From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-106401-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 17055 invoked by alias); 10 Dec 2004 16:21:07 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
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
Received: (qmail 16444 invoked from network); 10 Dec 2004 16:20:47 -0000
Received: from unknown (HELO smtp2.fuse.net) (216.68.8.175)
  by sourceware.org with SMTP; 10 Dec 2004 16:20:47 -0000
Received: from gx6.fuse.net ([66.42.246.223]) by smtp2.fuse.net
          (InterMail vM.6.01.03.04 201-2131-111-106-20040729) with ESMTP
          id <20041210161936.MGIR26919.smtp2.fuse.net@gx6.fuse.net>
          for <gcc@gcc.gnu.org>; Fri, 10 Dec 2004 11:19:36 -0500
Received: from dellpi.pinski.fam ([66.42.246.223]) by gx6.fuse.net
          (InterMail vG.1.00.00.00 201-2136-104-20040331) with ESMTP
          id <20041210161908.NNAP29070.gx6.fuse.net@dellpi.pinski.fam>;
          Fri, 10 Dec 2004 11:19:08 -0500
Received: from [10.0.0.80] (zhivago.i.pinski.fam [10.0.0.80])
	by dellpi.pinski.fam (8.12.2/8.12.1) with ESMTP id iBAGKfVT019159;
	Fri, 10 Dec 2004 11:20:41 -0500 (EST)
In-Reply-To: <Pine.LNX.4.44.0412101616530.4584-100000@alwazn.tat.physik.uni-tuebingen.de>
References: <Pine.LNX.4.44.0412101616530.4584-100000@alwazn.tat.physik.uni-tuebingen.de>
Mime-Version: 1.0 (Apple Message framework v619)
Content-Type: text/plain; charset=US-ASCII; format=flowed
Message-Id: <6E9B793A-4AC7-11D9-8186-000A95D692F4@physics.uc.edu>
Content-Transfer-Encoding: 7bit
Cc: gcc@gcc.gnu.org
From: Andrew Pinski <pinskia@physics.uc.edu>
Subject: Re: Simple loops not interchanged?
Date: Fri, 10 Dec 2004 16:21:00 -0000
To: Richard Guenther <rguenth@tat.physik.uni-tuebingen.de>
X-SW-Source: 2004-12/txt/msg00363.txt.bz2


On Dec 10, 2004, at 10:36 AM, Richard Guenther wrote:

> Hi!
>
> I expected -ftree-loop-linear to exchange loops in
>
> double foo(double *a)
> {
>         int i,j;
>         double r = 0.0;
> 	for (i=0; i<8; ++i)
>                 for (j=0; j<8; ++j)
> 			r += a[j*8+i];
>         return r;
> }
>
> but it tells me (regardless of loop order) that
> "Won't transform loop. Optimal transform is the identity transform"
> which I cannot believe, obviously.
>
> What's going wrong here?

First file a bug.
Second the loop linearizer loves ARRAY_REF and not INDIRECT_REF
for 4.1, we should be able to get MEM_REF which is like ARRAY_REF
and the loop lineaerizer will just do its job.

I tested the theory by having a local array instead of passing
the pointer.

Thanks,
Andrew Pinski