What Ever Happened To... The PDP-11 Architecture? (was: Re: Simulator and RSTS v7.0 Update)

[Aron Insinga <ainsinga@infomation.com>]

Bob Supnik was in charge of the J-11 ('Jaws'), the CMOS PDP-11-on-a-chip.  
It was to be the last PDP-11 CPU implementation, excluding PDP-11 mode in 
VAX-11s.  After that he was in charge of the MicroVAX chip (as opposed to 
the MicroVAX-I which was designed at DECwest as 1 board with some 
semi-custom chips, or the V-11  VLSI VAX chip set [VAX-11/8200??]; the 
later was the source of most of the design pieces that were rearranged to 
form the MicroVAX chip, leaving out the odder instructions to be emulated 
in software as was done for the MicroVAX-I).  I think the J-11 was in the 
PDP-11/8x systems.

He's a great speaker, if you get the chance to hear him.  The J-11 had this 
line of drivers across the output of the arithmetic unit, I think, and each 
one was carefully sized by hand to be big enough to drive the length of 
wire it had to drive in the amount of time needed, but no larger.  (Each of 
the wires was a different length due to going around corners.)  That marked 
a turning point at DEC towards Mead & Conway structured VLSI design 
whereby there were fewer cells designed and they were replicated more times 
each.

Before the J-11 there was the F-11 ('Fonz') which I think was a couple of 
chips and the T-11 ('Tiny') which was 1 chip with an bus like the Intel 
microprocessors of the day so that DEC could stop buying 8080s and whatnot 
from Intel for embedding in terminals and controllers and whatnot.

Except for the bit in DECwest [Bellvue, WA; founded when Dave Cutler's 
group left MA for Silicon Valley, discovered how expensive housing was, and 
then left there for WA], the above chips were designed and fabricated at 
the DEC facility in Hudson, MA [and perhaps other places like Scotland] 
which Intel is now taking over.  When Gordon Bell left DEC, he complained 
about 'our vanity, protected architectures.'  He wanted to let other 
companies make PDP-11 clones to get it into the running for industry 
standard microprocessor architectures.  DEC, as always, wanted to protect 
its hardware manufacturing business.  (The software engineering group in ZK 
got slapped down when we wanted to sell DEC's software development tools 
for Windows, before DEC got onto the WNT [VMS++] bandwagon and before there 
were many decent such tools for Windows.  The same thing happened to 
Apple's Claris.)

Being near Hercules, DuPont, Atlas, and of course the UofD, the local 
section of the IEEE Computer Society had (and I hope still does) a great 
speakers program.  I remember when the M6800 (not M68000) was new and 
someone from Motorola came to talk about it.  At the end he acknowleged DEC 
for the PDP-11 architecture, to which the M6800 architecture owed a lot, 
and also for the idea of the general-purpose computer that could be 
embedded in other systems (which was the PDP-8).

So does anybody know the patent # for GRPs & addressing modes?  That 
was how the PDP-11 was protected.  The PDP-8 was protected with a patent on 
the auto-indexing memory locations; I don't know that patent # either.  It 
would be amusing to get copies of them from the Patent & Trademark 
Office.

- Aron