Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S934121Ab0KPJd1 (ORCPT ); Tue, 16 Nov 2010 04:33:27 -0500 Received: from mga11.intel.com ([192.55.52.93]:13005 "EHLO mga11.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S934110Ab0KPJd0 (ORCPT ); Tue, 16 Nov 2010 04:33:26 -0500 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="4.59,205,1288594800"; d="scan'208";a="858053760" Subject: [performance bug] volanomark regression on 37-rc1 From: "Alex,Shi" To: ncrao@google.com, a.p.zijlstra@chello.nl, mingo@elte.hu Cc: linux-kernel@vger.kernel.org, "Chen, Tim C" , zheng.z.yan@intel.com Content-Type: text/plain; charset="UTF-8" Date: Tue, 16 Nov 2010 17:34:02 +0800 Message-ID: <1289900042.27424.253.camel@debian> Mime-Version: 1.0 X-Mailer: Evolution 2.28.1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2591 Lines: 60 When do performance testing on 37-rc1 kernel on Core2 machines, we find the volanomark loopback performance drop about 30%, that due to commit:fab476228ba37907ad7 Volanomark link: http://www.volano.com/benchmarks.html Our volanomark testing parameters as following: "-count 25000 -rooms 10 " JVM is jrockit-R27.3.1-jre1.5.0_11 java_options is "-Xmx1500m -Xms1500m -Xns750m -XXaggressive -Xlargepages -XXlazyUnlocking -Xgc:genpar -XXtlasize:min=16k,preferred=64k" and we set /proc/sys/kernel/sched_compat_yield as "1". We find if with the following patch, the regression can be recovered. Signed-off-by:Alex Shi diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index f4f6a83..5dca678 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c @@ -1766,7 +1766,6 @@ static void pull_task(struct rq *src_rq, struct task_struct *p, check_preempt_curr(this_rq, p, 0); /* re-arm NEWIDLE balancing when moving tasks */ - src_rq->avg_idle = this_rq->avg_idle = 2*sysctl_sched_migration_cost; this_rq->idle_stamp = 0; } It seems some of load_balance() is not necessary that caused by avg_idle setting. But do not know more details of the volano running. Anyone like to give a comments for this issue? Ncrao, I have no idea of your benchmarks, but just guess removing the avg_idle setting won't bring much wakeup delay for tasks. Could you like to show some data of this? The vmstat output for .36 and .37-rc1 kernel as below: 2.6.36 procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------ r b swpd free buff cache si so bi bo in cs us sy id wa st 394 0 0 4680728 3168 118972 0 0 0 3 5314 1348711 38 62 0 0 0 396 0 0 4680500 3184 118976 0 0 0 8 5345 1303237 38 62 0 0 0 413 0 0 4680252 3200 118976 0 0 0 3 5082 1326851 38 61 0 0 0 2.6.37-rc1 procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu------ r b swpd free buff cache si so bi bo in cs us sy id wa st 329 0 0 4653748 4600 134032 0 0 0 7 11649 918370 35 64 0 0 0 210 0 0 4653492 4608 134032 0 0 0 6 11957 898011 36 64 0 0 0 373 0 0 4653496 4624 134032 0 0 0 4 11736 912468 36 64 0 0 0 Regards Alex -- 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/