From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 23971 invoked by alias); 21 Oct 2014 03:21:17 -0000 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 Received: (qmail 23960 invoked by uid 89); 21 Oct 2014 03:21:15 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-2.4 required=5.0 tests=AWL,BAYES_00,RCVD_IN_DNSWL_LOW,SPF_PASS autolearn=ham version=3.3.2 X-HELO: mail-pa0-f42.google.com Received: from mail-pa0-f42.google.com (HELO mail-pa0-f42.google.com) (209.85.220.42) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with (AES128-SHA encrypted) ESMTPS; Tue, 21 Oct 2014 03:21:14 +0000 Received: by mail-pa0-f42.google.com with SMTP id bj1so418776pad.29 for ; Mon, 20 Oct 2014 20:21:12 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:from:content-type:subject:date:message-id:cc:to :mime-version; bh=SroRuBWya4E36g7s8D0YNj8+q187zFwKVU8Z/pF4/Jw=; b=GVvivqrPuXIquAS6N3RoJ8dDoxlMOLUu81p/l2zuOABygXpVfsOnhl+DkmMgsPe0vP MEQqsEExTxKyzzTbYOu7RkUGBo4UBnRefrcIUgnjujEbkutnmzc/sWCG7HofMBaeV5ml +qab4bPAYmFuxL6zm+xkeuIbiD0JSwwXgFjcwhlck59/ktPucldhkWCXImA+b2VbfNuq Vl+wMcx1LjkT+T2916fNBm43M3DrGmfkLIFbHALnTjgIVhJKLKZou5r6J75YUSLq1DBq obnXNLl0dce6kO7Kkm4rByk6DhxgbL3bsCm8NUo0H2d2+lvF593qCbOPQ/5jEDzhmpjR a+dw== X-Gm-Message-State: ALoCoQkQewBxjHuKyEwsmL7QPWlZlKxDjdwFw17oUAgtVcL17SpdEXp7bNuNqLf2gPM8iguFqsoW X-Received: by 10.70.60.167 with SMTP id i7mr31780878pdr.86.1413861672437; Mon, 20 Oct 2014 20:21:12 -0700 (PDT) Received: from [192.168.0.122] (121-99-179-86.bng1.nct.orcon.net.nz. [121.99.179.86]) by mx.google.com with ESMTPSA id zf5sm10375098pbc.44.2014.10.20.20.21.09 for (version=TLSv1 cipher=ECDHE-RSA-RC4-SHA bits=128/128); Mon, 20 Oct 2014 20:21:11 -0700 (PDT) From: Maxim Kuvyrkov Content-Type: multipart/mixed; boundary="Apple-Mail=_C9620461-374F-42C3-A9B6-9171518BD05C" Subject: [PATCH 4/8] Disable max_issue when scheduling for register pressure Date: Tue, 21 Oct 2014 03:31:00 -0000 Message-Id: Cc: Vladimir Makarov , Jeff Law , ramrad01@arm.com To: GCC Patches Mime-Version: 1.0 (Mac OS X Mail 7.3 \(1878.6\)) X-SW-Source: 2014-10/txt/msg01999.txt.bz2 --Apple-Mail=_C9620461-374F-42C3-A9B6-9171518BD05C Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset=us-ascii Content-length: 627 Hi, This patch disables max_issue-based lookahead multipass scheduling when reg= ister-pressure heuristics are active. The two approaches tend to undo each= others decisions and don't play well together. Currently this patch is a no-op, since no target uses both max_issue and re= gister-pressure scheduling. This will change soon with upcoming patches th= at will enable max_issue during 2nd scheduling pass for ARM and AArch64. Bootstrapped on x86_64-linux-gnu. Regression testing is in progress. OK t= o commit if no regressions? [ChangeLog is part of the git patch] Thank you, -- Maxim Kuvyrkov www.linaro.org --Apple-Mail=_C9620461-374F-42C3-A9B6-9171518BD05C Content-Disposition: attachment; filename=0004-Disable-max_issue-when-scheduling-for-register-press.patch Content-Type: application/octet-stream; name="0004-Disable-max_issue-when-scheduling-for-register-press.patch" Content-Transfer-Encoding: quoted-printable Content-length: 1182 >From caae8fd31b92ea63d91604df4d71a7866303ec2a Mon Sep 17 00:00:00 2001=0A= From: Maxim Kuvyrkov =0A= Date: Tue, 21 Oct 2014 00:25:11 +0100=0A= Subject: [PATCH 4/8] Disable max_issue when scheduling for register pressur= e=0A= =0A= * haifa-sched.c (sched_init): Disable max_issue when scheduling for=0A= register pressure.=0A= ---=0A= gcc/haifa-sched.c | 5 ++++-=0A= 1 file changed, 4 insertions(+), 1 deletion(-)=0A= =0A= diff --git a/gcc/haifa-sched.c b/gcc/haifa-sched.c=0A= index 4f9648c..26d9e29 100644=0A= --- a/gcc/haifa-sched.c=0A= +++ b/gcc/haifa-sched.c=0A= @@ -6867,7 +6867,10 @@ sched_init (void)=0A= else=0A= issue_rate =3D 1;=0A= =20=0A= - if (targetm.sched.first_cycle_multipass_dfa_lookahead)=0A= + if (targetm.sched.first_cycle_multipass_dfa_lookahead=0A= + /* Don't use max_issue with reg_pressure scheduling. Multipass=0A= + scheduling and reg_pressure scheduling undo each other's decisions. */= =0A= + && sched_pressure =3D=3D SCHED_PRESSURE_NONE)=0A= dfa_lookahead =3D targetm.sched.first_cycle_multipass_dfa_lookahead ()= ;=0A= else=0A= dfa_lookahead =3D 0;=0A= --=20=0A= 1.7.9.5=0A= =0A= --Apple-Mail=_C9620461-374F-42C3-A9B6-9171518BD05C--