From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 11825 invoked by alias); 2 May 2004 16:56:30 -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 11817 invoked from network); 2 May 2004 16:56:30 -0000 Received: from unknown (HELO dberlin.org) (69.3.5.6) by sources.redhat.com with SMTP; 2 May 2004 16:56:30 -0000 Received: from [192.168.1.7] (account dberlin HELO [192.168.1.7]) by dberlin.org (CommuniGate Pro SMTP 4.1.4) with ESMTP-TLS id 6680641; Sun, 02 May 2004 12:56:29 -0400 In-Reply-To: <20040502132733.GA24763@cse.unsw.EDU.AU> References: <20040502132733.GA24763@cse.unsw.EDU.AU> Mime-Version: 1.0 (Apple Message framework v617) Content-Type: text/plain; charset=US-ASCII; format=flowed Message-Id: Content-Transfer-Encoding: 7bit Cc: gcc@gcc.gnu.org From: Daniel Berlin Subject: Re: register allocation Date: Sun, 02 May 2004 16:56:00 -0000 To: qiongc@cse.unsw.edu.au (Qiong Cai) X-SW-Source: 2004-05/txt/msg00032.txt.bz2 On May 2, 2004, at 9:27 AM, Qiong Cai wrote: > Hi, > > I'm going to study register allocation in GCC. What do you mean "study"? To be honest, we already know a lot about GCC register allocation, it's shortcomings, etc. This is true for both allocators. > Currently GCC has > two register allocators. Here some questions: > * Is there any performance results(eg. spec2k results) available for > these two allocators? Yes, for various platforms. Check the mailing list archives. > * Besides those ra*.[c,h] files, is tthere any other source files > related > to register allocation? local-alloc, global-alloc (the "old" allocator) > * Which function, if available, calcuate the register pressure, > which is the defined as the max number of live variable for a basic > block? > None. > Many thanks. > > Qiong