[Simh] A possible Interdata roadmap.

[Davis Johnson]

I'm new here, so I'd like to introduce my self and what I'm up to.

My specific interest is the Interdata emulators. In the long run I would 
like to emulate a 3200 system running at least OS/32MT R09-02.2, or 
perhaps if I can get that far 3280 emulation R10-01. Unfortunatly the 
OS/32 licensencing situation may not be very good. I may have some more 
notes on that later, if there is an interest.  The Wolongong Unix should 
run just fine on a 3240, though. All of OS/16 was released to the user 
group (Interchange) long time ago - but I no longer have a copy.

I do have an actual 3203, and access to a 3280 to run on if I can't use 
the emulator for OS/32.

My current project roadmap is long, and don't expect hurried progress:

1) 8 register set 8/32 support.
The vast majority of 8/32 systems I've worked on have had 8 register 
sets. 8 register sets are required to run the more recent OS/32 revs. 
I've currently got this working but improperly. Right now all id32 
emulations get 8 register sets, even 7/32. It needs to be an option on 
8/32 systems only. When I get it right I'll send you a patch.

2) Run OS/32.
An older OS/32 will boot on the emulator. It uses more than two register 
sets. I've written a couple of small utilities to help build a boot tape 
file. They need a little polish, them you can have them if you are 
interested.

3) 3220 support.
The 3220 is an intermediate system between the 8/32 and the rest of the 
3200 family. It shares the same memory relocation/protection hardware as 
the 8/32, but implements the rest of the 3200 features, highlights include:

    * No hex display and keypad front pannel. The 3220 kept just a
      handfull of switches. Most front pannel operations are now handled
      by processor microcode that interacts with a console terminal at
      MUX address 0x10. This may not be worth emulating, might be worth
      doing for completness. The SCP does all this does and more.
    * New boot loader. It is still loaded by processor microcode from a
      ROM board ("LSU", Loader Storage Unit) and address 5. It is larger
      and interactive, using the same console as the microcode front
      pannel replacement.
    * A bunch of new instructions. Mainly commercial string and packed
      decimal sort of things.
    * Read and Write block instructions dropped.
    * Fault/exception handling is considerably cleaned up. Some faults
      give more information. Floating point faults now give the address
      of the faulting instruciton instead of the address of the next
      instruction.
    * Separate PSW bits for single and double precission floating point
      support.

4) 3210, 3212, 3230, 3240, 3250 support:
These probably represent the bulk of the 3200 systems shipped. As far as 
the OS was concerned there wasn't supposed to be any difference between 
them. They actually only have two distinct CPUs accross all those 
modles. The 3210 is a 3230 in a short cabinet with fewer backplane slots 
and no cache. The 3210 is a 3210 with cache added back in. The 3250 is a 
3240 with different memory power (5V instead of 12V). There is a little 
bit more to that, but not much. Unless you want to emulate memory error 
logging there may not be any differences worth emulating. There is a big 
difference from the 3220 tho:

    * New memory address translation/protection hardware, the MAT
      (Memory Address Translator). Much more complicated that the MAC
      (Memory Access Controller) in the 7/32, 8/32 and 3220. It makes
      access to 16MB of memory and virtual memory possible.

5a) 3260 Support:
The 3260 is a 3250 with yet another backplane (sort of). The 3260MPS was 
a multiprocessor system, with a 3250 CPU and modified 3230 CPUs as 
"APU"s. After the 3260 was a current product they started shipping 
uniprocessor systems with multi-processor capable backplanes. In theory 
you could have somthing like 10 processors. How would SIMH handle this?

5) 3203/3205 Support:
A single board CPU, with up to 2 MB of main memory and a SELCH on the 
CPU board. I've got a 3203 in my collection. The biggest diffence for 
the emulator is that the MAT allocates memory in 8K instead of 4K chunks.

6) 3280 and on.

 The 3280 was the first 3200 processor to address more than 16MB of ram. 
Needless to say there are a lot of differences. From Concurrent's 
(Interdata's successor's) standpoint the 3280 may be too close to 
current for comfort. Within the last several years they have sold a few 
million dollars worth of 3200-2000 processor upgrades.

***************************************************************************************
A few notes on my history, so as to give you some idea who just dropped 
in unannounced.

