Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1030332AbXAKNG6 (ORCPT ); Thu, 11 Jan 2007 08:06:58 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1030333AbXAKNG6 (ORCPT ); Thu, 11 Jan 2007 08:06:58 -0500 Received: from nikam-dmz.ms.mff.cuni.cz ([195.113.20.16]:32820 "EHLO nikam.ms.mff.cuni.cz" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1030332AbXAKNG5 (ORCPT ); Thu, 11 Jan 2007 08:06:57 -0500 Date: Thu, 11 Jan 2007 14:06:54 +0100 From: Martin Mares To: "linux-os (Dick Johnson)" Cc: Aubrey , Hua Zhong , Hugh Dickins , linux-kernel@vger.kernel.org, hch@infradead.org, kenneth.w.chen@intel.com, akpm@osdl.org, torvalds@osdl.org, mjt@tls.msk.ru Subject: Re: O_DIRECT question Message-ID: References: <6d6a94c50701101857v2af1e097xde69e592135e54ae@mail.gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: User-Agent: Mutt/1.5.9i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 903 Lines: 21 Hello! > Maybe you need to say why you want to use O_DIRECT with its terrible > performance? Incidentally, I was writing an external-memory radix-sort some time ago and it turned out that writing to 256 files at once is much faster with O_DIRECT than through the page cache, very likely because the page cache is flushing pages in essentially random order. Tweaking VM parameters and block device queue size helped, but only a little. Have a nice fortnight -- Martin `MJ' Mares http://mj.ucw.cz/ Faculty of Math and Physics, Charles University, Prague, Czech Rep., Earth Linux vs. Windows is a no-WIN situation. - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/