From owner-freebsd-chat  Mon Jan 20 10:41:24 2003
Delivered-To: freebsd-chat@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP
	id 0D3DF37B401; Mon, 20 Jan 2003 10:41:23 -0800 (PST)
Received: from haldjas.folklore.ee (Haldjas.folklore.ee [193.40.6.121])
	by mx1.FreeBSD.org (Postfix) with ESMTP
	id DAA4743F18; Mon, 20 Jan 2003 10:41:21 -0800 (PST)
	(envelope-from narvi@haldjas.folklore.ee)
Received: from haldjas.folklore.ee (localhost [127.0.0.1])
	by haldjas.folklore.ee (8.12.3/8.11.3) with ESMTP id h0KIfAOs019501;
	Mon, 20 Jan 2003 20:41:10 +0200 (EET)
	(envelope-from narvi@haldjas.folklore.ee)
Received: from localhost (narvi@localhost)
	by haldjas.folklore.ee (8.12.3/8.12.3/Submit) with ESMTP id h0KIevuN019497;
	Mon, 20 Jan 2003 20:41:02 +0200 (EET)
Date: Mon, 20 Jan 2003 20:40:57 +0200 (EET)
From: Narvi <narvi@haldjas.folklore.ee>
To: Rahul Siddharthan <rsidd@online.fr>
Cc: Brett Glass <brett@lariat.org>,
	Terry Lambert <tlambert2@mindspring.com>,
	"Greg 'groggy' Lehey" <grog@FreeBSD.ORG>, <chat@FreeBSD.ORG>
Subject: Re: Bystander shot by a spam filter
In-Reply-To: <20030105192556.GA526@papagena.rockefeller.edu>
Message-ID: <20030120203345.W43637-100000@haldjas.folklore.ee>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-freebsd-chat@FreeBSD.ORG
Precedence: bulk
List-ID: <freebsd-chat.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo@FreeBSD.ORG?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo@FreeBSD.ORG?subject=subscribe%20freebsd-chat>
List-Unsubscribe: <mailto:majordomo@FreeBSD.ORG?subject=unsubscribe%20freebsd-chat>
X-Loop: FreeBSD.org


[I understand its a bit late to answer - sorry about that]

On Sun, 5 Jan 2003, Rahul Siddharthan wrote:

> Perhaps Intel produces better "straight" code than gcc -- not
> surprising if it's their chip -- but gcc actually does better
> optimization, and therefore catches up on more complex code?
>

icc does optimisations like automaticly vectorising code to make
use of teh SSE2 regsiters for floating point. Its well possible
that it over-optimisticly converts code for which this is in
fact a  pessimisation to use SSE2. IIRC the benchmark used the
same compiler flags for all programs, and you can always run
into pessimistaions with that with sufficently aggressive options.

This is not an icc specific pitfall either.

> - Rahul
>
> To Unsubscribe: send mail to majordomo@FreeBSD.org
> with "unsubscribe freebsd-chat" in the body of the message
>


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message