Date: Wed, 11 Nov 1998 12:15:51 +0000 From: Tony Finch <dot@dotat.at> To: hackers@FreeBSD.ORG Subject: Re: SCSI vs. DMA33.. Message-ID: <E0zdZBf-0003dj-00@fanf.noc.demon.net> In-Reply-To: <19981111162000.O20374@freebie.lemis.com> References: <98Nov11.134648jst.21907@ns.isi.co.jp> <98Nov11.134648jst.21907@ns.isi.co.jp>
next in thread | previous in thread | raw e-mail | index | archive | help
Greg Lehey <grog@lemis.com> wrote:
>
[...]
>They say, though, that SCSI drives work better with multiple requests
>outstanding...
>
>$ dd if=/dev/rsd0c bs=32k count=1000 of=/dev/null & dd if=/dev/rsd1c bs=32k count=1000 of=/dev/null
[...]
>I must say, I'm surprised. This makes it look like there's more of a
>performance hit with concurrent requests on SCSI than on IDE.
[...]
Our canonical examples of applications that benefit from SCSI is a web
server or proxy cache, because they tend to exercise the disks'
ability to do random seeks. Given that disks frequently lie through
their teeth about their geometry and do things like transparent bad
block remapping, letting the disk do its own request scheduling is
probably a Good Thing. Obviously you have to have a controller that
can queue up a decent number of outstanding requests so care is
required when looking at the specs...
Sorry, I don't have performance numbers to hand right now.
Tony.
--
gg yhf**f.a.n.finch
dot@dotat.at
fanf@demon.net
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?E0zdZBf-0003dj-00>