From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 5713 invoked by alias); 21 Mar 2003 00:01:49 -0000 Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org Received: (qmail 5706 invoked from network); 21 Mar 2003 00:01:49 -0000 Received: from unknown (HELO hiauly1.hia.nrc.ca) (132.246.100.193) by sources.redhat.com with SMTP; 21 Mar 2003 00:01:49 -0000 Received: from hiauly1.hia.nrc.ca (localhost [127.0.0.1]) by hiauly1.hia.nrc.ca (8.12.8/8.12.8) with ESMTP id h2L01ksQ009341; Thu, 20 Mar 2003 19:01:47 -0500 (EST) Received: (from dave@localhost) by hiauly1.hia.nrc.ca (8.12.8/8.12.8/Submit) id h2L01ka9009337; Thu, 20 Mar 2003 19:01:46 -0500 (EST) Message-Id: <200303210001.h2L01ka9009337@hiauly1.hia.nrc.ca> Subject: Re: [PATCH]: PA long unconditional branch generation (PR10062) To: dave@hiauly1.hia.nrc.ca (John David Anglin) Date: Fri, 21 Mar 2003 00:02:00 -0000 From: "John David Anglin" Cc: mark@codesourcery.com, gcc-patches@gcc.gnu.org, china@thewrittenword.com In-Reply-To: from "John David Anglin" at Mar 20, 2003 06:49:58 pm MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-SW-Source: 2003-03/txt/msg01853.txt.bz2 > Without inlining, the length of the compare chain is 29. The default > inlining requires in excess of 768MB of virtual memory to compile > the testcase in the PR. Clearly, the default inlining parameters > need to be tweaked to keep the compilation with reasonable bounds > but I'm not sure how to do this. Possibly, the throttling algorithm > is broken. It would also be nice to use a jump table. I should note that once the size of a function exceeds ~240KB we need to use long calls and branches. These sequences are substantially longer and less efficient. So, we probably lose more than we gain at that point. Dave -- J. David Anglin dave.anglin@nrc-cnrc.gc.ca National Research Council of Canada (613) 990-0752 (FAX: 952-6602)