From owner-freebsd-current Sun Mar 23 9:55:19 2003 Delivered-To: freebsd-current@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 2A11D37B405 for ; Sun, 23 Mar 2003 09:55:18 -0800 (PST) Received: from grassmarket.ucs.ed.ac.uk (grassmarket.ucs.ed.ac.uk [129.215.166.64]) by mx1.FreeBSD.org (Postfix) with ESMTP id EEC5B43F75 for ; Sun, 23 Mar 2003 09:55:15 -0800 (PST) (envelope-from s9905155@sms.ed.ac.uk) Received: from fourtytwo.brucec.backnet (12266209.resnet.ed.ac.uk [10.6.0.100] (may be forged)) by grassmarket.ucs.ed.ac.uk (8.11.6/8.11.6) with ESMTP id h2NHt0T11643; Sun, 23 Mar 2003 17:55:00 GMT Received: from fourtytwo.brucec.backnet (localhost.brucec.backnet [127.0.0.1]) by fourtytwo.brucec.backnet (8.12.8/8.12.8) with ESMTP id h2NHuxDb002628; Sun, 23 Mar 2003 17:57:00 GMT (envelope-from bruce@fourtytwo.brucec.backnet) Received: (from bruce@localhost) by fourtytwo.brucec.backnet (8.12.8/8.12.8/Submit) id h2NHux5C002627; Sun, 23 Mar 2003 17:56:59 GMT Date: Sun, 23 Mar 2003 17:56:59 +0000 From: Bruce Cran To: Michael Nottebrock Cc: current@freebsd.org Subject: Re: libm problem Message-ID: <20030323175659.GA2558@fourtytwo.brucec.backnet> References: <20030318173051.GA2322@f111.hadiko.de> <20030323070808.0d3f63c0.fearow@attbi.com> <20030323170244.GA1911@f111.hadiko.de> <200303231843.16545.michaelnottebrock@gmx.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <200303231843.16545.michaelnottebrock@gmx.net> User-Agent: Mutt/1.5.4i Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG On Sun, Mar 23, 2003 at 06:43:16PM +0100, Michael Nottebrock wrote: > > Just out of curiousity, have you tried using -mfpmath=sse? I remember someone > on this list claiming that the SSE fpa-code works much better than the i387 > code which is used by default (even with -march=pentium4). > I don't know if SSE code runs faster, but it defitely seems to fix bad code generation when optimizing for athlon-xp architecture. For example, -O2 -march=athlon-xp produces broken code, whereas -O2 -march=athlon-xp -mfpmath=sse produces programs which work. Bruce Cran To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message