How the PDP-11 Was Born

Discussion:

(too old to reply)

Larry McGowan

2020-10-15 23:24:10 UTC

I just found this group and thought I would provide a link to a document I wrote some years ago about the birth of the PDP-11. In 1967 at DEC in Maynard, Massachusetts, I was the software manager for the paper tape software for the original PDP-11/20. I also wrote the first simulator, SIM-11, which ran on the PDP-6/10 and the first manual documenting the assembly language. Those were exciting times!

Here's a link to my document on how the PDP-11 was born:

http://hampage.hu/pdp-11/birth.html

also on my webpage at

https://www.dropbox.com/s/c9e6crm3rlo70re/pdp-11birth.pdf?dl=0

Regards,
Larry

Paul Rubin

2020-10-15 23:52:21 UTC

Permalink

Post by Larry McGowan
http://hampage.hu/pdp-11/birth.html

Thanks for the cool document! I remember that the 16-bit instruction
word could only encode a few two-address instructions (ADD, CMP)
with 6-bit (register # and mode) specifiers, and some of the others like
XOR had to make do with more constraints on the operands.

Maybe because of that, I heard claims that the -11 was originally
intended to be an 18-bit machine (conveniently a PDP-10 half-word), but
it then had to be shoehorned into 16 bits because that was what the
market wanted. With 18 bits the instruction encodings could have been
more uniform, and/or there could have been 16 registers instead of 8.

Was there any truth to that? Or was it 16 bits from the beginning? Did
the designers feel any pain over those issues? The later PDP-11-like
Texas Instruments MSP430 microcontroller had 16 bit words but 16
registers. I guess it did that through even more instruction encoding
hacks. I never programmed the MSP430 though, so don't remember any
specifics.

The MSP430 is still around, though somewhat fading under the ARM
juggernaut.

Don North

2020-10-15 23:56:40 UTC

Permalink

Post by Paul Rubin

Post by Larry McGowan
http://hampage.hu/pdp-11/birth.html

Thanks for the cool document! I remember that the 16-bit instruction
word could only encode a few two-address instructions (ADD, CMP)
with 6-bit (register # and mode) specifiers, and some of the others like
XOR had to make do with more constraints on the operands.
Maybe because of that, I heard claims that the -11 was originally
intended to be an 18-bit machine (conveniently a PDP-10 half-word), but
it then had to be shoehorned into 16 bits because that was what the
market wanted. With 18 bits the instruction encodings could have been
more uniform, and/or there could have been 16 registers instead of 8.
Was there any truth to that? Or was it 16 bits from the beginning? Did
the designers feel any pain over those issues? The later PDP-11-like
Texas Instruments MSP430 microcontroller had 16 bit words but 16
registers. I guess it did that through even more instruction encoding
hacks. I never programmed the MSP430 though, so don't remember any
specifics.
The MSP430 is still around, though somewhat fading under the ARM
juggernaut.

Lots of memos on PDP-11 origins and strategy here ...

http://www.bitsavers.org/pdf/dec/pdp11/memos/

Larry McGowan

2020-10-16 22:37:51 UTC

Permalink

Post by Larry McGowan
http://hampage.hu/pdp-11/birth.html

Thanks for the cool document! I remember that the 16-bit instruction
word could only encode a few two-address instructions (ADD, CMP)
with 6-bit (register # and mode) specifiers, and some of the others like
XOR had to make do with more constraints on the operands.
Maybe because of that, I heard claims that the -11 was originally
intended to be an 18-bit machine (conveniently a PDP-10 half-word), but
it then had to be shoehorned into 16 bits because that was what the
market wanted. With 18 bits the instruction encodings could have been
more uniform, and/or there could have been 16 registers instead of 8.
Was there any truth to that? Or was it 16 bits from the beginning? Did
the designers feel any pain over those issues? The later PDP-11-like
Texas Instruments MSP430 microcontroller had 16 bit words but 16
registers. I guess it did that through even more instruction encoding
hacks. I never programmed the MSP430 though, so don't remember any
specifics.
The MSP430 is still around, though somewhat fading under the ARM
juggernaut.

Paul, Originally we started out to do an 8-bit computer but noticed the world was going 16-bit. Never was intended to be 18 bits.
-- Larry

Rich Alderson

2020-10-16 02:10:37 UTC

Permalink

Post by Larry McGowan
I just found this group and thought I would provide a link to a document I
wrote some years ago about the birth of the PDP-11. In 1967 at DEC in
Maynard, Massachusetts, I was the software manager for the paper tape
software for the original PDP-11/20. I also wrote the first simulator,
SIM-11, which ran on the PDP-6/10 and the first manual documenting the
assembly language. Those were exciting times!
http://hampage.hu/pdp-11/birth.html
also on my webpage at
https://www.dropbox.com/s/c9e6crm3rlo70re/pdp-11birth.pdf?dl=3D0

Interesting. Do you still have the code for SIM-11? Speaking as a PDP-10
programmer, that would be very interesting!

--
Rich Alderson ***@alderson.users.panix.com
Audendum est, et veritas investiganda; quam etiamsi non assequamur,
omnino tamen proprius, quam nunc sumus, ad eam perveniemus.
--Galen

Lars Brinkhoff

2020-10-16 05:26:15 UTC

Permalink

Post by Rich Alderson

Post by Larry McGowan
I just found this group and thought I would provide a link to a
document I wrote some years ago about the birth of the PDP-11.

Thank you for posting this here!

Post by Rich Alderson
Do you still have the code for SIM-11? Speaking as a PDP-10
programmer, that would be very interesting!

