Date: Sun, 8 Oct 1995 18:01:35 -0400 (EDT) From: "Jonathan M. Bresler" <jmb@kryten.Atinc.COM> To: Bruce Evans <bde@zeta.org.au> Cc: julian@ref.tfs.com, roberto@keltia.freenix.fr, bde@zeta.org.au, dennis@etinc.com, hackers@freebsd.org Subject: Re: VLB Disk Controllers Message-ID: <Pine.3.89.9510081749.G13099-0100000@kryten.atinc.com> In-Reply-To: <199510082145.HAA11073@godzilla.zeta.org.au>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 9 Oct 1995, Bruce Evans wrote:
> >> > transfer speed is 1.25286e+07 bytes/sec
>
> >heh?
> >how do you get 12MB/sec on a 10MB/sec SCSI bus?
well how about these numbers (dont sweat them, i have not yet
created the sd1sX devices)
./disklatency /dev/rsd1f
Command overhead is 11092 usec (time_4096 = 11098, time_8192 = 11104)
transfer speed is 6.88982e+08 bytes/sec
Aspen: {258} !!
./disklatency /dev/rsd1f
Command overhead is 6270 usec (time_4096 = 4104, time_8192 = 1938)
transfer speed is -1.891e+06 bytes/sec
Aspen: {259} !!
./disklatency /dev/rsd1f
Command overhead is 10991 usec (time_4096 = 11098, time_8192 = 11204)
transfer speed is 3.84637e+07 bytes/sec
Aspen: {260} !!
./disklatency /dev/rsd1f
Command overhead is 11065 usec (time_4096 = 11120, time_8192 = 11175)
transfer speed is 7.4424e+07 bytes/sec
>
> It's the relative transfer speed for transferring 4K blocks and 8K
> blocks. If 4K blocks are transferred at 5MB/sec and 8K blocks are
> transferred at 10MB/sec, then the relative speed is infinite :-).
> This shows that the test is biased. It prints the speed mainly as
> a sanity test. If the speed isn't close to the SCSI bus speed, then
> you shouldn't trust any of the results of the test.
>
> Bruce
>
Jonathan M. Bresler jmb@kryten.atinc.com | Analysis & Technology, Inc.
FreeBSD Postmaster jmb@FreeBSD.Org | 2341 Jeff Davis Hwy
play go. | Arlington, VA 22202
ride bike. hack FreeBSD.--ah the good life | 703-418-2800 x346
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.3.89.9510081749.G13099-0100000>