From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 12914 invoked by alias); 4 Dec 2003 00:06:48 -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 12901 invoked from network); 4 Dec 2003 00:06:48 -0000 Received: from unknown (HELO fencepost.gnu.org) (199.232.76.164) by sources.redhat.com with SMTP; 4 Dec 2003 00:06:48 -0000 Received: from monty-python.gnu.org ([199.232.76.173]) by fencepost.gnu.org with esmtp (Exim 4.24) id 1ARh0w-000214-3s for gcc@gnu.org; Wed, 03 Dec 2003 19:06:38 -0500 Received: from mail by monty-python.gnu.org with spam-scanned (Exim 4.24) id 1ARhxn-00083u-3X for gcc@gnu.org; Wed, 03 Dec 2003 20:07:59 -0500 Received: from [195.113.31.123] (helo=atrey.karlin.mff.cuni.cz) by monty-python.gnu.org with esmtp (Exim 4.24) id 1ARhxm-00083e-HW for gcc@gnu.org; Wed, 03 Dec 2003 20:07:26 -0500 Received: by atrey.karlin.mff.cuni.cz (Postfix, from userid 4018) id E8C0D4C0323; Thu, 4 Dec 2003 01:05:58 +0100 (CET) Date: Thu, 04 Dec 2003 00:13:00 -0000 From: Jan Hubicka To: "Vladimir N. Makarov" Cc: Ghassan Shobaki , gcc-help@gcc.gnu.org, gcc@gnu.org Subject: Re: Superblock Instruction Scheduling in GCC Message-ID: <20031204000558.GD23084@atrey.karlin.mff.cuni.cz> References: <3FCDEE2C.19E04DE3@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3FCDEE2C.19E04DE3@redhat.com> User-Agent: Mutt/1.5.4i X-Spam-Status: No, hits=-4.5 required=5.0 tests=IN_REP_TO,QUOTED_EMAIL_TEXT,REFERENCES,REPLY_WITH_QUOTES, USER_AGENT_MUTT version=2.55 X-Spam-Level: X-Spam-Checker-Version: SpamAssassin 2.55 (1.174.2.19-2003-05-19-exp) X-SW-Source: 2003-12/txt/msg00289.txt.bz2 > Ghassan Shobaki wrote: > > > I know how to get gcc to form superblocks (by using the -ftracer > > command-line switch), but is there a way to get it to use these > > superblocks as scheduling regions in the instruction scheduling pass? > > Currently, the instruction scheduling module forms regions that are totally > > different from the superblocks that are formed in the tracer module > > even though each superblock is a valid scheduling region. > > Any idea how I can achieve this? Or are there any plans to do superblock > > instruction scheduling in the near future? > > There was Jan Hubicka's patch for this. Please look at it > > http://gcc.gnu.org/ml/gcc-patches/2003-02/msg00499.html > > This patch should work for all platforms except for IA64 whose the second > scheduling is made on EBB. This patch is currently in the mainline tree, so you can simply use -fsched2-use-traces / -fsched2-use-superblocks > > I tried trace scheduling for IA64 (but I did not post the patch for ia64). > Here the results are > > http://gcc.gnu.org/ml/gcc-patches/2003-02/msg00499.html > > The problem with trace scheduling is that the generated code is bigger, the > compiler is slower and the code improvement is insignificant. > > If you manage to achieve an improvement for a platform on a credible > benchmark (SPEC95, SPEC2000), we could consider to add the patch to gcc at > least for given platform for -O3. Because the compiler changed since the > patch was posted, there is a probability that you could achieve this. Yes, we need experimenting here. I was quite surprised that the benefits wasn't too noticeable on in-order architecture and I would like to hear about any results (positive or negative). -fsched2-use-superblocks should bring most of benefits at no code size costs, while -fsched2-use-traces is more experimental and probably needs profile feedback to do somethign usefull. (I managed to get some speedups using this on Athlon but the benefits wasn't considerable enought to discuss inclusion in -O3 -fbranch-probabilities combination) Honza > > > Vlad >