Date: Wed, 3 Apr 1996 18:36:34 +0930 (CST) From: Michael Smith <msmith@atrad.adelaide.edu.au> To: luigi@labinfo.iet.unipi.it (Luigi Rizzo) Cc: msmith@atrad.adelaide.edu.au, bde@zeta.org.au, davidg@Root.COM, dutchman@spase.nl, freebsd-hackers@freebsd.org Subject: Re: HDD cpu usage (IDE vs. SCSI). Message-ID: <199604030906.SAA19932@genesis.atrad.adelaide.edu.au> In-Reply-To: <199604030812.KAA01334@labinfo.iet.unipi.it> from "Luigi Rizzo" at Apr 3, 96 10:12:34 am
next in thread | previous in thread | raw e-mail | index | archive | help
Luigi Rizzo stands accused of saying: >> Presuming you only have one application making requests in a linear fashion, >> that's fine. Tagged queueing and disconnect rapidly improves things >> once you start to get busy though. > > Disconnections are only useful to avoid locking the IO (SCSI) bus > during (implicit) seeks -- it's nothing different from getting an > interrupt when the transfer is complete. It allows you to issue commands to other devices, or to issue more commands to the same device. > Queueing requests _within_ the disk is useful if the OS does not > know the actual geometry of the disk, otherwise the OS can probably > do a more sensible work of reordering requests. Ah, so you want the system to know how each and every drive works, and to keep track of actuator position and velocity, as well as the rotational position of the media. Funny ha ha. The only part of the system in a position to make an _informed_ decision about which of several transactions is the easiest to perform next is the disk. With ZBR, hidden geometry and 'invisible' sector sparing, the OS doesn't have a hope. (Yes Terry, I know, RAID-X) > On a busy system with a single disk the only difference is really the > PIO overhead vs. the SCSI controller overhead (and many of them are > much slower than your main processor). On a busy system the ability to have multiple requests available for the disk to optimise and is a big win, and I'd say that at 5+M/sec even the $60 NCR controller is quite quick enough 8) PIO wins on _quiet_ systems where there's lots of free CPU. > Unless you want to say that SCSI supports many devices on the same bus, > which is certainly a big advantage. That too. Along with a standard that most vendors actually follow fairly closely. (Ever read any of the ATA specifications? It's not necessarily surprising that the devices are often total mutants...) > Luigi Rizzo Dip. di Ingegneria dell'Informazione -- ]] Mike Smith, Software Engineer msmith@atrad.adelaide.edu.au [[ ]] Genesis Software genesis@atrad.adelaide.edu.au [[ ]] High-speed data acquisition and (GSM mobile) 0411-222-496 [[ ]] realtime instrument control (ph/fax) +61-8-267-3039 [[ ]] Collector of old Unix hardware. "Where are your PEZ?" The Tick [[
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199604030906.SAA19932>