From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261884AbVFWAKr (ORCPT ); Wed, 22 Jun 2005 20:10:47 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261842AbVFWAKJ (ORCPT ); Wed, 22 Jun 2005 20:10:09 -0400 Received: from fmr22.intel.com ([143.183.121.14]:38838 "EHLO scsfmr002.sc.intel.com") by vger.kernel.org with ESMTP id S261156AbVFWAJU (ORCPT ); Wed, 22 Jun 2005 20:09:20 -0400 Message-Id: <200506230009.j5N09Hg11772@unix-os.sc.intel.com> From: "Chen, Kenneth W" To: Cc: Subject: RE: Update: Industry db benchmark result Date: Wed, 22 Jun 2005 17:09:17 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Mailer: Microsoft Office Outlook, Build 11.0.6353 Thread-Index: AcV3hDaw5pwRRpShSHu/LFhAbB+HeAAAgsIg X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180 In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Andi Kleen wrote on Wednesday, June 22, 2005 4:44 PM > "Chen, Kenneth W" writes: > > > > It's been hovering around -13%. > > Do you have an educated guess where the 13% loss is comming from? Process scheduler is the prime suspect. Earlier scheduler experiments on top of 2.6.11 showed that it was about 10% regression. Similar experiments repeated on top of 2.6.12-rc4 indicated about a much smaller 4%. Other area are O_DIRECT (2%), and possibly scsi. - Ken