From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 31361 invoked by alias); 5 Jun 2008 23:38:09 -0000 Received: (qmail 31350 invoked by uid 22791); 5 Jun 2008 23:38:08 -0000 X-Spam-Check-By: sourceware.org Received: from rv-out-0708.google.com (HELO rv-out-0708.google.com) (209.85.198.243) by sourceware.org (qpsmtpd/0.31) with ESMTP; Thu, 05 Jun 2008 23:37:40 +0000 Received: by rv-out-0708.google.com with SMTP id c5so864007rvf.56 for ; Thu, 05 Jun 2008 16:37:38 -0700 (PDT) Received: by 10.141.15.19 with SMTP id s19mr1308121rvi.75.1212709058145; Thu, 05 Jun 2008 16:37:38 -0700 (PDT) Received: by 10.140.128.6 with HTTP; Thu, 5 Jun 2008 16:37:38 -0700 (PDT) Message-ID: <571f6b510806051637k765e7d65k7466381ee18e0350@mail.gmail.com> Date: Thu, 05 Jun 2008 23:38:00 -0000 From: "Steven Bosscher" To: "Andrey Belevantsev" Subject: Re: [RFC] Selective scheduling pass Cc: gcc-patches , "Jim Wilson" , "Vladimir Makarov" In-Reply-To: <48483071.4010906@ispras.ru> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <571f6b510806042343y599633e0kd2ef40e9a649e924@mail.gmail.com> <48483071.4010906@ispras.ru> Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm Precedence: bulk List-Id: List-Archive: List-Post: List-Help: Sender: gcc-patches-owner@gcc.gnu.org X-SW-Source: 2008-06/txt/msg00282.txt.bz2 On Thu, Jun 5, 2008 at 8:29 PM, Andrey Belevantsev wrote: >> Also, you post a new scheduler and a set of target tunings in one set >> of patches. I would like to know what the performance impact is of >> just the target changes alone. That is, what happens to e.g. SPEC >> scores for ia64 with just the tweaks and tunings patch >> (http://gcc.gnu.org/ml/gcc-patches/2008-06/msg00117.html)? I assume >> those target changes alone (the ones not related to sel-sched) also >> have a positive performance impact. Since you've globbed everything >> into one patch set, it's impossible to tell how much of the >> performance changes can be attributed to sel-sched, and how much is >> just target tweaks... > > AFAIR, the target tunings gave around 1% when we have tested it on -O2. > That was a couple of months ago. We will retest tonight to get fresh > numbers for this. Great. Another thought/suggestion/comment: Does the selective scheduler also make the register renaming pass obsolete (pass_regrename and pass_cprop_hardreg)? I would expect it does, since the scheduler handles the register renaming itself, as far as I understand. These two passes are quite expensive (or at least they were when I last looked at them). You could perhaps buy yourself some compile time back if you can demonstrate you don't need to run these passes if you do post-regalloc selective scheduling. Gr. Steven