Date: 10 Nov 1997 07:22:44 GMT From: chris@netmonger.net (Christopher Masto) To: freebsd-chat@freebsd.org Subject: IDE performance - benchmark numbers Message-ID: <646co4$n3l$1@schenectady.netmonger.net>
next in thread | raw e-mail | index | archive | help
I got around to reading the documentation, so now I know about the wd flags. I thought I'd try a little experiment. First, here's what I tested this on: (built from sources supped on Nov 5) Copyright (c) 1992-1997 FreeBSD Inc. Copyright (c) 1982, 1986, 1989, 1991, 1993 The Regents of the University of California. All rights reserved. FreeBSD 3.0-CURRENT #0: Sun Nov 9 21:33:37 EST 1997 chris@kaboom.masto.com:/usr/src/sys/compile/KABOOM CPU: Pentium (132.96-MHz 586-class CPU) Origin = "GenuineIntel" Id = 0x52c Stepping=12 Features=0x1bf<FPU,VME,DE,PSE,TSC,MSR,MCE,CX8> real memory = 33554432 (32768K bytes) avail memory = 30183424 (29476K bytes) Probing for devices on PCI bus 0: chip0: <Intel 82437FX PCI cache memory controller> rev 0x02 on pci0.0.0 chip1: <Intel 82371FB PCI to ISA bridge> rev 0x02 on pci0.7.0 ide_pci0: <Intel PIIX Bus-master IDE controller> rev 0x02 on pci0.7.1 bktr0: <BrookTree 848> rev 0x11 int a irq 11 on pci0.9.0 Hauppauge WinCast/TV, Philips NTSC tuner, dbx stereo. vga0: <VGA-compatible display device> rev 0x00 int a irq 10 on pci0.11.0 Probing for PnP devices: CSN 1 Vendor ID: CTL0044 [0x44008c0e] Serial 0x001274d8 CSN 2 Vendor ID: TCM5094 [0x94506d50] Serial 0x24d36110 Probing for devices on the ISA bus: sc0 at 0x60-0x6f irq 1 on motherboard sc0: VGA color <16 virtual consoles, flags=0x0> sio0 at 0x3f8-0x3ff irq 4 flags 0x10 on isa sio0: type 16550A sio1 at 0x2f8-0x2ff irq 3 on isa sio1: type 16550A lpt0 at 0x378-0x37f irq 7 on isa lpt0: Interrupt-driven port lp0: TCP/IP capable interface psm0 at 0x60-0x64 irq 12 on motherboard psm0: device ID 0 fdc0 at 0x3f0-0x3f7 irq 6 drq 2 on isa fdc0: FIFO enabled, 8 bytes threshold fd0: 1.44MB 3.5in wdc0 at 0x1f0-0x1f7 irq 14 flags 0xa0ffa0ff on isa wdc0: unit 0 (wd0): <Maxtor 82560 A4 ->, 32-bit, multi-block-16 wd0: 2442MB (5001696 sectors), 4962 cyls, 16 heads, 63 S/T, 512 B/S wdc0: unit 1 (wd1): <WDC AC33100H>, 32-bit, multi-block-16 wd1: 3020MB (6185088 sectors), 6136 cyls, 16 heads, 63 S/T, 512 B/S wdc1 at 0x170-0x177 irq 15 flags 0xa0ffa0ff on isa wdc1: unit 0 (atapi): <FX600S/P01>, removable, intr, dma, iordis wcd0: 1033Kb/sec, 256Kb cache, audio play, 255 volume levels, ejectable tray wcd0: no disc inside, unlocked 16 3C5x9 board(s) on ISA found at 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 0x200 ep0 not found at 0x300 npx0 on motherboard npx0: INT 16 interface sb0 at 0x220 irq 5 drq 1 on isa snd0: <SoundBlaster 16 4.13> sbxvi0 at ? drq 5 on isa snd0: <SoundBlaster 16 4.13> sbmidi0 not found at 0x330 awe0 at 0x620 on isa AWE32 not found AWE32: not detected I know IOZONE isn't a great measure of performance, but it was the first to download, so here are the numbers I got from it. This is all with a 64MB test file, since it says to make it at least twice the available RAM. There was a comment made here that mode 4 isn't a good idea, so I tried it with mode 2 for comparison. (I didn't see any messages about DMA, so I'm wondering if my controller actually supports it). Anyway: MODE BLOCK SEC-WRITE SEC-READ MB/S-W MB/S-R 2 512 35.7 31.3 1.8 2.1 2 4096 29.6 29.5 2.3 2.3 2 10240 29.2 27.9 2.3 2.4 4 512 15.5 142.6 4.3 0.5 4 4096 12.2 141.8 5.5 0.5 4 10240 12.3 142.0 5.5 0.5 Obviously, there's something incredibly wrong with the read performance. IMHO, all of the numbers are pretty lousy. Just for comparison, I set the flags back to 0x0 and got these results: MODE BLOCK SEC-WRITE SEC-READ MB/S-W MB/S-R 4 512 21.6 17.7 3.1 3.8 4 4096 15.7 15.6 4.3 4.3 4 10240 15.2 16.0 4.4 4.2 Conclusion: I don't know. There doesn't seem to be an obvious combination that gives the best results all around. I still suspect that an Adaptec 2940UW and a nice SCSI drive will be a lot better. I don't have an apples-to-apples comparison, though.. the closest is a P166 with the same amount of RAM (32MB) and an Adaptec 2940UW: Writing the 64 Megabyte file, 'iozone.tmp'...8.453125 seconds Reading the file...7.070312 seconds IOZONE performance measurements: 7938941 bytes/second for writing the file 9491640 bytes/second for reading the file Anyway, now there are some numbers to argue over. -- = Christopher Masto = chris@netmonger.net = http://www.netmonger.net/ = = NetMonger Communications = finger for PGP key = $19.95/mo unlimited access = = Director of Operations = (516) 221-6664 = mailto:info@netmonger.net =
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?646co4$n3l$1>