Date: Tue, 21 Mar 1995 01:32:52 -0600 (CST) From: Jim Bryant <jbryant@server.iadfw.net> To: kuku@gilberto.physik.rwth-aachen.de (Christoph P. Kukulies) Cc: freebsd-current@FreeBSD.org Subject: Re: iozone read vs. write Message-ID: <199503210732.BAA02285@news.iadfw.net> In-Reply-To: <199503201602.RAA02319@gilberto.physik.rwth-aachen.de> from "Christoph P. Kukulies" at Mar 20, 95 05:02:31 pm
next in thread | previous in thread | raw e-mail | index | archive | help
In reply:
> Date: Mon, 20 Mar 1995 17:02:31 +0100
> From: "Christoph P. Kukulies" <kuku@gilberto.physik.rwth-aachen.de>
> To: freebsd-current@freefall.cdrom.com
> Subject: iozone read vs. write
>
> iozone 13 ...
>
> Writing the 13 Megabyte file, 'iozone.tmp'...15.085938 seconds
> Reading the file...4.000000 seconds
>
> IOZONE performance measurements:
> 903589 bytes/second for writing the file
> 3407872 bytes/second for reading the file
>
> Does anone have an explanation why the number diverge to such an extent?
> (System is a 32 MB 486DX2/66 EIDE 2 Quantum 540 system)
caching. Try modifying the sources to change the number of iterations in
auto-mode. I find on a 64 meg system that the read and write times
converge to hardware speed when it hits a file size of 32 megs. The test
is essentially useless until you get past the cache, that is unless you
are measuring caching performance.
Jim
--
All opinions expressed are mine, if you | "I will not be pushed, stamped,
think otherwise, then go jump into turbid | briefed, debriefed, indexed, or
radioactive waters and yell WAHOO !!! | numbered!" - #1, "The Prisoner"
jbryant@server.iadfw.net, System administrator, Internet America
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199503210732.BAA02285>