Agreed!

(There is also the simliarly named 11SIM by Eastlake, Stallman, et al.
Developed on ITS and ported to TENEX and, I believe, TOPS-20.)

Paul Rubin

2020-10-16 08:56:48 UTC

Permalink

Post by Lars Brinkhoff
(There is also the simliarly named 11SIM by Eastlake, Stallman, et al.
Developed on ITS and ported to TENEX and, I believe, TOPS-20.)

Speaking of ITS, does anyone know anything about MINITS? It was
apparently a PDP-11 multitasker (not a full OS) written by ITS hackers
to run some of what would now be considered embedded applications around
the AI lab, I think. I did locate some info about it a while back, but
have forgotten everything specific by now, except that it was less fancy
than I had hoped.

Lars Brinkhoff

2020-10-16 13:26:05 UTC

Permalink

Post by Paul Rubin
Speaking of ITS, does anyone know anything about MINITS?

Yes, it's part of the ITS repository on GitHub. Assembled with PALX
from sources.

Post by Paul Rubin
It was apparently a PDP-11 multitasker (not a full OS) written by ITS
hackers to run some of what would now be considered embedded
applications around the AI lab, I think.

Most prominently Chaosnet routers, but also some graphical workstations,
etc. Source code comments point to it being in use outside MIT at Yale,
University of Texas, Symbolics, Palo Alto, Atari Cambridge, and
Sunnyvale.

Post by Paul Rubin
I did locate some info about it a while back, but have forgotten
everything specific by now, except that it was less fancy than I had
hoped.

Think of it as an RTOS with strong Chaosnet support, and some bundled
applications. It has a cousin in NSWIT, also a PDP-11 "operating system
slash router" (vibes of OpenWRT?) with Chaosnet routing.

Other PDP-11 systems inspired by ITS, written at or near MIT, include
SITS (Logo PDP-11/45 OS) and CamExec.

Larry McGowan

2020-10-16 22:43:27 UTC

Permalink

Interesting. Do you still have the code for SIM-11? Speaking as a PDP-10
programmer, that would be very interesting!
--
Audendum est, et veritas investiganda; quam etiamsi non assequamur,
omnino tamen proprius, quam nunc sumus, ad eam perveniemus.
--Galen

Rich, I left DEC in Maynard in late 1971 and then worked for DEC in California as the Western Region SW Specialist. No software was brought with me. SIM-11 was written in FORTRAN for the PDP-10. I would have probably coded it in PDP-10 assembly language but my boss wanted it in FORTRAN. Easier to read I guess.
-- Larry

Three Jeeps

2020-10-17 13:49:17 UTC

Permalink

Fortran for the PDP10 - Just curious, I assume it was Fortran IV ? with or w/o DEC extensions?

Yes, *WHAT* were you thinking wrt the console?!?!?
Again, thanks for the insight, especially the CMU connection. I did graduate school at CMU beginning 1976 through early eighties. My thesis advisor was one of the primary investigators of CM* (as well a c.mmp-Bill Wulf) which was one of a number of PDP11 based computing research projects. I had two 11/45s in my control lab and a 34a in my 'office'...fun times....

John

Larry McGowan

2020-10-17 16:37:58 UTC

Permalink

Fortran for the PDP10 - Just curious, I assume it was Fortran IV ? with or w/o DEC extensions?
Yes, *WHAT* were you thinking wrt the console?!?!?
Again, thanks for the insight, especially the CMU connection. I did graduate school at CMU beginning 1976 through early eighties. My thesis advisor was one of the primary investigators of CM* (as well a c.mmp-Bill Wulf) which was one of a number of PDP11 based computing research projects. I had two 11/45s in my control lab and a 34a in my 'office'...fun times....
John

John,

About the Fortran - I just recall it was Fortran. It was a long time ago.....

We spent a lot of time with Gordon Bell and Bill Wulf from CMU designing the extensions to the basic machine (floating point, memory management, etc.).

-- Larry

jjh...@gmail.com

2020-10-16 18:52:47 UTC

Permalink

I just found this group and thought I would provide a link to a document I wrote some years ago about the birth of the PDP-11. In 1967 at DEC in Maynard, Massachusetts, I was the software manager for the paper tape software for the original PDP-11/20. I also wrote the first simulator, SIM-11, which ran on the PDP-6/10 and the first manual documenting the assembly language. Those were exciting times!
http://hampage.hu/pdp-11/birth.html
also on my webpage at
https://www.dropbox.com/s/c9e6crm3rlo70re/pdp-11birth.pdf?dl=0
Regards,
Larry

Very interesting and thank you for sharing. I hope you kept the front console!!!
Is SIM-11 still around?/Available?
John

Larry McGowan

2020-10-16 22:46:44 UTC

Permalink

I just found this group and thought I would provide a link to a document I wrote some years ago about the birth of the PDP-11. In 1967 at DEC in Maynard, Massachusetts, I was the software manager for the paper tape software for the original PDP-11/20. I also wrote the first simulator, SIM-11, which ran on the PDP-6/10 and the first manual documenting the assembly language. Those were exciting times!
http://hampage.hu/pdp-11/birth.html
also on my webpage at
https://www.dropbox.com/s/c9e6crm3rlo70re/pdp-11birth.pdf?dl=0
Regards,
Larry

Very interesting and thank you for sharing. I hope you kept the front console!!!
Is SIM-11 still around?/Available?
John

John, I didn't take the front console. What was I thinking? I have no idea about the existence of SIM-11. I believe it was in DECUS so if that still exists maybe it is in there.
-- Larry