From owner-freebsd-current@FreeBSD.ORG Sat May 26 16:46:17 2007 Return-Path: X-Original-To: freebsd-current@freebsd.org Delivered-To: freebsd-current@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id D6FC816A400 for ; Sat, 26 May 2007 16:46:17 +0000 (UTC) (envelope-from sgk@troutmask.apl.washington.edu) Received: from troutmask.apl.washington.edu (troutmask.apl.washington.edu [128.208.78.105]) by mx1.freebsd.org (Postfix) with ESMTP id B9AF713C455 for ; Sat, 26 May 2007 16:46:17 +0000 (UTC) (envelope-from sgk@troutmask.apl.washington.edu) Received: from troutmask.apl.washington.edu (localhost.apl.washington.edu [127.0.0.1]) by troutmask.apl.washington.edu (8.14.1/8.13.8) with ESMTP id l4QGijGa053628; Sat, 26 May 2007 09:44:45 -0700 (PDT) (envelope-from sgk@troutmask.apl.washington.edu) Received: (from sgk@localhost) by troutmask.apl.washington.edu (8.14.1/8.13.8/Submit) id l4QGij00053627; Sat, 26 May 2007 09:44:45 -0700 (PDT) (envelope-from sgk) Date: Sat, 26 May 2007 09:44:45 -0700 From: Steve Kargl To: Michiel Boland Message-ID: <20070526164445.GA53570@troutmask.apl.washington.edu> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.4.2.2i Cc: freebsd-current@freebsd.org, Claus Guttesen Subject: Re: gcc memory consumption: amd64 v i386 X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Sat, 26 May 2007 16:46:17 -0000 On Sat, May 26, 2007 at 04:19:38PM +0200, Michiel Boland wrote: > >>Hi. I noticed that compilation of xorg-server on i386 with the new gcc > >>proceeds normally, whereas compilation on amd64 would crash because the > >>compiler would consume all memory. The i386 and amd64 boxen have the same > >>amount of RAM and swap, obviously. And they run, give or take a few hours, > >>more or less same version of -CURRENT. > > > >It does not crash if you have enough swap. I have 2 GB swap and it > >proceeded fine after some swapping. > > The point I was trying to make (although perhaps not clearly enough) is > that there is no reason that a trivial source file takes up such a huge > amount of memory to compile. Especially since gcc 3.4.6 does not blow up > like that. Major portions of the middle and back end of gcc were rewritten in going from 3.4.6 to 4.x. The TREE-SSA representation supposely offers the possibility of better optimization passes, but it comes at the moment with some memory pressure. The GCC developers are accutely aware of this issue, and are working on the problem. -- Steve