From: "J. Bruce Fields" Subject: Re: [PATCH 0/11] nfsd: Summary of "Improve NFS server performance" Date: Wed, 22 Apr 2009 15:41:21 -0400 Message-ID: <20090422194121.GF9541@fieldses.org> References: <20090325133607.16437.33288.sendpatchset@localhost.localdomain> <20090421225728.GI27411@fieldses.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: jlayton@redhat.com, Krishna Kumar , linux-nfs@vger.kernel.org To: Krishna Kumar2 Return-path: Received: from mail.fieldses.org ([141.211.133.115]:34488 "EHLO pickle.fieldses.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753271AbZDVTl2 (ORCPT ); Wed, 22 Apr 2009 15:41:28 -0400 In-Reply-To: Sender: linux-nfs-owner@vger.kernel.org List-ID: On Wed, Apr 22, 2009 at 11:05:14AM +0530, Krishna Kumar2 wrote: > "J. Bruce Fields" wrote on 04/22/2009 04:27:28 AM: > > > The numbers look promising, but very noisy. Maybe an average of a few > > tests would give more stable numbers? > > Is it enough if I cut out some "popular" bufsizes and #processes out of > that > table and post it? If so, here is a smaller list: By "noisy" I just mean that they look like there's a lot of randomness. So, yes, I'm fine with a subset, but I'm curious what the variance is on repeated runs of each test. --b. > > ---------------------------------------------------------------------- > #Processes Bufsize OrgBW KB/s New BW KB/s % > ---------------------------------------------------------------------- > 4 4096 60660.67 64309.37 6.01 > 4 16384 60796.86 64069.82 5.38 > 4 65536 60774.45 63735.24 4.87 > 4 131072 61369.66 65261.85 6.34 > > 8 4096 50650.45 55400.01 9.37 > 8 16384 51114.76 56025.31 9.60 > 8 65536 51000.23 54285.63 6.44 > 8 131072 52996.73 54021.11 1.93 > > 16 4096 43897.69 46519.89 5.97 > 16 16384 43883.62 46726.03 6.47 > 16 65536 43804.33 44865.20 2.42 > 16 131072 44525.30 43752.62 -1.73 > > 32 4096 39913.14 42279.21 5.92 > 32 16384 38094.95 42645.32 11.94 > 32 65536 41438.05 41794.76 .86 > 32 131072 41869.06 43644.07 4.23 > > 64 4096 35468.42 35529.29 .17 > 64 16384 31926.51 32694.91 2.40 > 64 65536 36066.08 36359.77 .81 > 64 131072 35969.04 37462.86 4.15 > > 128 4096 29311.11 30097.65 2.68 > 128 16384 28283.58 29071.45 2.78 > 128 65536 32617.13 33778.46 3.56 > 128 131072 32972.71 34160.82 3.60 > > 192 4096 27368.48 29318.49 7.12 > 192 16384 26388.54 29719.15 12.62 > 192 65536 33374.85 33607.14 .69 > 192 131072 33523.48 32601.93 -2.74 > ----------------------------------------------------------------------- > 1170059.88 1223737.36 4.58% > ----------------------------------------------------------------------- > > Thanks, > > - KK >