Subject: Re: ATAPI CD performance (update)
To: None <Federico_Lupi@www.datasys.it>
From: Manuel Bouyer <bouyer@antioche.lip6.fr>
List: port-i386
Date: 10/14/1999 21:07:30
On Thu, Oct 14, 1999 at 11:17:58AM +0200, Federico_Lupi@www.datasys.it wrote:
> Are you using 'cd..' or 'rcd..' for your tests?
> 
> I made some more tests with NetBSD 1.4.1
> 
> ASUS 40x:
>      dd if=/dev/cd0d  of=/dev/null bs=32k   ==>  360 K/s
>      dd if=/dev/rcd0d of=/dev/null bs=32k   ==> 4000 K/s
> HP 24x:
>      dd if=/dev/cd1d  of=/dev/null bs=32k   ==>  860 K/s
>      dd if=/dev/rcd1d of=/dev/null bs=32k   ==> 2600 K/s
> 
> I wonder why rcd0d is an order of magnitude faster than cd0d.

Hum maybe because the block device will issue several 2k commands for a
32k read() ? Just a though ...

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--