From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 14483 invoked by alias); 7 Sep 2013 11:55:00 -0000 Mailing-List: contact libc-ports-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Post: List-Help: , Sender: libc-ports-owner@sourceware.org Received: (qmail 14472 invoked by uid 89); 7 Sep 2013 11:55:00 -0000 Received: from popelka.ms.mff.cuni.cz (HELO popelka.ms.mff.cuni.cz) (195.113.20.131) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Sat, 07 Sep 2013 11:55:00 +0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-0.7 required=5.0 tests=AWL,BAYES_00,FREEMAIL_FROM,SPF_NEUTRAL autolearn=no version=3.3.2 X-HELO: popelka.ms.mff.cuni.cz Received: from domone.kolej.mff.cuni.cz (popelka.ms.mff.cuni.cz [195.113.20.131]) by popelka.ms.mff.cuni.cz (Postfix) with ESMTPS id 724816A1A5; Sat, 7 Sep 2013 13:54:36 +0200 (CEST) Received: by domone.kolej.mff.cuni.cz (Postfix, from userid 1000) id 4C08B5F822; Sat, 7 Sep 2013 13:54:36 +0200 (CEST) Date: Sat, 07 Sep 2013 11:55:00 -0000 From: =?utf-8?B?T25kxZllaiBCw61sa2E=?= To: "Ryan S. Arnold" Cc: Carlos O'Donell , Will Newton , "libc-ports@sourceware.org" , Patch Tracking , Siddhesh Poyarekar Subject: Re: [PATCH] sysdeps/arm/armv7/multiarch/memcpy_impl.S: Improve performance. Message-ID: <20130907115434.GA13391@domone.kolej.mff.cuni.cz> References: <520894D5.7060207@linaro.org> <5220D30B.9080306@redhat.com> <5220F1F0.80501@redhat.com> <52260BD0.6090805@redhat.com> <20130903173710.GA2028@domone.kolej.mff.cuni.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: User-Agent: Mutt/1.5.20 (2009-06-14) X-IsSubscribed: yes X-SW-Source: 2013-09/txt/msg00068.txt.bz2 On Tue, Sep 03, 2013 at 02:34:54PM -0500, Ryan S. Arnold wrote: > On Tue, Sep 3, 2013 at 12:37 PM, Ondřej Bílka wrote: > >> I disagree strongly. You *must* come up with a measurable answer and > >> looking at a graph is never a solution I'm going to accept. > >> > > You can have that opinion. > > Looking at performance graphs is most powerful technique how to > > understand performance. I got most of my improvements from analyzing > > these. > > Are there any open source pipeline analysis tools? I've found the one > I've used (proprietary) to be a pretty good indicator of general > instruction selection optimization/correctness. > After bit of googling I found http://marss86.org/~marss86/index.php/Home Is anyone familiar with it? This reminded me to work on project in my backlog. Writing tool that can legaly shuffle assembler instructions. Then I run simulated annaling to get optimal sheduling for given processor and benchmark. With dryrun data this will be close to real performance on small size so I could get some percents there.