[pups] Unix and PDP11/20 (was PDP9?)

[pups] Unix and PDP11/20 (was PDP9?)

[Carl Lowenstein]

> From: Dennis Ritchie <dmr at plan9.bell-labs.com>
> Subject: [pups] Unix and PDP11/20 (was PDP9?)
> Date: Wed, 4 Sep 2002 00:21:40 -0400
> 
> I'm not positive about the logo on our first PDP-11.

The text accompanying the picture "Ken and Den" somewhere on your web site
says that the logo on your first PDP-11 was just "PDP-11" without the /20.

> On the earliest handbook I have, the front panel photo
> just shows "PDP11", though inside the handbook
> it does talk about the two models (11/10 and 11/20).
> Both had the same KA11 processor, but the basic
> 11/10 sported 1024Kw ROM memory plus a generous
> 128 words of RAM, while the 11-20 had
> 4096Kw core RAM, and the ASR33 Teletype was included.
> You could add more RAM to the 11/20.

I fear that you have suffered a "units slip" saying 1024Kw
and 4096Kw when you meant 1Kw and 4Kw respectively.

> Incidentally, the machine's handbook was a wonder.
> In 104 pages (each 5.25x8 inches), it described the whole
> system: not only the instruction set but the theory
> of the Unibus (including some logic diagrams) together with
> programming specifications for the TTY, the clock,
> and the paper tape reader.
> 
> 	Dennis

Agreed, "PDP11 Handbook Second Edition" was a really good book
Occasionally I wonder if I ever had my hands on a "First Edition"
and threw it away when the second edition came out.  Not knowing
that both the computer and the handbook would become classics.

    carl
-- 
    carl lowenstein         marine physical lab     u.c. san diego
                                                 clowenst at ucsd.edu

[pups] Unix and PDP11/20 (was PDP9?)

[Jeffrey Sharp]

On Wednesday, September 4, 2002, Johnny Billquist wrote:

> On Wed, 4 Sep 2002, Dennis Ritchie wrote:
>
> > I'm not positive about the logo on our first PDP-11.
>
> Early PDP-11/20 just said PDP11 on the front, I believe.

Definitely. My 11/20 (SN #786) used to have one of the earlier panels
without the "/20". That panel was damaged and subsequently replaced with the
newer version that did include the "/20".

-- 
Jeffrey Sharp

The email address lists at subatomix.com is for mailing list traffic. Please
send off-list mail to roach jss at wasp subatomix beetle dot com. You may
need to remove some bugs first.

[pups] Unix and PDP11/20 (was PDP9?)

[John Holden]

> Dennis Ritchie wrote:-
>
> Early on, for fun, we tried assembling the DEC-supplied
> assembler, which came on at least one (maybe more) long
> fan-folded paper tapes.  I don't think we ever succeeded; it had to
> be fed in twice for the two passes, and enough characters
> were dropped that phase errors occurred.

The early high speed tape readers used the clock pulse off the stepper motor
(with suitable delay) to strobe the data from the photo-transistors. It
was somewhat unreliable. Latter models added a ninth detector under the sprocket
holes, which being smaller, neatly strobed the data in the middle of the punched
data.

The 11/20 I first used only had the paper tape software. The pdp-11 instruction
set (at that point) was nicely orthogonal, so we often hand coded patches rather
than use the assembler/editor, since it was faster.

[pups] Unix and PDP11/20 (was PDP9?)

[Carl Lowenstein]

> From: Dennis Ritchie <dmr at plan9.bell-labs.com>
> Subject: [pups] Unix and PDP11/20 (was PDP9?)
> Date: Fri, 6 Sep 2002 02:08:02 -0400
> 
> Holden's link,
> 
> http://www.psych.usyd.edu.au/pdp-11/11_20.html
> 
> reinforces my guess that our first -11 probably did
> have just "PDP11" on the bezel.  The one in my photo
> (which has the 20) is doubtless our second -11.
> I've looked at this page before, but it slipped my mind.
> 
> Our first -11 was very early, and its disk took several
> months to arrive: it had TTY33 and high-speed paper tape
> as its only peripherals besides the clock.
> 
> Early on, for fun, we tried assembling the DEC-supplied
> assembler, which came on at least one (maybe more) long
> fan-folded paper tapes.  I don't think we ever succeeded; it had to
> be fed in twice for the two passes, and enough characters
> were dropped that phase errors occurred.

