From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754551AbZINDbj (ORCPT ); Sun, 13 Sep 2009 23:31:39 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754028AbZINDbi (ORCPT ); Sun, 13 Sep 2009 23:31:38 -0400 Received: from mga06.intel.com ([134.134.136.21]:21836 "EHLO orsmga101.jf.intel.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1753603AbZINDbi (ORCPT ); Sun, 13 Sep 2009 23:31:38 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.44,381,1249282800"; d="scan'208";a="448189138" Subject: Re: PATCH] cpuidle: A new variant of the menu governor to boost IO performance From: "Zhang, Yanmin" To: John Stoffel Cc: Arjan van de Ven , linux-kernel@vger.kernel.org, lenb@kernel.org, mingo@elte.hu, akpm@linux-foundation.org, peterz@infradead.org, torvalds@linux-foundation.org, jens.axboe@oracle.com In-Reply-To: <19114.41505.479086.782442@stoffel.org> References: <20090911174019.1ed02737@infradead.org> <19114.41505.479086.782442@stoffel.org> Content-Type: text/plain Date: Mon, 14 Sep 2009 11:30:10 +0800 Message-Id: <1252899010.2606.25.camel@ymzhang> Mime-Version: 1.0 X-Mailer: Evolution 2.22.1 (2.22.1-2.fc9) Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 2009-09-11 at 15:16 -0400, John Stoffel wrote: > >>>>> "Arjan" == Arjan van de Ven writes: > > Arjan> From: Arjan van de Ven > Arjan> Subject: [PATCH] cpuidle: A new variant of the menu governor > > Arjan> This patch adds a new idle governor which balances power savings, > Arjan> energy efficiency and performance impact. > > Arjan> The reason for a reworked governor is that there have been > Arjan> serious performance issues reported with the existing code > Arjan> on Nehalem server systems. > > Arjan> To show this I'm sure Andrew wants to see benchmark results: > Arjan> (benchmark is "fio", "no cstates" is using "idle=poll") > > Arjan> no cstates current linux new algorithm > Arjan> 1 disk 107 Mb/s 85 Mb/s 105 Mb/s > Arjan> 2 disks 215 Mb/s 123 Mb/s 209 Mb/s > Arjan> 12 disks 590 Mb/s 320 Mb/s 585 Mb/s > > Don't you need another row or three where you show a) how much time > each test took, We start fio and always ask it running for 15 minutes. > and b) how much (or average) power used for the > duration of the test? The power consumption with the patch almost is equal to the one without the patch. But the fio result is far better with the patch. > > I'm just curious if the new algorithm (or even the current one!) saves > any appreciable power over the 'no cstates' case. It's not clear what > the savings are. > > Also, latency in terms of switching to higher power and then back down > would be nice to see. > > Cheers, > John