From owner-freebsd-scsi Wed Aug 20 00:32:42 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.5/8.8.5) id AAA05799 for freebsd-scsi-outgoing; Wed, 20 Aug 1997 00:32:42 -0700 (PDT) Received: from sendero-ppp.i-connect.net (sendero-ppp.i-Connect.Net [206.190.143.100]) by hub.freebsd.org (8.8.5/8.8.5) with SMTP id AAA05767 for ; Wed, 20 Aug 1997 00:32:34 -0700 (PDT) Received: (qmail 5018 invoked by uid 1000); 20 Aug 1997 07:32:51 -0000 Message-ID: X-Mailer: XFMail 1.2-alpha [p0] on FreeBSD Content-Type: text/plain; charset=iso-8859-8 Content-Transfer-Encoding: 8bit MIME-Version: 1.0 In-Reply-To: <9708190633.AA20893@mail.kcwc.com> Date: Wed, 20 Aug 1997 00:32:51 -0700 (PDT) Organization: Atlas Telecom From: Simon Shapiro To: (Curt Welch) Subject: Re: expected performance with DPT RAID-5... Cc: scsi@FreeBSD.ORG, Tom Samplonius Sender: owner-freebsd-scsi@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk Hi Curt Welch; On 19-Aug-97 you wrote: > Tom wrote: > > "cd" to one of the RAID-5 file systems and do a "dd > > if=/dev/zero of=testfile bs=64k count=1000". That will > > create a file called "testfile" 64MB in length (64k x > > 1000), and tell you how long it took to write it. > > > Then reboot (the only way I know of clearing the FreeBSD > > file cache, and the DPT cache), and do a "dd if=testfile > > bs=64k of=/dev/null" to read the file back. > > > On the write test I get about 3500000 bps, and the > > write test about 9200000 bps. These results don't seem > > very good too me. > > On bus 0 I have fast/wide drives but I currently have > the speed limited to 5Mhz (slow/wide?). > > On bus 1 I have ultra/wide drives and the bus is currently > limted to 5Mhz (trying to work out SCSI problems), but I > put it back to 20 Mhz for some of these tests. > > Numbers are all bytes per second (output from dd). > > On write, I see 1,800,000 on bus 0 at 5 Mhz > 2,500,000 on bus 1 at 20 Mhz. > 2,400,000 on bus 1 at 5 Mhz Read the original Berkeley papers on RAID-5 and then the DPT implementation notes. > > On read, I see 6,300,000 on bus 0 at 5 Mhz. > 5,400,000 on bus 1 at 5 Mhz. > > This is with 2.2.2-RELEASE and the DPT 1.1.10 driver with > the following options set: > > options DPT_USE_SINTR=1 > options DPT_MEASURE_PERFORMANCE > options DPT_TRACK_CCB_STATES ^ If you do not have crashes and other nasties, disable ths option. > options DPT_HANDLE_TIMEOUTS > options DPT_COMMAND_SPLHIGH > options DPT_INTR_CHECK_SOFTC Definitely disable this option. As a matter of fact, get 1.2.1 and test again. > > The ultra-wide drives are Quantum Vikings. I also have one > of these drives connected to an Adaptec 9240UW card in the > same system. Your test on that drive produces: > > On write, I see 8,400,000 on AHA at 20 Mhz > > read is about the same - but I don't want to reboot again to > do a "clean" no cache test. I suspect it would produce about > the same number however. > > This test: dd if=/dev/rsd1c bs=8k of=/dev/null count=10000 > > produces a rate of 9.3 MB/S on the AHA. > > On the DPT (bus 1 at 5Mhz), it produces a rate of 4.0 MB/S. > > I'm not over impressed with these numbers for the DPT, but > then I'm still dealing with unknown SCSI problems with this > set-up. You are comparing apples and oranges. RAID-5 is not RAID-0, nor is it a streight disk. There is no free lunch. Also, running the bus at 5MHz is not quite as fast as 20. Simon P.S. Let me try and help you with these problems. Call me if email is not the solution. 503.799.2313 PDT, my day starts at 1000 :-) Simon