From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Sat, 1 Sep 2001 08:40:45 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Sat, 1 Sep 2001 08:40:36 -0400 Received: from lightning.swansea.linux.org.uk ([194.168.151.1]:58384 "EHLO the-village.bc.nu") by vger.kernel.org with ESMTP id ; Sat, 1 Sep 2001 08:40:26 -0400 Subject: Re: lilo vs other OS bootloaders was: FreeBSD makes progress To: pwaechtler@loewe-komp.de (Peter =?iso-8859-1?Q?W=E4chtler?=) Date: Sat, 1 Sep 2001 13:39:10 +0100 (BST) Cc: _deepfire@mail.ru (Samium Gromoff), linux-kernel@vger.kernel.org In-Reply-To: <3B90CE7E.F4F54D9@loewe-komp.de> from "Peter =?iso-8859-1?Q?W=E4chtler?=" at Sep 01, 2001 02:03:10 PM X-Mailer: ELM [version 2.5 PL6] MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Message-Id: From: Alan Cox Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > Did someone any benchmarking? Yes > I expect the loss of performance per application a none issue. > What do you think: >0.5%? It depends on the actual application > Are you considering interrupt latency in the first place? IRQ latency isnt the issue. Older x86 processors have dreadful locking performance. On an Athlon or later pIII there seems to be very little difference, but other stuff you can measure it. Locks cost. On older CPU's atomic operations go back to main memory and on newer processors they still cause pipeline stalls. Calling back to non inlined locks has a small hit too. > Then obviously BeOS is also engineered from idiots... > Oh, and QNX/RTP has separate kernels for UP/SMP. And they > don't need UP/SMP versions of "modules". I doubt they were idiots. I suspect they had different engineering constraints like "base system must fit on one floppy disk". Alan