25 Feb
2021
25 Feb
'21
10:43 a.m.
Oh this is fun stuff. Is there a specification write-up anywhere on the
MASSBUS overall?
For example I wonder if the RH70 could do a transfer >128kb at a time.
Another is around the RH11: There were two models, the traditional RH11
(which could only do so many words on a DMA transfer) and the RH11-C
which could do more words per transaction by basically running the
Unibus in "Hog mode". That allowed the 2020 to run RM03's at a full 3600
RPM (and I assume can allow the 2020 to run things like RM80's).
Another item I always wondered about was the RH11's support of two
unibuses. I think the idea was to do the data transfers on the second
bus right to an 11/45's FASTBUS memory without worrying about DMA
timeouts while running the control and status registers on the normal
UNIBUS (which wouldn't block other devices). I wonder, does the MBA on
an 11/70 use similar cards to an RH11?
C
On 2/25/2021 1:30 PM, Noel Chiappa via cctalk wrote:
From: Paul
Koning
There's a good reason why the big disks on
many DEC machines were Massbus
devices until MSCP arrived. It's quite clear on Unibus PDP-11s, which
needed Massbus both for speed and for a cleaner answer to more-than-18
bit addressing.
I follow the first sentence, but I'm confused by the second, especially "a
cleaner answer to more-than-18 bit addressing". The UNIBUS MASSBUS
controller/adapter, the RH11, only has 18-bit addressing on the main memory
side. It does have more than 18-bit addressing on the device side, but so does
the RP11 (sort of). Are you thinking of the RH70? That does have access to
more than 2^18 bytes of main memory, but that's because it connects to the
-11/70 memory bus (as well as the UNIBUS, which is only used for control, not
data).
Similar questions about the speed point; passing data through an RH11 doesn't
increase the speed of the UNIBUS? Yes, the RH70 is faster, but that's because
of its connection to the -11/70 memory bus.
Noel