Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753966Ab2FSIsb (ORCPT ); Tue, 19 Jun 2012 04:48:31 -0400 Received: from mail-vc0-f174.google.com ([209.85.220.174]:56868 "EHLO mail-vc0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753718Ab2FSIs3 convert rfc822-to-8bit (ORCPT ); Tue, 19 Jun 2012 04:48:29 -0400 MIME-Version: 1.0 In-Reply-To: <1339435331.31548.19.camel@twins> References: <1337857490.7300.19.camel@marge.simpson.net> <1337865431.9783.148.camel@laptop> <1338906621.6110.30.camel@marge.simpson.net> <1339433835.7358.30.camel@marge.simpson.net> <1339435331.31548.19.camel@twins> From: Paul Turner Date: Tue, 19 Jun 2012 01:47:57 -0700 Message-ID: Subject: Re: [patch v3] sched: fix select_idle_sibling() induced bouncing To: Peter Zijlstra Cc: Mike Galbraith , lkml , Suresh Siddha , Arjan Van De Ven Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT X-System-Of-Record: true Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1600 Lines: 36 On Mon, Jun 11, 2012 at 10:22 AM, Peter Zijlstra wrote: > On Mon, 2012-06-11 at 18:57 +0200, Mike Galbraith wrote: > >> Traversing an entire package is not only expensive, it also leads to tasks >> bouncing all over a partially idle and possible quite large package. ?Fix >> that up by assigning a 'buddy' CPU to try to motivate. ?Each buddy may try >> to motivate that one other CPU, if it's busy, tough, it may then try it's >> SMT sibling, but that's all this optimization is allowed to cost. >> >> Sibling cache buddies are cross-wired to prevent bouncing. >> >> Signed-off-by: Mike Galbraith > > The patch could do with a little comment on how you achieve the > cross-wiring because staring at the code I go cross-eyed again ;-) > > Anyway, I think I'll grab it since nobody seems to have any objections > and the numbers seem good. > > PJT any progress on your load-tracking stuff? Arjan is interested in the > avg runtime estimation it has to make the whole wake an idle thing > conditional on. I'm still pretty much completely clobbered with internal stuff :( but it's on my agenda to put some cycles towards this week (and weekend). I'll say by Monday/Tuesday with a hopefully by-Friday. We've gotten a lot of of test hours on it internally so I'm comfortable posting a non-RFC mergable series. - Paul -- 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/