On Wed, 7 Aug 2002, Wojciech Puchar wrote:

> > 	Just a datapoint. Switching a Dell poweredge 350 (entry level
> > 	celeron 850) running RAIDFRAME on two 111GB disks from both
> > 	disks on the same channel to individual channels changed the
> > 	bonnie++ numbers as below:
> >
> > Version  1.01       ------Sequential Output------ --Sequential Input- --Random-
> >                     -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
> > Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
> > same-channel   300M  9239  31 10083  18  4019   7  7402  29  8841   8  89.2   3
> > separate-chan  300M  9441  31 10214  18  5370  10  8271  34 10943  11 140.0   4
>
> quite large difference excluding writes, where disk write buffers accept
> data immediately.
>
> anyway 8-10MB/s is similar to my Pentium 133 with old 20GB seagate.
> is your system well configured at all?

	Dell cut all the corners and equipped the PE350 with a UDMA2
	disk interface (See my "Dell PE350 third DIMM slot and NetBSD"
	for other PE350 fun).

	The raw disk gives 27.4MB/s read speed, and the raw RAID gives
	25.3MB/s. What does bonnie++ gives on your box? (the above
	bonnie++ numbers were with 128MB)


-- 
		David/absolute		-- www.netbsd.org: No hype required --