From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750800AbWDKM1r (ORCPT ); Tue, 11 Apr 2006 08:27:47 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750801AbWDKM1r (ORCPT ); Tue, 11 Apr 2006 08:27:47 -0400 Received: from mailhub.sw.ru ([195.214.233.200]:59523 "EHLO relay.sw.ru") by vger.kernel.org with ESMTP id S1750800AbWDKM1r (ORCPT ); Tue, 11 Apr 2006 08:27:47 -0400 Message-ID: <443BA1D3.1070200@sw.ru> Date: Tue, 11 Apr 2006 16:32:19 +0400 From: Kirill Korotaev User-Agent: Mozilla/5.0 (X11; U; Linux i686; ru-RU; rv:1.2.1) Gecko/20030426 X-Accept-Language: ru-ru, en MIME-Version: 1.0 To: "Serge E. Hallyn" CC: linux-kernel@vger.kernel.org, herbert@13thfloor.at, devel@openvz.org, sam@vilain.net, "Eric W. Biederman" , xemul@sw.ru, James Morris Subject: Re: [RFC][PATCH 0/5] uts namespaces: Introduction References: <20060407095132.455784000@sergelap> In-Reply-To: <20060407095132.455784000@sergelap> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Serge, > This patchset is based on Kirill Korotaev's Mar 24 submission, taking > comments (in particular from James Morris and Eric Biederman) into > account. thanks a lot for doing this! > Some performance results are attached. I was mainly curious whether > it would be worth putting the task_struct->uts_ns pointer inside > a #ifdef CONFIG_UTS_NS. The result show that leaving it in when > CONFIG_UTS_NS=n has negligable performance impact, so that is the > approach this patch takes. Serge, your testing approach looks really strange for me. First of all, you selected the worst namespace to check performance overhead on. 1) uts_ns is rarely used and never used on hot paths, 2) also all these test suites below doesn't test the code paths you modified. So I wonder what was the goal of these tests, especially dbench?! Thanks, Kirill > > -serge > > Performance testing was done on a 2-cpu hyperthreaded > x86 box with 16G ram. The following tests were run: > dbench (20 times, four clients, on reiser fs non-isolated partition) > tbench (20 times, 5 connections) > kernbench (20 times) > reaim (20 times ranging from 1 to 15 users) > > They were run on 2.6.17-rc1: > pristine > patched, but with !CONFIG_UTS_NS ("disabled") > patched with CONFIG_UTS_NS=y ("enabled") > > All results are presented as means +/- 95% confidence interval. > > Dbench results: > pristine: 387.080727 +/- 9.344585 > patched disabled: 389.524364 +/- 9.574921 > patched enabled: 370.155600 +/- 30.127808 > > Tbench results: > pristine: 388.940100 +/- 18.095104 > patched disabled: 389.173700 +/- 23.658035 > patched enabled: 394.333200 +/- 25.813393 > > Kernbench results: > pristine: 70.317500 +/- 0.210833 > patched, disabled: 70.860000 +/- 0.179292 > patched, enabled: 70.346500 +/- 0.184784 > > Reaim results: > pristine: > Nclients Mean 95% CI > 1 106080.000000 11327.896029 > 3 236057.142000 18205.544810 > 5 247867.136000 23536.800062 > 7 265370.000000 21284.335743 > 9 262969.936000 18225.497529 > 11 278256.000000 6230.342816 > 13 284288.016000 8924.589388 > 15 286987.170000 7881.034658 > > patched, disabled: > Nclients Mean 95% CI > 1 105400.000000 8739.978241 > 3 229500.000000 0.000000 > 5 252325.176667 16685.663423 > 7 265125.000000 6747.777319 > 9 271258.645000 11715.635212 > 11 280662.608333 7775.229351 > 13 277719.706667 8173.390359 > 15 278515.421667 10963.211450 > > patched, enabled: > Nclients Mean 95% CI > 1 102000.000000 0.000000 > 3 224400.000000 14159.870036 > 5 242963.288000 40529.490781 > 7 255150.000000 8745.802081 > 9 270154.284000 8918.863136 > 11 283134.260000 12239.361252 > 13 288497.540000 11336.550964 > 15 280022.728000 8804.882369 > >