From owner-freebsd-current@FreeBSD.ORG Thu Sep 6 12:57:19 2012 Return-Path: Delivered-To: freebsd-current@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 5DBDB106566B; Thu, 6 Sep 2012 12:57:19 +0000 (UTC) (envelope-from dimitry@andric.com) Received: from tensor.andric.com (cl-327.ede-01.nl.sixxs.net [IPv6:2001:7b8:2ff:146::2]) by mx1.freebsd.org (Postfix) with ESMTP id 110388FC08; Thu, 6 Sep 2012 12:57:19 +0000 (UTC) Received: from [IPv6:2001:7b8:3a7:0:f8a2:c245:f52d:c435] (unknown [IPv6:2001:7b8:3a7:0:f8a2:c245:f52d:c435]) (using TLSv1 with cipher DHE-RSA-CAMELLIA256-SHA (256/256 bits)) (No client certificate requested) by tensor.andric.com (Postfix) with ESMTPSA id 541A65C37; Thu, 6 Sep 2012 14:57:18 +0200 (CEST) Message-ID: <50489DB1.4070102@andric.com> Date: Thu, 06 Sep 2012 14:57:21 +0200 From: Dimitry Andric User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:16.0) Gecko/20120828 Thunderbird/16.0 MIME-Version: 1.0 To: David Chisnall References: <5046670C.6050500@andric.com> <20120904214344.GA17723@troutmask.apl.washington.edu> <504679CB.90204@andric.com> <20120904221413.GA19395@troutmask.apl.washington.edu> <50471BEE.6030708@andric.com> <20120905221310.GA97847@troutmask.apl.washington.edu> <20120906084312.GA13223@freebsd.org> <122361C5-7B83-494B-A580-970FAFF83CAA@FreeBSD.org> In-Reply-To: <122361C5-7B83-494B-A580-970FAFF83CAA@FreeBSD.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: Garrett Cooper , Roman Divacky , freebsd-current@FreeBSD.org, Steve Kargl Subject: Re: Compiler performance tests on FreeBSD 10.0-CURRENT X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 06 Sep 2012 12:57:19 -0000 On 2012-09-06 12:20, David Chisnall wrote: ... > There may also be a difference in whether -ffast-math is the default on each compiler. On x86, this will replace a number of libm calls with (much faster, but less accurate) SSE or x87 instructions. If this is enabled by default with clang and not with gcc, it would account for the difference. No, -ffast-math is not enabled by default in clang, as far as I can tell. Also, the help text for the option says: "Enable the *frontend*'s 'fast-math' mode. This has no effect on optimizations, but provides a preprocessor macro __FAST_MATH__ the same as GCC's -ffast-math flag."