On Wed, Jun 20, 2001 at 11:23:40PM +0200, Markus W Kilbinger wrote:
> >>>>> "Manuel" == Manuel Bouyer <bouyer@antioche.lip6.fr> writes:
> 
>     >> Attachment!? Couldn't find one...
> 
>     Manuel> Ops, sorry. Here it is
> 
> Ok... 'time ./tst /dev/rwd0d 2000' was the call, rigth?
> 
> Hmm, I get different results ranging from
> 
>  # time ./tst /dev/rwd0c 2000
> 
>  real    0m2.842s
>  user    0m0.000s
>  sys     0m0.663s

This would be 46MB/s, meaning the bus runs at Ultra/66
> 
> and
> 
>  # time ./tst /dev/rwd0d 2000
>  
>  real    0m17.274s
>  user    0m0.000s
>  sys     0m0.667s
> 
> The fastest counts?

Was the  machine idle at the time you tested ?
This measure bandwith between the disks's cache and host memory. This is
only valid if there isn't any other disk access.

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