36:00 -0000 X-Bugzilla-Reason: CC X-Bugzilla-Type: changed X-Bugzilla-Watch-Reason: None X-Bugzilla-Product: glibc X-Bugzilla-Component: math X-Bugzilla-Keywords: X-Bugzilla-Severity: normal X-Bugzilla-Who: vincent-srcware at vinc17 dot net X-Bugzilla-Status: NEW X-Bugzilla-Priority: P2 X-Bugzilla-Assigned-To: unassigned at sourceware dot org X-Bugzilla-Target-Milestone: --- X-Bugzilla-Changed-Fields: Summary Message-ID: In-Reply-To: References: X-Bugzilla-URL: http://sourceware.org/bugzilla/ Auto-Submitted: auto-generated Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Mailing-List: contact glibc-bugs-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Post: List-Help: , Sender: glibc-bugs-owner@sourceware.org X-SW-Source: 2012-02/txt/msg00532.txt.bz2 http://sourceware.org/bugzilla/show_bug.cgi?id=3D706 Vincent Lef=C3=A8vre changed: What |Removed |Added ---------------------------------------------------------------------------- Summary|pow() produces inaccurate |pow() produces inaccurate |results for base ~ 1.0, and |results for base ~ 1.0, and |large exponent on 32-bit |large integer exponent on |x86 |32-bit x86 --- Comment #14 from Vincent Lef=C3=A8vre 2012-02-28 16:36:13 UTC --- I've done some tests, and it seems that pow(x,y) for large y is inaccurate = only when y is an integer. So, if I understand correctly, an iterative algorithm= is used for integers y, and a log-exp algorithm (which seems accurate enough on the few values I've tested) is used for other values. So, a possible easy f= ix would be to always use the log-exp algorithm, possibly except for small integers y. Note: I've updated the bug summary following this observation (large expone= nt =E2=86=92 large integer exponent). --=20 Configure bugmail: http://sourceware.org/bugzilla/userprefs.cgi?tab=3Demail ------- You are receiving this mail because: ------- You are on the CC list for the bug.