27 Feb
2020
27 Feb
'20
12:24 a.m.
Yes, I couldn't remember how the DTE was connected on the KL side other
than it did use PI.? It was unibus side that we had speculated about at
Columbia.? That theoretically could have been plugged into the KS
somewhere, but this was a conversation a long time ago with somebody who
was not an DEC FE.? It occurs that there may have been issues with 18
vs. 16 bit data paths.
I don't remember the maximum number of RH's being 4.? As far as I
recall, the maximum number of RH's for the KL is 8 and not four. Towards
end of life, slot 5 was occupied by the NI and slots 6 and 7 were
occupied by the CI which took two (I believe because of power
consumption reasons).? At Columbia, we kept our tape drives on slot 1
with the rest of the disks on slots 0, 2, 3 and 4.? Both the monitor
code and KLH10 assume a possibility for 8 RH20's.
I may have it backwards, but I thought it was Tops-10 that complained
about tape and disk devices being on the same channel. In any event,
that was nothing you wanted to do.? Having the disks and tape on
different channels allowed for concurrent I/O which meant you could
stream the tape drive.
The RH20 was different enough from the RH10 as to provoke a famous
saga--Tony in RH20 Land; you can pull this anonymously from VENTI2::.
I cannot for the life of me remember a blessed thing about a DL10.? Sigh...
On 2/26/20 6:10 PM, Robert Armstrong wrote:
You would
likely have never seen a DTE on a KS processor ?
The DTEs plugged into dedicated
slots on the KL. There's no way you could have ever seen one on another model CPU -
it's physically not possible. The KL could hold up to four DTEs, and the first one
was "special" in that it could poke around in the KLs microstore and internal
datapaths. The 11/40 CFE was connected to that one and that's how the KL was started
up. All the microcode was stored in RAM and immediately after a power on the KL was
little more than a big heater. The other three DTEs were more general purpose and were
used for communications interfaces.
There was, however, a DL10 for the KI that interfaced up to four PDP-11s to the I/O
and memory busses. Conceptually the DL10 was similar to the DTE, although I don't
know how close they were programmatically. Of course there was no equivalent to the CFE
on the KI and all four -11 ports on the DL10 were identical. And where as every KL had at
least one DTE for the CFE, the DL10 was strictly optional.
It's a similar story for the RH20s - they were dedicated options for the KL only.
A KL could have a maximum of 4 RH20s and every one needed at least two - one for disk and
one for tape. Although I believe on TOPS10 you could mix disk and tape on the same
MASSBUS - maybe, I'm not sure about that.
And likewise there was an RH10 MASSBUS controller for the KI which was similar to, but
not the same as, the RH20.
Bob