On Mon, Apr 30, 2001 at 10:23:57AM +0200, wojtek@3miasto.net wrote:
> > > > any HDA being manufactured, and the interface speed is just kind of
> > > > silly.
> > > 
> > > unless you put 10 drives on one SCSI-3 bus. 
> > 
> > Sure.  And for half the price, I can put twice as much IDE storage on
> > the same machine, even if I have to add some plugin IDE controllers.
> > And for many applications, it will be faster.
> 
> unfortunately no. the ware some posts about performance problems with many
> IDE disks even on separate controllers. it's because IDE chip saturates
> completely PCI bus when transferring data so making CPU stalls when it
> tries to send command to another IDE chip

That's simply not the case for any bus-mastering IDE controller I've
tested.  Perhaps there is some lame implementation, or perhaps the post
in question is referring to a chip running non-bus-mastering mode, but
then I could complain about all those NCR 53c80 controllers that hog
the CPU, too.

I encourage you to post SPECIFIC numbers on SPECIFIC machines with
SPECIFIC benchmarks.