From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756167Ab3A1LVw (ORCPT ); Mon, 28 Jan 2013 06:21:52 -0500 Received: from mail-bk0-f49.google.com ([209.85.214.49]:37908 "EHLO mail-bk0-f49.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751035Ab3A1LVu (ORCPT ); Mon, 28 Jan 2013 06:21:50 -0500 Date: Mon, 28 Jan 2013 12:21:45 +0100 From: Ingo Molnar To: Mike Galbraith Cc: LKML , Ingo Molnar , Peter Zijlstra Subject: Re: [patch] sched: minimalist select_idle_sibling() bouncing cow syndrome fix Message-ID: <20130128112145.GA23495@gmail.com> References: <1359273021.5803.80.camel@marge.simpson.net> <20130128105304.GC20263@gmail.com> <1359371965.5783.127.camel@marge.simpson.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1359371965.5783.127.camel@marge.simpson.net> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org * Mike Galbraith wrote: > On Mon, 2013-01-28 at 11:53 +0100, Ingo Molnar wrote: > > * Mike Galbraith wrote: > > > > > If the previous CPU is cache affine and idle, select it. > > > > No objections in principle - but would be nice to have a > > changelog with numbers, % of improvement included and so? > > Well, that like was my changelog, guess it needs improvement. > > Take 2. > > sched: minimalist select_idle_sibling() bouncing cow syndrome fix > > If the previous CPU is cache affine and idle, select it. > > The current implementation simply traverses the sd_llc domain, > taking the first idle CPU encountered, which walks buddy pairs > hand in hand over the package, inflicting excruciating pain. > > 1 tbench pair (worst case) in a 10 core + SMT package: > > pre 15.22 MB/sec 1 procs > post 252.01 MB/sec 1 procs Drool ... :-) What would be a 'contrarian' test - i.e. a test where this could hurt most? Thanks, Ingo