From owner-cvs-all Fri Oct 26 19:46:37 2001 Delivered-To: cvs-all@freebsd.org Received: from beppo.feral.com (beppo.feral.com [192.67.166.79]) by hub.freebsd.org (Postfix) with ESMTP id F2B1737B406; Fri, 26 Oct 2001 19:46:28 -0700 (PDT) Received: from mailhost.feral.com (mjacob@mailhost.feral.com [192.67.166.1]) by beppo.feral.com (8.11.3/8.11.3) with ESMTP id f9R2kRH13705; Fri, 26 Oct 2001 19:46:27 -0700 (PDT) (envelope-from mjacob@feral.com) Date: Fri, 26 Oct 2001 19:46:27 -0700 (PDT) From: Matthew Jacob X-Sender: mjacob@beppo Reply-To: mjacob@feral.com To: Greg Lehey Cc: Doug Rabson , Luigi Rizzo , John Baldwin , Jonathan Lemon , cvs-all@FreeBSD.org, cvs-committers@FreeBSD.org Subject: Re: RAID-5 parity calculations (was: cvs commit: src/sys/dev/fxp if_fx) In-Reply-To: <20011027111145.A7846@wantadilla.lemis.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=X-UNKNOWN Content-Transfer-Encoding: 8BIT Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG > >> Sorry, that seems wrong to me. > > Have you done measurements? You know, I have, but it's now been 5 years- I'd have to dig the notebook out of storage. > > >> Typical RAID write performance for something like an Sun A1000 > >> which has a pentium in it is about 50% of theoretical. > > That doesn't mean that's because of the parity calculations. *blush* You may be right there! > > > I guess the real question is: 'can you get the parity calculations > > done in time so that the entire stripe can go out together'. > > Why? Because it's best if you can push stripes in some configurations out together- essentially a software spindle sync. > > > This obviously doesn't really work for the first request unless you > > delay it. If you have a hugely deep queue, you will burn your > > central processor doing things that are not germane to regular > > systems work- you can't help but be assisted by a coprocessor doing > > that work (it's like bcopy h/w). > > I don't see much similarity. As you say, it's the steady state > performance that's important. Basically, the resource you want to > optimize is disk bandwidth. Even if you do the calculations > instantaneously, the data doesn't get written out until the disk has > time to do it, and in a heavy load situation that will mean queueing > behind other requests. > > Instantaneously? Well, how long does it take to checksum 6 kB (the > average request size)? That's 1500 words, say 10000 instructions. At > 1 GHz, that's 10 µs, which is completely negligible compared to the > minimum four I/O transactions needed, which add up to about 25 ms. > Hmm. I think I've fallen victim to 'old assumptions' syndrome. I keep on forgetting how fast processors are getting. Apologies. I'll go rethink some now. -matt To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message