27 Nov
2014
27 Nov
'14
8 a.m.
From: Phil Budne
I think MIT-MC was a "1080" and only had
RH10s.
That sounds right.
I think they used a DL10 for PDP11 interfacing
(including chaosnet and
some disks)
I don't recall anymore, but I think there was only one -11 attached to the
DL10 (couldn't they support up to 4?)? And yes, the Tridents (along with,
IIRC, some serial lines, and the CHAOSNET) were attached to that -11.
and ran KLDCP (pre RSX20) on the console front end....
Yup. Although it had been hacked locally. After briefly glancing at the
sources (see the KLDCP; directory, and IOELEV > in SYSTEM;), it was a real
kludge; the IOELEV program ran in _both_ the DL10 -11, and the console -11.
(In the latter case, along with KLDCP!!) There was some stuff attached to the
console -11 too, apparently; a bunch of serial lines on DH11's.
ISTR a disk farm labeled "Moon's
Laundromat"
Yes, but I don't think that was on MC. I vaguely recall that it was on a
bunch of drives which were mounted in vertical pairs, which must have been
CalChomps, which I think must have been the original AI? (IIRC, both ML and
DM had DEC RP drives.)
Noel
2 Dec
2 Dec
3:21 a.m.
New subject: "There's no need to resurrect MC" (KL10 ITS)
?There?s no need to resurrect MC? says the klh10 source, but since when has that ever
stopped me from doing something?
Is there some other reason I *can?t* resurrect KL10 ITS by adding the relevant missing
parts to klh10? I think all I?d have to add is ITS paging in the KL version, the RH10, and
simulation of the IO PDP11 and stuff that was attached to it. There?s references to model
B hardware in UCODE; (Did MC become a model-B KL at some point?) so I might not have to
deal with model-A vs. model-B differences. Or I can simply assume it was intended to run
ITS on model B hardware and hack in support where required.
7:06 a.m.
New subject: "There's no need to resurrect MC" (KL10 ITS)
Daniel Seagraves wrote:
(Did MC become a model-B KL at some point?) so I might
not have to deal with model-A vs. model-B differences. Or I can simply assume it was
intended to run ITS on model B hardware and hack in support where required.
I'm 99.44% sure MC was born, and stayed a "model A"
ISTR someone looking at putting ITS support into the KL model B
microcode (I can't remember if it was Alan Bawden at MIT, or Stu
Grossman while at DEC). They decided not to, but I can't speak to why.
The BIG difference between "A" and "B" is that the model B implements
Extended Addressing. Section zero of an "Extended" KL should be
largely identical to a non-extended KL. Later on One Word Global byte
pointers were made to work in section zero.
3672
days inactive
3677
days old
2 comments
3 participants
participants (3)
-
dseagrav＠lunar-tokyo.net -
jnc＠mercury.lcs.mit.edu -
phil＠ultimate.com