From: "Rick Bensene" <rickb at
bensene.com>
There was dedicated hardware in the display console
that did CDC
character set (a 6-bit code) conversion to vector characters. Vector
graphics were possible, within the limitations of the speed of the PPU.
On later 6x00-series systems, such as the CYBER-73, the PPUs ran fast
enough to generate a nice looking all-vector chessboard on the left
screen, and a text-based transcript of the moves on the right screen.
There were also a number of other cute programs, one being a pair of eyes
(one on each screen) which would look around and blink.
I think you probably mean the CYBER 173 PPU--the CYBER 7x PPUs were
essentially unchanged from the 6x00 series. 1 usec. cycle time. The
character generator, IIRC, was an interesting array of LRC circuits,
with the character size varied by simply increasing the amplitude of
the drive waveform. This led to odd-shaped large characters.
Chess was an oddball program. It was possible for a user program to
request control of the display from the operator. There were two
ways to do this. The first was to write your own display driver and
request the display channel from DSD (DSD monitored R.RCH and R.DCH
requests). An early version of chess I saw used a driver called (I
think) CHD for this.
There was also what was called the "T" display in DSD. This was a
display mode that allowed a user-mode CP program provide a list of
display directions in CM to be read by DSD. One of the directions
was "jump", so the user program could allow DSD to regenerate the
display without having to made additional system requests (which
would have been too slow). I saw a later version of CHESS that used
this method. The restriction on the T display was that it could
generate a display for one screen or the other, but not both
simultaneously.
There was a very popular "programmer's display" driver called DIS
that could attach to a control point and allow for job control entry,
exchange package display, and numerous dump formats as well as a
disassembler and CP breakpoint.
There was a keypunch simulator called O26 (note the alpha "o") to
allow the enterprising operator or programmer to generate a file of
card images. A very expensive keypunch, indeed.
The operating > system was called KRONOS, and I
clearly remember that
the console command > to run the "eye" program was "X.EYES". >
Some
privileged users could > remotely fire off this EYES program from a
terminal, and it'd make for a bit of a surprise, especially when we
gave tours of the computer center to schoolchildren.
KRONOS was a direct descendent of MACE, an alternative OS for the
6x00 written largely by Dave Callender and Greg Mansfield as a
bootleg project at Arden Hills. Greg and Dave would visit the QA
floor and cable together a system to do development on late at night,
after the daytime shift had gone home. (MACE, if I remember
correctly, stood for "Mansfield's answer to Customer Engineering".
The "eye" display was part of every CE's deadstart (not
"coldstart")
tape.
The "official" OS for the 6000 series was SCOPE--a system that was
located largely in the PPUs, and pretty much oriented toward batch
processing.
Pat might want to investigate Dave and Greg's contribution to the OS
that Purdue ran on their 6500 in the late 60's. I seem to recall
that it was largely MACE with a lot of OS functionality moved into
the CPMTR portion.
A cultural war broke out in CDC over the two systems. SCOPE was
pretty much a west coast Sunnyvale operation and KRONOS (i.e.
enhanced MACE) an Arden Hills affair. Things really got pitched when
PLATO elected to use KRONOS instead of SCOPE. The solution was to
offer both as renamed entities. SCOPE 3.4 was followed by NOS/BE 1.0
(BE for batch environment) and KRONOS became NOS 1.0. A lousy move
on the part of management, in my opinion and a huge duplication of
resources. I have no particular ax to grind here, as I was Sunnyvale
Ops, but Greg was a good friend, too.
This doesn't include the "special" operating systems such as TCM,
Zodiac, Rover, etc.
One very positive thing that I can say about MACE was that Dave
authored a set of fairly straightforward coding standards and adhered
to them. So MACE systems code was easy to read; SCOPE was an
agglutinative affiar with no particular coding standard being used
anywhere.
I had the distinct privilege of being a systems
operator on a Tektronix
Cyber 73 (which was a derivative of the 6600).
The 73 was basically a 6400 with added goo, such as the Compare/Move
Unit; the 74 was basically a 6600 (wherein a CMU was not possible).
There was a Cyber 72, which was a slowed-down (via jumper) 6400. The
Cyber 76 was the follow-on to the 7600. This is from recollection,
but I'm pretty sure that I'm very close. The Cyber 7x's had the fake
wood-and-blue-glass look, while the 6000 series was dark grey and
beige-ish.
The machine had an array of toggle switches called the
"Coldstart Panel"
in one of the CPU bays into which a small (I think it was something like
twelve twelve-bit words) PPU bootstrap program could be toggled in. The
system could be booted from one of the "washing machine" drives, which was
the usual mode of startup, or from magtape.
You had an unusual shop. At Sunnyvale, deadstart was almost always
from tape. The disk drives could be anything from the
old 6603
Bryant disk, to a big 808 (4-spindle hydraulic positioner, bit-
parallel drive) to one of the "little" 844 removable pack drives.
I know--more than anyone asked for. But questions about old iron are
getting less frequent.
Cheers,
Chuck