This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: [PATCH] Fix PR18754: add early loop pass, 2nd try

From: Richard Guenther <richard dot guenther at gmail dot com>
To: Richard Guenther <rguenth at tat dot physik dot uni-tuebingen dot de>
Cc: Giovanni Bajo <giovannibajo at libero dot it>, gcc-patches at gcc dot gnu dot org
Date: Thu, 20 Jan 2005 16:29:07 +0100
Subject: Re: [PATCH] Fix PR18754: add early loop pass, 2nd try
References: <Pine.LNX.4.44.0501201448030.16464-100000@alwazn.tat.physik.uni-tuebingen.de> <Pine.LNX.4.44.0501201526150.16464-100000@alwazn.tat.physik.uni-tuebingen.de>
Reply-to: Richard Guenther <richard dot guenther at gmail dot com>

On Thu, 20 Jan 2005 15:26:49 +0100 (CET), Richard Guenther
<rguenth@tat.physik.uni-tuebingen.de> wrote:
> On Thu, 20 Jan 2005, Richard Guenther wrote:
> 
> > From inspecting generated assembly code for POOMA, we don't do
> > very good.  It seems the reduced C++ testcase from 18754 does
> > not contain all difficulties.
> >
> > Back to experimenting...
> 
> It seems those are ivopt problems - I'll apply some of the pending
> ivopt patches.

Applying #7 and #10 (just picked them as potentially profitable) improves
the runtime with early loop to 11.8s while not changing numbers for
the run without.  That's 10% total gain for early loop - I guess I keep that
patch locally.

Maybe we can get some of the more obvious patches from
http://gcc.gnu.org/ml/gcc-patches/2004-12/msg01381.html
applied.

Richard.

References:
- Re: [PATCH] Fix PR18754: add early loop pass, 2nd try
  - From: Richard Guenther
- Re: [PATCH] Fix PR18754: add early loop pass, 2nd try
  - From: Richard Guenther

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]