From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 3836 invoked by alias); 3 Jan 2008 09:01:27 -0000 Received: (qmail 3819 invoked by uid 22791); 3 Jan 2008 09:01:26 -0000 X-Spam-Check-By: sourceware.org Received: from divok.RZ-Berlin.MPG.DE (HELO divok.rz-berlin.mpg.de) (141.14.131.15) by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 03 Jan 2008 09:00:58 +0000 Received: from divok.rz-berlin.mpg.de (localhost [127.0.0.1]) by localhost (Postfix) with SMTP id 8175A2A103 for ; Thu, 3 Jan 2008 10:00:54 +0100 (CET) Received: from koelsch.rz-berlin.mpg.de (koelsch.rz-berlin.mpg.de [141.14.142.142]) (using TLSv1 with cipher AES128-SHA (128/128 bits)) (No client certificate requested) by divok.rz-berlin.mpg.de (Postfix) with ESMTP id 67F2F2A102 for ; Thu, 3 Jan 2008 10:00:54 +0100 (CET) Message-Id: <9AFC9075-1EA3-4E26-9BE9-EB9C475FE3D9@fhi-berlin.mpg.de> From: =?ISO-8859-1?Q?Jochen_K=FCpper?= To: gsl-discuss@sourceware.org In-Reply-To: <477BDBEE.2000603@googlemail.com> Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg=pgp-sha1; boundary="Apple-Mail-5--147161078" Content-Transfer-Encoding: 7bit Mime-Version: 1.0 (Apple Message framework v915) Subject: Unwanted compiler optimization (was: Bug report and fix for adaptive step size control in ODE solving) Date: Thu, 03 Jan 2008 09:01:00 -0000 References: <4772D239.9070206@googlemail.com> <0E55B479-624E-4F88-B511-CDFF93435D9E@fhi-berlin.mpg.de> <477BDBEE.2000603@googlemail.com> X-Pgp-Agent: GPGMail d51 (Leopard) X-Mailer: Apple Mail (2.915) 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: 2008-q1/txt/msg00004.txt.bz2 This is an OpenPGP/MIME signed message (RFC 2440 and 3156) --Apple-Mail-5--147161078 Content-Type: text/plain; charset=ISO-8859-1; format=flowed; delsp=yes Content-Transfer-Encoding: quoted-printable Content-length: 985 Dear Optimization specialists, On 02.01.2008, at 19:46, Frank Reininghaus wrote: > Jochen K=FCpper wrote: > >> However, why don't you test what you really describe your comment?=20=20 >> This would look like the following (untested): >> >> h0 =3D GSL_MAX_DBL(GSL_DBL_MIN, (t0 !=3D t0+h0) ? h0 :=20=20 >> GSL_DBL_EPSILON*t0); > > Actually, I had been thinking of something like that first, but then=20=20 > I was afraid that an optimising compiler might transform 't0 !=3D=20=20 > t0+h0' into '0 !=3D h0', making the patch useless for the 'h0 <=20=20 > GSL_DBL_EPSILON*t0' problem. Is this an optimization that is considered valid even for floating-=20 point calculations in compiler building? Maybe we should ask our friends from the GCC development about it... Greetings, Jochen --=20 Einigkeit und Recht und Freiheit http://www.Jochen-Kuepper.de Libert=E9, =C9galit=E9, Fraternit=E9 GnuPG key: CC1B0B4D Sex, drugs and rock-n-roll --Apple-Mail-5--147161078 content-type: application/pgp-signature; x-mac-type=70674453; name=PGP.sig content-description: This is a digitally signed message part content-disposition: inline; filename=PGP.sig content-transfer-encoding: 7bit Content-length: 186 -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (Darwin) iD8DBQFHfKRG7eYXdswbC00RAvCrAJ91FeXGSifZuSSF/gAAqxKPvJYbcQCfR+x3 +8ZQ2YynDHaOcWO70mDcX9A= =m2x2 -----END PGP SIGNATURE----- --Apple-Mail-5--147161078--