From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 17 Oct 2002 08:33:20 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 17 Oct 2002 08:33:20 -0400 Received: from p50847823.dip.t-dialin.net ([80.132.120.35]:59013 "HELO hermes.zoffice.de") by vger.kernel.org with SMTP id ; Thu, 17 Oct 2002 08:33:18 -0400 Subject: increasing blocksizes decrease performance in Gbit From: Zoltan Bogdan To: linux-kernel@vger.kernel.org Content-Type: text/plain Content-Transfer-Encoding: 7bit X-Mailer: Ximian Evolution 1.0.8 Date: 17 Oct 2002 14:38:59 +0200 Message-Id: <1034858339.13137.3.camel@hermes> Mime-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi, does anyone know about the performace impact whereabouts of different blocksizes? I'm using kernel 2.4.18 on a machine with 512MB Ram and two 3Ware Escalade 7850 controllers with an 8-Disk (hardware- )RAID 5 array on each. A softraid 0 Array spans those two Raidarrays. While doing some benchmarks over Gbit Network I experianced the following: Block | Read Lin | Read Rnd | Write Lin kB | MB/sec | MB/sec | MB/sec ------+----------+----------+----------- 16 | 25.9 | 3.3 | 12.1 32 | 31.5 | 6.6 | 17.8 256 | 15.4 | 13.2 | 21.3 1024 | 13.7 | 20.8 | 21.4 Any clue why the value for linear read decreases when the Blocksizes get higher? I thought the contrary to be true. I apologize if my question I to off topic. Thanks z