From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-156901-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 9568 invoked by alias); 29 Sep 2009 18:23:56 -0000
Received: (qmail 9557 invoked by uid 22791); 29 Sep 2009 18:23:55 -0000
X-SWARE-Spam-Status: No, hits=-1.8 required=5.0 	tests=AWL,BAYES_00,SARE_MSGID_LONG40,SPF_PASS
X-Spam-Check-By: sourceware.org
Received: from fg-out-1718.google.com (HELO fg-out-1718.google.com) (72.14.220.155)     by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Tue, 29 Sep 2009 18:23:51 +0000
Received: by fg-out-1718.google.com with SMTP id d23so1664382fga.8         for <gcc@gcc.gnu.org>; Tue, 29 Sep 2009 11:23:49 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.239.181.67 with SMTP id l3mr418943hbg.201.1254248629274; Tue,  	29 Sep 2009 11:23:49 -0700 (PDT)
In-Reply-To: <84fc9c000908200148p57979f3cp5e440f778b8fc484@mail.gmail.com>
References: <E1MdhYW-0005WQ-00.aserg2004-list-ru@f143.mail.ru> 	 <4A8BE7B2.7080708@inria.fr> 	 <84fc9c000908190507x708772cdueb8193a0ff7517aa@mail.gmail.com> 	 <4A8BFF5F.5070708@inria.fr> <4A8C0405.9030008@inria.fr> 	 <84fc9c000908190656p7ec6cc4aja93ff860a5f61960@mail.gmail.com> 	 <84fc9c000908190658k23a13034s64cac8be7b2de827@mail.gmail.com> 	 <4A8CA4BC.4020807@inria.fr> 	 <84fc9c000908200148p57979f3cp5e440f778b8fc484@mail.gmail.com>
Date: Tue, 29 Sep 2009 18:54:00 -0000
Message-ID: <303e1d290909291123t480939f6oe83d76c7536419e1@mail.gmail.com>
Subject: Re: complete_unrolli / complete_unroll
From: David Edelsohn <dje.gcc@gmail.com>
To: Richard Guenther <richard.guenther@gmail.com>
Cc: Albert Cohen <Albert.Cohen@inria.fr>, gcc@gcc.gnu.org
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
X-IsSubscribed: yes
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
X-SW-Source: 2009-09/txt/msg00610.txt.bz2

On Thu, Aug 20, 2009 at 4:48 AM, Richard Guenther
<richard.guenther@gmail.com> wrote:

> Can't we use graphite to re-roll loops? =A0That is, compress the
> polyhedron by introducing a new parameter? =A0But maybe I am
> not good at guessing what your initial bloat issue looks like.
>
> The reason I'm asking is that there is enough code out in the
> wild that has loops with manually unrolled bodies - I have seen
> up to 16 times here.

Do we want to try to address this partially in GCC 4.5?  Providing
some way to disable early unrolling either explicitly or implicitly
when Graphite is enabled?

Early unrolling can cause two problems:

1) Increase the size of SCoPs, which increases memory consumption and
analysis time.

2) Confusing SCoP analysis.

Separate from re-rolling and other long-term solutions, it would be
helpful for Graphite if there was some explicit control over early
unrolling to help with experimentation.

Thanks, David