🚀 go-pugleaf

uunet!dg!lewine writes:

>jallen@eeserv1.ic.sunysb.edu (Joseph Allen) writes:

>> Cool project!  Maybe you can skip the extended address instructions (is that
>> what you mean?), I don't think TOPS-20 used them (although I could be wrong
>> about this).  The byte instructions must be fun.

>	(1) TOPS-20 *does* make use of the extended addressing stuff.  I
>        think that TOPS-10 does also.  In any event, it is not that 
>        difficult to include.  I don't know of any compilers that 
>        generate code that runs outside of section zero, but I have
>        not looked in a long time.

TOPS-10 definitely did support the extended-address stuff, but not until
one of the 7.0n releases, unless I am mistaken.  And, of course, TOPS-10
continued to run on KA's and KI's, so the extended-address instructions
would not be totally necessary for TOPS-10.

When I finally left DEC-10 land (not without a great deal of regret), DEC
had promised that the FORTRAN compiler would be able to generate non-section-0
code.

>    (2) The byte instructions are fairly easy.  The biggest problem
>        is dealing with 36-bits on a machine that has 32-bit longs.

I just wrote 'em the other day for my own emulator (which is in a fairly
embryonic stage at this point).  They were "fairly" easy, but definitely
fun, as they were non-trivial due to the 32/36 bit stuff.  I can't even
force myself to THINK about 36-bit (or 71-bit - there's a trivia question
for you) floating point; I would despair of ever getting the low-order
bit EXACTLY right in all situations ...

>> Once I heard a rumor that someone was going to put a DEC20 on a chip...
>> anyone here about this?  I have no other details.
>     No.  I designed a one chip PDP-10 as a test for some CAD tools
>     and gate array technology.  It was never built.  I think I 
>     still have the design someplace.  Took me a whole weekend!

That's very impressive.  I considered building one out of bit-slice at one
time, but this sounds like a far better deal.  I'm surprised it went this
quickly, especially with paging and the cache to do ...

>     The real questions is: would anyone want one badly enough to
>     to pay to get the gate array built?

At one time, I'd have paid good money for a DEC-10 on my desk; but I'm
addicted to Unix and superfast workstations now (though not without some
(more) regret).

>     Also, DEC still has the
>     rights to the software and I don't think they are willing to
>     give it away.

Worse, I hear that they are not even going to give the TOPS-20 sources to
the customers they are abandoning.  At least the TOPS-10 folks have sources
and can do their own support.  I can think of nothing but unprintable
adjectives (complete with raising questions about their parentage) for the
DEC folks who made THAT decision.
 
>Anyway, I spent from about 1969-1980 building -10 software, hardware,
>and microcode.  I would love to see a PDP-10 emulator, but not enough
>to do any work on it.  I will be happy to answer questions on how the
>machine, or Tops-10 or Tops-20 worked.

I was a TOPS-10 hacker from 1976 to 1985; never did any hardware stuff
beyond using a soldering iron on a KA-10 a few times.  I am writing an
emulator, mostly just for fun; I'm not really motivated enough to try
to actually make the I/O work (or the floating-point, for that matter).
But I also would be quite happy to entertain questions.

>--------------------------------------------------------------------
>Donald A. Lewine                (508) 870-9008 Voice
>Data General Corporation        (508) 366-0750 FAX
>4400 Computer Drive. MS D112A
>Westboro, MA 01580  U.S.A.
>
>uucp: uunet!dg!lewine   Internet: lewine@cheshirecat.webo.dg.com

Terry R. Friedrichsen

terry@venus.sunquest.com  (Internet)
uunet!sunquest!terry	  (Usenet)
terry@sds.sdsc.edu        (alternate address; I live in Tucson)

Quote:  "Do, or do not.  There is no 'try'." - Yoda, The Empire Strikes Back