From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 20238 invoked by alias); 22 Apr 2003 15:19:33 -0000 Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-owner@gcc.gnu.org Received: (qmail 20230 invoked from network); 22 Apr 2003 15:19:32 -0000 Received: from unknown (HELO rose.csi.cam.ac.uk) (131.111.8.13) by sources.redhat.com with SMTP; 22 Apr 2003 15:19:32 -0000 Received: from cass41.ast.cam.ac.uk ([131.111.69.186]) by rose.csi.cam.ac.uk with esmtp (Exim 4.12) id 197zYR-0002hH-00; Tue, 22 Apr 2003 16:19:31 +0100 Received: from xserv1.ast.cam.ac.uk (IDENT:ZuRPyAViSY3tp3o38cMhQFUu13wRptiM@xserv1.ast.cam.ac.uk [131.111.69.235]) by cass41.ast.cam.ac.uk (8.12.9+Sun/8.12.9) with ESMTP id h3MFJTgD021450; Tue, 22 Apr 2003 16:19:29 +0100 (BST) Received: from xpc5.ast.cam.ac.uk (IDENT:l1eKU4F3VLnRGmc7GFzSB4jtv9O7NHqh@xpc5.ast.cam.ac.uk [131.111.68.220]) by xserv1.ast.cam.ac.uk (8.11.6/8.11.6) with ESMTP id h3MFJTl19933; Tue, 22 Apr 2003 16:19:29 +0100 Date: Tue, 22 Apr 2003 16:11:00 -0000 From: Jeremy Sanders To: Daniel Berlin cc: "S. Bosscher" , "'gcc@gcc.gnu.org '" Subject: Re: benchmarking (or almabench) In-Reply-To: Message-ID: References: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-SW-Source: 2003-04/txt/msg01064.txt.bz2 On Tue, 22 Apr 2003, Daniel Berlin wrote: > On Tuesday, April 22, 2003, at 11:08 AM, S. Bosscher wrote: > > > -march=pentium4 is known to pessimise code compared to -march=i686 for > > some > > benchmarks, see PR 8474. Maybe you're seeing the same problem? > > Actually, if i had to guess, i'd put my money on the vectorization. > Notice ICC vectorized two loops in his example, and obviously, we > vectorized 0. > :) The intel compiler doesn't seem to vectorize with just "-O2" (by default it should report whether it is using vectorization), and that's still 88% faster than gcc. I can't absolutely confirm there's no vectorization as I can't see a switch to turn it off. icc says it's vectorizing when the P4 specific options are enabled (which gcc can't do yet). If I turn off any optimization on icc, then it's still faster than gcc!!! xpc5:/<3>almabench-1.0.1/cpp> make icc -o almabench.o -O0 -c almabench.cpp icc -o almabench -O0 almabench.o xpc5:/<3>almabench-1.0.1/cpp> time ./almabench 23.853u 0.134s 0:25.82 92.8% 0+0k 0+0io 121pf+0w Jeremy -- Jeremy Sanders http://www-xray.ast.cam.ac.uk/~jss/ X-Ray Group, Institute of Astronomy, University of Cambridge, UK. Public Key Server PGP Key ID: E1AAE053