Subject: Re: gettimeofday() system call timing curiosity
To: pjd@fred001.dynip.com
Date: Sat, 9 Mar 2002 18:51:19 +0000 (GMT)
Cc: alan@lxorguk.ukuu.org.uk (Alan Cox), root@chaos.analogic.com,
        johan.adolfsson@axis.com, lk@tantalophile.demon.co.uk (Jamie Lokier),
        hpa@zytor.com (H. Peter Anvin), linux-kernel@vger.kernel.org,
        terje.eggestad@scali.com (Terje Eggestad),
        greearb@candelatech.com (Ben Greear),
        davidel@xmailserver.org (Davide Libenzi),
        george@mvista.com (george anzinger)
In-Reply-To: <200203090303.g29332R14018@fred.cambridge.ma.us> from "pjd@fred.cambridge.ma.us" at Mar 08, 2002 10:03:02 PM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-Id: <E16jlw7-0002BC-00@the-village.bc.nu>
From: Alan Cox <alan@lxorguk.ukuu.org.uk>
Sender: linux-kernel-owner@vger.kernel.org

> Curiously, on a 1.7GHz P4 running RH7.1 (2.4.2 i686) it never hits
> the same case, and takes slightly over a second to call gettimeofday a

Nothing curious about that I'm afraid. Every benchmark I've run on a PIV
shows that kind of behaviour - anything but perfectly predicted linear
code flow seems to hurt it. Locked operations really seem to make it suffer.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/