From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 130838 invoked by alias); 20 Sep 2016 02:20:04 -0000 Mailing-List: contact gdb-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-owner@sourceware.org Received: (qmail 130818 invoked by uid 89); 20 Sep 2016 02:20:03 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-0.9 required=5.0 tests=BAYES_00,KAM_LAZY_DOMAIN_SECURITY autolearn=no version=3.3.2 spammy=H*r:2002, H*r:Debian-4, H*RU:200, Hx-spam-relays-external:200 X-HELO: moltke.seatribe.se Received: from moltke.seatribe.se (HELO moltke.seatribe.se) (178.63.100.209) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Tue, 20 Sep 2016 02:20:01 +0000 Received: from [2002:2f8:a79::1] (helo=nerv.dolda2000.com) by moltke.seatribe.se with esmtp (Exim 4.80) (envelope-from ) id 1bmAf7-0004nF-D1; Tue, 20 Sep 2016 04:19:57 +0200 Received: from [2002:2f8:a79:200:d63d:7eff:feb4:da50] ([IPv6:2002:2f8:a79:200:d63d:7eff:feb4:da50]) (authenticated bits=0) by nerv.dolda2000.com (8.14.4/8.14.4/Debian-4+deb7u1) with ESMTP id u8K2JsWa026172 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Tue, 20 Sep 2016 04:19:55 +0200 Date: Tue, 20 Sep 2016 02:20:00 -0000 From: Fredrik Tolf To: Pedro Alves , gdb@sourceware.org Subject: Re: Bad performance in updating JIT debug symbols In-Reply-To: Message-ID: References: User-Agent: Alpine 2.11 (DEB 23 2013-08-11) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed X-SW-Source: 2016-09/txt/msg00061.txt.bz2 On Mon, 19 Sep 2016, Pedro Alves wrote: > I've now rebased and force-pushed part of what I had to the > users/palves/jit-speedup branch. It's regression-free for me > on x86_64 Fedora 23. I'm also attaching qsort patch below for > reference. I tested the branch, and can verify that it seems to be working really well for me. I no longer any noticable slowdown at compilations whatsoever, so that's really great for me. I'll probably be running this until the code gets its way into release, so thanks a lot! I haven't tested it a whole lot yet, but I too haven't noticed any regressions thus far. > If you don't have any breakpoint set, then with that branch, > JIT debugging is snappy. The problem is that the next bottleneck > triggers as soon as you have some breakpoint set. For the record, I'm not noticing this. I tried setting a breakpoint, and there's still no problem whatsoever when new code is JITted. -- Fredrik Tolf