AGP, vs PCI, and throw in a little UPA was Re: [SunRescue] NetBSD v. Linux v. Solaris 2.51

[James Lockwood]

On Thu, 7 Dec 2000, Christopher Byrne wrote:

> Moving to 64 bit and 66 mhz helps a lot, increasing maximum bandwitdh to
> about 500 MB/s, but that is stil inadequate for the upcoming generation of
> I/O controllers with speeds into the Gigabyte per second (FC-AL is expected
> to be available at that speed within 18 months). What we really need is a
> revolutionary rather than evolutionary jump in peripheral and component
> interfaces. Some people are suggesting that serial I/O is the way of the
> future, but I don't find myself convinced.

Before we're capable of handling bulk I/O rates of that magnitude, memory
busses are going to have to speed up significantly.  There's little point
to being able to transfer data at that rate if you can't process it.  Some
significant lessons can be learned from mainframe design, distributing I/O
processing into front end processors and channel controllers allows a host
system to handle far faster transfers.

> Which machines have 120mhz UPA slots? The U80 only has 112.5 (of course it
> does have two of them), and I cant find UPA speed specs on all systems

All Sun systems which are "natively" PCI (U5/10/AXi/30/60/80/2x0/4x0...) 
support a 120MHz UPA.  UPA frequency is normally an integer fraction of
CPU clock speed, the U80 has a 112.5MHz UPA bus when in its high-end
configuration because it uses 450MHz CPU's with a 4x multiplier. 

Systems which run at 360MHz with a 3x multiplier (such as UltraSparc-IIi
boxes with Sabre 360/2MB's or UltraSparc-II boxes with Sapphire Black
360/4MB CPU's) will run the UPA at 120MHz.  Systems which run at 480MHz
with a 4x multiplier (such as UltraSparc-IIi boxes with Sapphire Red
480/2MB's) will also run the UPA at 120MHz.

I posted a summary of CPU/bus speeds to SunRescue about 6 months ago (I
believe), more information is in the archives.

-James