From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752134AbeA3Nk4 (ORCPT ); Tue, 30 Jan 2018 08:40:56 -0500 Received: from bombadil.infradead.org ([65.50.211.133]:47605 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751674AbeA3Nkz (ORCPT ); Tue, 30 Jan 2018 08:40:55 -0500 Date: Tue, 30 Jan 2018 14:40:49 +0100 From: Peter Zijlstra To: Mel Gorman Cc: Mike Galbraith , Matt Fleming , LKML , rjw@rjwysocki.net, srinivas.pandruvada@linux.intel.com Subject: Re: [PATCH 4/4] sched/fair: Use a recently used CPU as an idle candidate and the basis for SIS Message-ID: <20180130134049.GU2249@hirez.programming.kicks-ass.net> References: <20180130104555.4125-1-mgorman@techsingularity.net> <20180130104555.4125-5-mgorman@techsingularity.net> <20180130115054.GA2269@hirez.programming.kicks-ass.net> <20180130125718.iwntjuvcp3yplvdx@techsingularity.net> <20180130131531.GD2269@hirez.programming.kicks-ass.net> <20180130132527.tjc6vea76sm5zm76@techsingularity.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180130132527.tjc6vea76sm5zm76@techsingularity.net> User-Agent: Mutt/1.9.2 (2017-12-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Tue, Jan 30, 2018 at 01:25:27PM +0000, Mel Gorman wrote: > > Because, esp. in this scenario; a task migrating; the hardware really > > can't do anything sensible, whereas the OS _knows_. > > Potentially yes. One option without HWP would be to track utilisation > for a task or artifically boost it for a short period after migration so > a higher p-state is potentially selected. With HWP, a hint would have to > be given to the hardware to try select a higher frequency but I've no idea > how expensive that is or how it would behave on different implementations > of HWP. It may also be a game of whack-a-mole trying to get every cpufreq > configuration correct. We have much of this infrastructure and have been working on improving it [1]. We already track per task utilization and feed it into a cpufreq governor (schedutil). > One advantage of using fewer cores is that it should > work regardless of cpufreq driver. Sure, and I started out by saying this patch isn't necessarily bad; but I think our 'use' [2] of HWP _is_. [1] https://lkml.kernel.org/r/20180123180847.4477-1-patrick.bellasi@arm.com [2] IIRC we basically don't do _anything_ when HWP.