Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754643AbYJFGAW (ORCPT ); Mon, 6 Oct 2008 02:00:22 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752712AbYJFGAE (ORCPT ); Mon, 6 Oct 2008 02:00:04 -0400 Received: from mga10.intel.com ([192.55.52.92]:3727 "EHLO fmsmga102.fm.intel.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752640AbYJFGAB (ORCPT ); Mon, 6 Oct 2008 02:00:01 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.33,366,1220252400"; d="scan'208";a="623754589" Subject: Re: aim7 47% regression with 2.6.27-rc8 From: "Zhang, Yanmin" To: Mike Galbraith Cc: LKML In-Reply-To: <1223271482.7794.1.camel@marge.simson.net> References: <1223269135.1685.11.camel@ymzhang> <1223271482.7794.1.camel@marge.simson.net> Content-Type: text/plain Date: Mon, 06 Oct 2008 13:52:58 +0800 Message-Id: <1223272378.1685.12.camel@ymzhang> Mime-Version: 1.0 X-Mailer: Evolution 2.21.5 (2.21.5-2.fc9) Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2447 Lines: 67 On Mon, 2008-10-06 at 07:38 +0200, Mike Galbraith wrote: > On Mon, 2008-10-06 at 12:58 +0800, Zhang, Yanmin wrote: > > Comparing with 2.6.27-rc8, aim7 result has about 47% regression with 2.6.27-rc8 on > > my 16-core tigerton and 8 core+HyperThreading x86_64 machine. > > > > I bisected it down to patch: > > > > 302745699c1b675b5d2a1af87271de10e4d96b6a is first bad commit > > > After I reverted the patch against 2.6.27-rc8, the regression disappears. > > It's interesting that the regression doesn't exist on 8-core stoakley. > > Looks like it's been fixed meanwhile. I did a new quick test and the patch does fix it. Thanks, Yanmin > > commit 07454bfff151d2465ada809bbaddf3548cc1097c > Author: Thomas Gleixner > Date: Sat Oct 4 10:51:07 2008 +0200 > > clockevents: check broadcast tick device not the clock events device > > Impact: jiffies increment too fast. > > Hugh Dickins noted that with NOHZ=n and HIGHRES=n jiffies get > incremented too fast. The reason is a wrong check in the broadcast > enter/exit code, which keeps the local apic timer in periodic mode > when the switch happens. > > Signed-off-by: Thomas Gleixner > > diff --git a/kernel/time/tick-broadcast.c b/kernel/time/tick-broadcast.c > index bd70345..cb01cd8 100644 > --- a/kernel/time/tick-broadcast.c > +++ b/kernel/time/tick-broadcast.c > @@ -235,7 +235,8 @@ static void tick_do_broadcast_on_off(void *why) > case CLOCK_EVT_NOTIFY_BROADCAST_FORCE: > if (!cpu_isset(cpu, tick_broadcast_mask)) { > cpu_set(cpu, tick_broadcast_mask); > - if (bc->mode == TICKDEV_MODE_PERIODIC) > + if (tick_broadcast_device.mode == > + TICKDEV_MODE_PERIODIC) > clockevents_shutdown(dev); > } > if (*reason == CLOCK_EVT_NOTIFY_BROADCAST_FORCE) > @@ -245,7 +246,8 @@ static void tick_do_broadcast_on_off(void *why) > if (!tick_broadcast_force && > cpu_isset(cpu, tick_broadcast_mask)) { > cpu_clear(cpu, tick_broadcast_mask); > - if (bc->mode == TICKDEV_MODE_PERIODIC) > + if (tick_broadcast_device.mode == > + TICKDEV_MODE_PERIODIC) > tick_setup_periodic(dev, 0); > } > break; > > -- 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/