From owner-freebsd-scsi Mon Oct 25 9: 8:44 1999 Delivered-To: freebsd-scsi@freebsd.org Received: from mail.tvol.com (mail.wgate.com [38.219.83.4]) by hub.freebsd.org (Postfix) with ESMTP id 77CB5150B4 for ; Mon, 25 Oct 1999 09:08:36 -0700 (PDT) (envelope-from rjesup@wgate.com) Received: from jesup.eng.tvol.net (jesup.eng.tvol.net [10.32.2.26]) by mail.tvol.com (8.8.8/8.8.3) with ESMTP id MAA15653; Mon, 25 Oct 1999 12:03:04 -0400 (EDT) Reply-To: Randell Jesup To: "Michael Sinz" Cc: "Geoff Buckingham" , "Gerard Roudier" , "scsi@FreeBSD.ORG" Subject: Re: FreeBSD 3.2 / Slow SCSI Dell PowerEdge 4300 References: <199910251239.IAA28752@vixen.sinz.org> From: Randell Jesup Date: 25 Oct 1999 12:04:37 +0000 In-Reply-To: "Michael Sinz"'s message of "Mon, 25 Oct 1999 08:40:43 -0400" Message-ID: X-Mailer: Gnus v5.6.43/Emacs 20.4 Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org "Michael Sinz" writes: >>> For a general purpose server, one must assume that the special case of >>> the I/O working out to be single threaded will not happen. Multiple things >>> will be going on and the working set will be larger than the cache size. >>> A bit of overhead added to the "simple" cases will make the general >>> operation better. Benchmarks, however, may well show this as slower >>> since some extra overhead had to be added. Benchmarks would need to >>> become much more complex in order to show the real benefit or lack of >>> benefit for any one technique. >>> >>Which brings us back to the question as to wether or not disabling TAGs for >>WDE * is the correct thing to do? IMHO it is not. > >For the general case I would say that disabling TAGS is not worth the >few specific case performance improvements (most of which are benchmark-only) I must agree for a server that Tags are good in most cases. Even for a single-user machine, I'm not certain I'd take a significant hit to multitasking (multiple-accessor) disk performance to get a 5% improvement to pure sequential IO. Now, there may be issues with the disksort algorithm that is loosing the benefits of Tags for multi-threaded IO as well; remember that most disksort algorithms were originally designed with cacheless (and single-threaded) drives in mind. Have there been any good ACM/etc papers on disksort and multi-threaded IO systems in the last few years? -- Randell Jesup, Worldgate Communications, ex-Scala, ex-Amiga OS team ('88-94) rjesup@wgate.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message