From owner-freebsd-chat Fri Dec 10 2:13:26 1999 Delivered-To: freebsd-chat@freebsd.org Received: from smtp01.primenet.com (smtp01.primenet.com [206.165.6.131]) by hub.freebsd.org (Postfix) with ESMTP id 0E93715372 for ; Fri, 10 Dec 1999 02:13:21 -0800 (PST) (envelope-from tlambert@usr02.primenet.com) Received: (from daemon@localhost) by smtp01.primenet.com (8.9.3/8.9.3) id TAA22667; Thu, 9 Dec 1999 19:54:27 -0700 (MST) Received: from usr02.primenet.com(206.165.6.202) via SMTP by smtp01.primenet.com, id smtpdAAAXWaOaj; Thu Dec 9 19:13:50 1999 Received: (from tlambert@localhost) by usr02.primenet.com (8.8.5/8.8.5) id TAA04264; Thu, 9 Dec 1999 19:13:35 -0700 (MST) From: Terry Lambert Message-Id: <199912100213.TAA04264@usr02.primenet.com> Subject: Re: dual 400 -> dual 600 worth it? To: des@flood.ping.uio.no (Dag-Erling Smorgrav) Date: Fri, 10 Dec 1999 02:13:34 +0000 (GMT) Cc: andrews@technologist.com, Doug@gorean.org, bright@wintelcom.net, chat@FreeBSD.ORG, doconnor@gsoft.com.au, dscheidt@enteract.com In-Reply-To: from "Dag-Erling Smorgrav" at Dec 8, 99 03:19:23 pm X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-chat@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org > > > Yeah, the new box I'm evaluating has SCA LVD SCSI, and it goes a > > > lot faster. I'm compiling -Stable and so far -j 6, 8 and 12 have all > > It _SHOULD_ go faster with SCSI as opposed to (E)IDE/UDMA/etc. > > Why, because "Scuzzy" is a cooler name than "Eye-dee-ee"? SCSI has > higher overhead than IDE, so for a single-disk system (or a two-disk > system, provided each is on a separate IDE bus), IDE wins (given > otherwise identical disks, of course). FWIW, while the IDE specification supports tagged command queues to allow more than one disk transaction to be outstanding, there are no IDE drives currently available that support this (IBM has run some in some labs, but there was no real interest in getting them out, and I am told the project was scrapped for lack of controller support on other than lab-based controllers). This means that for server systems, A SCSI drive with a tagged command queue depth of 128 (common on a number of IBM drives, just to keep the vendor the same) can support 128 times as much concurrency as an IDE drive, everything else about the drive being equal. It's constantly amazing to me that the same people who state that FreeBSD should not go after the desktop and should not have graphical logins and other destop workstation fluff, are the same people who claim that IDE is as good as, or better than, SCSI. Perhaps for a single user workstation, IDE _is_ better than SCSI. All of the benchmarks that claim this are non-concurrent, after all, just like the one application likely to be running at a time on a single user workstation. For heavily loaded servers, howwever, there is absolutely no comparison: SCSI wins because of concurrency, and latency for single-user, single-threaded operations be damned. PS: My SCSI-based, mirrored NOC disk array on my NOC is capable of handling all of BEST Internet Inc.'s mail for a full month in just under 48 hours... for in excess of 10,000 transiently connected servers sending it ETRNs; what's your IDE based NOC capable of? Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message