I was there once myself.  The problem was fuzzy holes in the DEC-punched
fan-fold paper tape.  So I toggled in a small utility program
"wait, read, wait, punch, loop" to copy from the TTY reader to the
high-speed punch.  The sensing pins of the TTY had no trouble with the
fuzzy holes, and I got paper tapes that worked in the high-speed reader.

    carl
-- 
    carl lowenstein         marine physical lab     u.c. san diego
                                                 clowenst at ucsd.edu

[pups] Unix and PDP11/20 (was PDP9?)

[Dennis Ritchie]

Holden's link,

http://www.psych.usyd.edu.au/pdp-11/11_20.html

reinforces my guess that our first -11 probably did
have just "PDP11" on the bezel.  The one in my photo
(which has the 20) is doubtless our second -11.
I've looked at this page before, but it slipped my mind.

Our first -11 was very early, and its disk took several
months to arrive: it had TTY33 and high-speed paper tape
as its only peripherals besides the clock.

Early on, for fun, we tried assembling the DEC-supplied
assembler, which came on at least one (maybe more) long
fan-folded paper tapes.  I don't think we ever succeeded; it had to
be fed in twice for the two passes, and enough characters
were dropped that phase errors occurred.

Incidentally, B programs could be run on this first pre-disk
-11, using cross-compilation from GECOS.  There was
a stand-alone predecessor of dc!

BTW, apologies for the units slip in the earlier posting.
Indeed 128 words of RAM on the 11/10, 4096 words
standard on the 11/20 (we splurged with 12K).

Also BTW, the young woman on p. 104 of the first
manual has a just-so-1969 hairdo!  She has her
index finger on one of the console switches, is
holding a Unibus jumper in the other hand, and
the caption is "The PDP-11 provides Direct Device
Addressing...."  The Unibus address pin assignments
that replaced herwere probably more useful, but not
so redolent of history.

	Dennis

[pups] Unix and PDP11/20 (was PDP9?)

[John Holden]

On Wed, 4 Sep 2002, Dennis Ritchie wrote:-

> I'm not positive about the logo on our first PDP-11.

The picture of Dennis and Ken in front ASR-33's hints at a pdp11/20 logo
on the console. The 11/20 I have (built 29/1/71, SN 821) has just plain 'pdp11'.
The lead time on getting the machine was about 6 months. I suspect that the
/20 was added as other models were in the pipeline (/05,/45).

For a picture, see http://www.psych.usyd.edu.au/pdp-11/11_20.html

> Incidentally, the machine's handbook was a wonder.

Indead it was

The front cover is interesting, in that it shows a table top version of the
11/20. It was quite possible to run one just with paper tape and an ASR33 with
the reader/punch option. There was a similar option for the pdp8/e.

The only obvious change between the first and second edition handbooks was that
the latter changed the last page from a picture of a young lady in front of
a machine to a table of Unibus pin assignments.

[pups] Unix and PDP11/20 (was PDP9?)

[Johnny Billquist]

On Wed, 4 Sep 2002, Dennis Ritchie wrote:

> I'm not positive about the logo on our first PDP-11.
> On the earliest handbook I have, the front panel photo
> just shows "PDP11", though inside the handbook
> it does talk about the two models (11/10 and 11/20).

Early PDP-11/20 just said PDP11 on the front, I believe.

	Johnny

Johnny Billquist                  || "I'm on a bus
                                  ||  on a psychedelic trip
email: bqt at update.uu.se           ||  Reading murder books
pdp is alive!                     ||  tryin' to stay hip" - B. Idol

[pups] Unix and PDP11/20 (was PDP9?)

[Dennis Ritchie]

As Billquist, Wilson and others pointed out, the first
versions of Unix did run on machines with no memory
mapping or protection, and indeed context-switching
was accomplished by swapping.  By 1973 we had the luxurious
11/45, to considerable relief.

I'm not positive about the logo on our first PDP-11.
On the earliest handbook I have, the front panel photo
just shows "PDP11", though inside the handbook
it does talk about the two models (11/10 and 11/20).
Both had the same KA11 processor, but the basic
11/10 sported 1024Kw ROM memory plus a generous
128 words of RAM, while the 11-20 had
4096Kw core RAM, and the ASR33 Teletype was included.
You could add more RAM to the 11/20.

Incidentally, the machine's handbook was a wonder.
In 104 pages (each 5.25x8 inches), it described the whole
system: not only the instruction set but the theory
of the Unibus (including some logic diagrams) together with
programming specifications for the TTY, the clock,
and the paper tape reader.

	Dennis