From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 6932 invoked by alias); 14 Apr 2011 20:13:06 -0000 Received: (qmail 6909 invoked by uid 22791); 14 Apr 2011 20:13:02 -0000 X-SWARE-Spam-Status: No, hits=-1.8 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_NONE X-Spam-Check-By: sourceware.org Received: from mx01.qsc.de (HELO mx01.qsc.de) (213.148.129.14) by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 14 Apr 2011 20:12:38 +0000 Received: from [192.168.178.22] (port-92-204-66-174.dynamic.qsc.de [92.204.66.174]) by mx01.qsc.de (Postfix) with ESMTP id 55F563CC4A; Thu, 14 Apr 2011 22:12:37 +0200 (CEST) Message-ID: <4DA75534.4000706@net-b.de> Date: Thu, 14 Apr 2011 20:30:00 -0000 From: Tobias Burnus User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; de; rv:1.9.2.13) Gecko/20101206 SUSE/3.1.7 Thunderbird/3.1.7 MIME-Version: 1.0 To: Michael Matz CC: fortran@gcc.gnu.org, gcc-patches@gcc.gnu.org Subject: Re: Implement stack arrays even for unknown sizes References: <20110409100809.BF9CD3BE18@mailhost.lps.ens.fr> <4DA75102.2080500@net-b.de> In-Reply-To: <4DA75102.2080500@net-b.de> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2011-04/txt/msg01134.txt.bz2 Tobias Burnus wrote: > no stack-arrays with stack-arrays > + -fwhole-program -flto: 10.1s 8.9s > + -fwhole-program -flto -finline-limit=600 4.8s 3.6s I wonder whether the following is special to my system* or generally true. I use: gfortran -O3 -march=native -ffast-math -funroll-loops -fwhole-program -finline-limit=600 fatigue.f90 no stack-arrays with stack-arrays 0m6.622s 0m8.174s -flto 0m8.444s 0m8.174s Thus, the non "-flto" version is faster (in particular without stack arrays). I assume that it has to do with the declaration issues of the front end. However, the last time I tried to find the problem, I failed to spot anything which looked wrong - especially, the UIDs seemed to be OK. (Besides, I would like to have the -fno-lto performance also with .-flto ... ;-) Tobias * AMD Athlon(tm) 64 X2 Dual Core Processor 4800+ (2.4 GHz), x86-64 Linux