Subject: Re: RiscBSD News Server
To: Duncan Ferguson <d.ferguson@worc.ac.uk>
From: Mark Brinicombe <amb@physig.ph.kcl.ac.uk>
List: port-arm32
Date: 01/13/1997 11:56:26
On Mon, 13 Jan 1997, Duncan Ferguson wrote:

> I seem to remember that SCSI disks would be faster overall than IDE 
> in most cases, although they are more expensive.  Since an IDE podule 
> could only take 2 disks extra, whereas SCSI can go up to 7 i would 
> have thought SCSI better, but then than depends again on the disks i 
> get - 2x4.5g or 1x9g.  Can two smaller disks be meshed together as 
> suggested?  The way you said 'several IDE disks' above seems like it 
> can, but i am a novice to UNIX (although i have a good book to use as 
> reference) and there is no mention in my book on single partitions 
> over multiple disks.

Ok the NetBSD ccd device should work fine under RiscBSD although I have 
not had the opertunity to test this yet. (The root partition cannot be 
striped as it is needed to get the ccd config).

These days IDE is also quite fast. OK Fast SCSI is faster but both IDE 
and SCSI2 can go faster that the max xfer rate across the podule bus.

Try man ccd for more info on the ccd device.

The card is going to be the important bit. Both the cumana and powertec 
cards will get fast when DMA gets implemented.
The internal IDE and IDE podules (ICS,simtec etc.) will provide 1MB/sec 
xfer rates. This is limited by these cards being 16 bit PIO interfaces.
Currently the only faster IDE is the RapIDE which is not yet supported.
However from conversations I have last week RapIDE a high performance 
RapIDE driver *may* become a reality very soon.
 
Cheers,
				Mark

Mark Brinicombe				amb@physig.ph.kcl.ac.uk
Research Associate			http://www.ph.kcl.ac.uk/~amb/
Department of Physics			tel: 0171 873 2894
King's College London			fax: 0171 873 2716