On 2012-12-30 19:00, Christian Gauger-Cosgrove<captainkirk359 at gmail.com>
wrote:
On 29 December 2012 21:08, Toby Thain<toby at
telegraphics.com.au> wrote:
>Does that follow? PMI is a non-bus (private!)
ribbon cable jumper from CPU
>to RAM.
>
You're mistaking uVAX memory with PDP-11 memory. uVAXen have the
ribbon cables, and the CD slot. On the PDP-11 processors and RAM (and
QBUS to UNIBUS adapter -- KTJ11, I think it is called -- in the 84 and
94) the PMI is all on only the CD slots of the board.
Correct, PMI memory on PDP-11 machines is just communicating on the CD
slots. No ribbon cables.
However, I'd like to point out that the 11/83 also can use PMI memory.
The thing to understand is that for the memory to be connected on the
PMI, it needs to set *before* the CPU on the Qbus, and be PMI capable.
The same memory placed *after* the CPU means it acts as normal Qbus
memory. Some difference in performance.
When sitting in the 11/84, the memory sits after the CPU, but all the
first four slots are PMI slots in that box, always.
Actually, PMI is also not only on the CD slots. A few signals in the AB
slots are working differently than in a Qbus as well.
And to point out one last, obvious thing, the PMI on the VAX is a bus,
even through it's on a ribbon cable. Why would that make it not a bus?
Johnny