From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 4491 invoked by alias); 28 Jul 2008 10:57:37 -0000 Received: (qmail 4483 invoked by uid 22791); 28 Jul 2008 10:57:36 -0000 X-Spam-Check-By: sourceware.org Received: from smtp.fullrate.dk (HELO dns2.fullrate.dk) (89.150.129.5) by sourceware.org (qpsmtpd/0.31) with ESMTP; Mon, 28 Jul 2008 10:57:11 +0000 Received: from [192.168.1.33] (3604ds3-fb.0.fullrate.dk [90.184.27.253]) by dns2.fullrate.dk (Postfix) with ESMTP id 19A3A5CC99; Mon, 28 Jul 2008 12:57:07 +0200 (CEST) Message-ID: <488DA5F9.5050203@agner.org> Date: Mon, 28 Jul 2008 11:45:00 -0000 From: Agner Fog User-Agent: Thunderbird 2.0.0.16 (Windows/20080708) MIME-Version: 1.0 To: Michael Meissner , Agner Fog , Raksit Ashok , dclarke@opensolaris.org, gcc@gcc.gnu.org, TimothyPrince@sbcglobal.net, Tarjei Knapstad Subject: Re: gcc will become the best optimizing x86 compiler References: <2E073B3ABB3F664DBA1D1C4D5FB47EF40EBDAD8E@NT-IRVA-0752.brcm.ad.broadcom.com> <4887592E.4040804@agner.org> <4888375A.30601@agner.org> <9718380a0807241009m563e8da1x2983145a93e3ce7b@mail.gmail.com> <48897BFA.8000008@agner.org> <20080725220958.GA4900@tiktok.the-meissners.org> In-Reply-To: <20080725220958.GA4900@tiktok.the-meissners.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-IsSubscribed: yes Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-owner@gcc.gnu.org X-SW-Source: 2008-07/txt/msg00569.txt.bz2 Michael Meissner wrote: >Memcpy/memset optimizations were added to glibc 2.8, though when your favorite >distribution will provide it is a different question: >http://sourceware.org/ml/libc-alpha/2008-04/msg00050.html I finally got a SUSE with glibc 2.8. I can see that 32-bit memcpy has been modified with an extra misalignment branch, but no significant improvement. Glibc 2.8 is NOT faster than glibc 2.7 in my tests. It still doesn't use XMM registers. Glibc 2.8 is still almost 5 times slower than the best function libraries for unaligned data on Intel Core 2, and the default builtin function is slower than any other implementation I have seen (copies 1 byte at a time!). Tarjei Knapstad wrote: >2008/7/26 Agner Fog : >>I have libc version 2.7. Can't find version 2.8 >It's in Fedora 9, I have no idea why the source isn't directly >available from the glibc homepage. 2.8 is not an official final release yet.