Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 26 Dec 2000 09:29:59 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 26 Dec 2000 09:29:49 -0500 Received: from brutus.conectiva.com.br ([200.250.58.146]:52471 "EHLO brutus.conectiva.com.br") by vger.kernel.org with ESMTP id ; Tue, 26 Dec 2000 09:29:42 -0500 Date: Tue, 26 Dec 2000 11:58:34 -0200 (BRDT) From: Rik van Riel To: Felix von Leitner cc: linux-kernel@vger.kernel.org Subject: Re: Abysmal RAID 0 performance on 2.4.0-test10 for IDE? In-Reply-To: <20001226004843.A6103@convergence.de> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Tue, 26 Dec 2000, Felix von Leitner wrote: > Thus spake Felix von Leitner (leitner@convergence.de): > > Here is the result of my test program on the strip set: > > # rb < /dev/md/0 > > 30.3 meg/sec > > # > > One more detail: top says the CPU is 50% system when reading from either > one of the disk or raid devices. That seems awfully high considering > that the Promise controller claims to do UDMA. > > Any comments? Your program reads in data at 30MB/second, on a memory bus that most likely supports something like 60 to 100MB/second. Part of this memory bandwidth is needed for the UDMA controller to push the data to memory, probably between 30% and 50%. Every time the UDMA controller has the memory bus for itself the CPU will busy-wait on memory, which shows up as CPU busy time. regards, Rik -- Hollywood goes for world dumbination, Trailer at 11. http://www.surriel.com/ http://www.conectiva.com/ http://distro.conectiva.com.br/ - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/