From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-2.9 required=5.0 tests=BAYES_00,FREEMAIL_FROM,
	MAILING_LIST_MULTI autolearn=unavailable autolearn_force=no
	version=3.4.4
X-Google-Thread: 103376,acba876b1e3c9639
X-Google-Attributes: gid103376,public
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news1.google.com!news3.google.com!proxad.net!cleanfeed3-a.proxad.net!nnrp8-1.free.fr!not-for-mail
Return-Path: <baldrick@free.fr>
From: Duncan Sands <baldrick@free.fr>
To: comp.lang.ada@ada-france.org
Subject: Re: GNAT Optimization of Constant Expressions
Date: Fri, 18 May 2007 19:49:53 +0200
User-Agent: KMail/1.9.6
References: <1179355028.624745.258370@q75g2000hsh.googlegroups.com>
	<f2kn4p$av6$1@jacob-sparre.dk>
	<200705181935.23877.baldrick@free.fr>
In-Reply-To: <200705181935.23877.baldrick@free.fr>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
X-Virus-Scanned: amavisd-new at ada-france.org
Cc: Randy Brukardt <randy@rrsoftware.com>
X-BeenThere: comp.lang.ada@ada-france.org
X-Mailman-Version: 2.1.9rc1
Precedence: list
List-Id: "Gateway to the comp.lang.ada Usenet newsgroup"
	<comp.lang.ada.ada-france.org>
List-Unsubscribe: <http://www.ada-france.org/mailman/listinfo/comp.lang.ada>,
	<mailto:comp.lang.ada-request@ada-france.org?subject=unsubscribe>
List-Post: <mailto:comp.lang.ada@ada-france.org>
List-Help: <mailto:comp.lang.ada-request@ada-france.org?subject=help>
List-Subscribe: <http://www.ada-france.org/mailman/listinfo/comp.lang.ada>,
	<mailto:comp.lang.ada-request@ada-france.org?subject=subscribe>
Newsgroups: comp.lang.ada
Message-ID: <mailman.154.1179510602.18371.comp.lang.ada@ada-france.org>
X-Leafnode-NNTP-Posting-Host: 88.191.17.134
Organization: Guest of ProXad - France
NNTP-Posting-Date: 18 May 2007 19:50:09 MEST
NNTP-Posting-Host: 88.191.14.223
X-Trace: 1179510609 news-2.free.fr 25709 88.191.14.223:48015
X-Complaints-To: abuse@proxad.net
Xref: g2news1.google.com comp.lang.ada:15840
Date: 2007-05-18T19:50:09+02:00

> while what you say is very logical, unfortunately it seems to be wrong: I
> checked and these functions are not being marked pure.  So how come declaring
> a constant helps?  I don't know - I didn't look into it yet.

I took a quick look.  At -O2 defining constants makes no difference (in the
C case "const" functions are hoisted out of the loop at -O2) but with -O3
the calls are hoisted.  This is presumably because gcc inlines the bodies
for sin etc at -O3 and sees that in fact they have no side-effects, but
for some reason doesn't hoist them unless the constants have been defined.

Ciao,

Duncan.