From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756564Ab2ADRRR (ORCPT ); Wed, 4 Jan 2012 12:17:17 -0500 Received: from mx1.redhat.com ([209.132.183.28]:15394 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756447Ab2ADRRQ (ORCPT ); Wed, 4 Jan 2012 12:17:16 -0500 Message-ID: <4F04898B.1080600@redhat.com> Date: Wed, 04 Jan 2012 19:16:59 +0200 From: Avi Kivity User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:8.0) Gecko/20111115 Thunderbird/8.0 MIME-Version: 1.0 To: Rik van Riel CC: Nikunj A Dadhania , Ingo Molnar , peterz@infradead.org, linux-kernel@vger.kernel.org, vatsa@linux.vnet.ibm.com, bharata@linux.vnet.ibm.com Subject: Re: [RFC PATCH 0/4] Gang scheduling in CFS References: <20111219083141.32311.9429.stgit@abhimanyu.in.ibm.com> <20111219112326.GA15090@elte.hu> <87sjke1a53.fsf@abhimanyu.in.ibm.com> <4EF1B85F.7060105@redhat.com> <877h1o9dp7.fsf@linux.vnet.ibm.com> <20111223103620.GD4749@elte.hu> <4EF701C7.9080907@redhat.com> <20111230095147.GA10543@elte.hu> <878vlu4bgh.fsf@linux.vnet.ibm.com> <87pqf5mqg4.fsf@abhimanyu.in.ibm.com> <4F017AD2.3090504@redhat.com> <87mxa3zqm1.fsf@abhimanyu.in.ibm.com> <4F046536.5080207@redhat.com> <4F048295.1050907@redhat.com> In-Reply-To: <4F048295.1050907@redhat.com> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 01/04/2012 06:47 PM, Rik van Riel wrote: >> So it looks like the default is optimal, at least wrt the cases you >> tested and your test workload. > > > It depends on the workload. > > I believe ebizzy synchronously bounces messages around between > userland threads, and may benefit from lower latency preemption > and re-scheduling. > > Workloads like AMQP do asynchronous messaging, and are likely > to benefit from having a lower number of switches. > > I do not know which kind of workload is more prevalent. > > Another worry with gang scheduling is scalability. One of > the reasons Linux scales well to larger systems is that a > lot of things are done CPU local, without communicating > things with other CPUs. Making the scheduling algorithm > system-global has the potential to add in a lot of overhead. > > Likewise, removing the ability to migrate workloads to idle > CPUs is likely to hurt a lot of real world workloads. > > Benchmarks don't care, because they run full-out. However, > users do not run benchmarks nearly as much as they run > actual workloads... > I think we can solve it at the guest level. The paravirt ticketlock stuff introduces wait/wake calls (actually wait is just a HLT instruction); we could spin for a while, then HLT until the other side wakes us. We should do this for all sites that busy wait. -- error compiling committee.c: too many arguments to function