From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-79952-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 16991 invoked by alias); 29 Jul 2003 14:35:42 -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 16881 invoked from network); 29 Jul 2003 14:35:38 -0000
Received: from unknown (HELO mx01.uni-tuebingen.de) (134.2.3.11)
  by sources.redhat.com with SMTP; 29 Jul 2003 14:35:38 -0000
Received: from bellatrix.tat.physik.uni-tuebingen.de (bellatrix.tat.physik.uni-tuebingen.de [134.2.170.113])
	by mx01.uni-tuebingen.de (8.12.3/8.12.3) with ESMTP id h6TEZa0w026385;
	Tue, 29 Jul 2003 16:35:36 +0200
Received: from bellatrix.tat.physik.uni-tuebingen.de (localhost [127.0.0.1])
	by bellatrix.tat.physik.uni-tuebingen.de (8.12.3/8.12.2/SuSE Linux 0.6) with ESMTP id h6TEZaKA020063;
	Tue, 29 Jul 2003 16:35:36 +0200
Received: from localhost (rguenth@localhost)
	by bellatrix.tat.physik.uni-tuebingen.de (8.12.3/8.12.3/Submit) with ESMTP id h6TEZaDu020059;
	Tue, 29 Jul 2003 16:35:36 +0200
X-Authentication-Warning: bellatrix.tat.physik.uni-tuebingen.de: rguenth owned process doing -bs
Date: Tue, 29 Jul 2003 15:24:00 -0000
From: Richard Guenther <rguenth@tat.physik.uni-tuebingen.de>
To: Gabriel Dos Reis <gdr@integrable-solutions.net>
cc: Steven Bosscher <s.bosscher@student.tudelft.nl>, <gcc@gcc.gnu.org>
Subject: Re: std::pow implementation
In-Reply-To: <m3vftlo1ik.fsf@uniton.integrable-solutions.net>
Message-ID: <Pine.LNX.4.44.0307291624410.14031-100000@bellatrix.tat.physik.uni-tuebingen.de>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-AntiVirus: checked by AntiVir Milter 1.0.2; AVE 6.20.0.1; VDF 6.20.0.50
X-SW-Source: 2003-07/txt/msg02002.txt.bz2

If we go back to the original question, wether std::__pow_helper and
std::__cmath_power are good implementations, please.

It has been shown that the current gcc unroller is not able to
unroll the while loop in std::__cmath_power() and such even inlining
does not help to create good code. Also inlining the loop if it is
not unrolled and its contents not sufficiently cprop'ed is harmful.

Isnt what one would expect to inline for constant and small __n only?
This we can make work using __builtin_constant_p() as proposed. This would
work for gcc3.3, too. For gcc3.4 we may want to rely on __builtin_pow()
instead (which calls libc pow() for not expanded cases which may be
slower than the pow(T, int) overload for small n).

Richard.

--
Richard Guenther <richard dot guenther at uni-tuebingen dot de>
WWW: http://www.tat.physik.uni-tuebingen.de/~rguenth/