> PCI max speed is 133 MB/s I have read.  If I put
> another ATA card in my system so each drive has its
> own channel, would that reproduce a similar issue to
> the controllers limited speed, or are the issues with
> two drives on a single controller more controller
> based then bandwidth based?

There are several layers of bandwith here:

at the lowest layer, there is drive bandwidth, and that's limited simply by 
physics. There is nothing you can do about that, short of increasing the 
rotational speed of the disks (platters) inside of the drive, which of course 
isn't really feasible.

Then there is the matter of the (E)IDE/ATA bus width.

Finally, the (E)IDE/ATA bus is tied into your PCI bus, in some way or another. 
A 133MHz PCI bus can handle (8B per cycle * 133 million cycles per second) = 
1064000000B/s = 0.99GB/s transfers, assuming that this is a 64-bit PCI bus, 
hence the 8 Bytes per cycle.

So what does that tell us? Well, it tells us that, for your (E)IDE/ATA bus to 
be able to saturate the PCI bus, you'd have to have quite a bit of controllers 
with quite a bit of very fast ATA disks on them, pumping data at full speed the 
whole time.
 
 
This message posted from opensolaris.org
_______________________________________________
opensolaris-discuss mailing list
[email protected]

Reply via email to