In middle school they tried to teach us some programming using an 
Olivetti Programma 101 programmable calculator. An interesting machine 
but a bear to work with. Needless to say little success was achieved.

We had a "Computer Science" curriculum in high school that was taught 
using dial up access to an HP2000 time sharing BASIC system. I was an 
instant computer junkie.

I graduated from high school in 1977, and did a little time at the local 
community colledge. I took Electrical Engineering Technology courses and 
a few MIS courses where I learned the rudiments of IBM 370 assembler and 
FORTRAN.

I then went to Old Dominion University as an Electrical Engineering 
student. Student assignments were done on the University's Decsystem-10 
(KL1090, 1MW) along with all other student uses and adminstrative 
functions. After just barely making it through second semester circuit 
analisys with a passing grade (first semester started in an auditorium, 
second semester finished in a small room with fewer than 20 students - 
I'm not ashamed of my performance) I switched to Computer Science. I 
learned Macro-10, Pascal, Algorithms etc on that same KL1090 until the 
computer science department actualy got its own computer, a PDP-11/44 
running V. 7 Unix. There I learned PDP-11 assembler (which I liked).

I wanted my own computer so I bought a 9900 based process controler from 
a company in Columbia Maryland. I had already figured out that I didn't 
want an 8 bit machine. I realy wanted a PDP-11, but the Heathkit H-11 
was too expensive for me. The 9900 has a similar instruction set and 
this process controler was a lot cheaper.

I left school for medical reasons (I'm better now, thanks) and 
unfortunatly never made it back.

I got a job working for Perkin-Elmer in their Rockville MD field office. 
There I provided pre- and post-sales technical support for Maryland, DC, 
North Carolina and, in theory, parts of West Virginia. I never did 
figure out which parts of WV, never had to go there. For about 5 years 
if you bought a new P-E system in those areas I'd probably be the one to 
do the initial on-site software install for you -- OS/32 or Unix. Didn't 
do much Unix. If you needed on site software support, or if the customer 
engineers needed software support that was probably me, also. I also did 
custom device drivers, occaisonal hardware integration - what ever was 
needed.

After that I went to work for a customer. There they were in the process 
of replacing a number of PDP-15s and implementing some new functions 
with 3260s and 3230s. The PDP-15s had replaced PDP-1s many years before. 
The front pannel of one of the PDP-1s is still in the office.

My tenure there was interrupted by three years at elsewhere, where I 
supported two 3200 systems, including a 3280, providing scientific 
timesharing services for their engineers.

Back to what has become my long term job, I haven't done much 3200 
anything professionaly since Y2K. We still have 8 3280 systems. Six on 
line, six on standby/programming support, two for hardware maintenance 
support.

Somewhere along the line I started collecting computers. Right after I 
stopped seeing interesting (toggle switches and lights) old machines at 
hamfests for next to nothing. Today the front pannel alone for some of 
these things can set you back a chunk of change.

I've got:

Interdata model 74 (front pannel and a CPU boards).
Interdata 6/16 (works, no periperals).
Perkin-Elmer 3203 (worked until the last time I tried to turn it on, 
probably a power supply problem)

Data General Nova 1200 - Possibly restorable, probably not.

DEC PDP-11/05 - Probably restorable. Was stripped of I/O controlers and 
perhaps some unibus terminators. Would probably work with the addition 
of some grant cards and terminators. I've found the maintenance manual. 
I'm crossing my fingers.

Digital DECsystem 5000/240 - Works, boots Linux.

Digital AlphaStation 200 4/166 - Works, boots Digital Unix or Linux.

A Tadpole Alphabook 1 laptop - Works, boots Linux.

I've been looking for a VAXstation.

Sun SPARCstation 2 - works, boots Solaris or Linux.

Sun SPARCstation 10 (4 processor) - works boots Solaris or Linux

An RDI Powerlite laptop - works, boots Solaris or Linux.

I've been looking for an Ultrasparc system, probably an Ultra 5.

Various working MACs.
    68000 represented by a MAC portable and a Macintosh IIfx (68030).
    68000 with PPC upgrade card represented by a Quadra 605. (68LC40 or 
601?)
    PPC represented by a Beige G3. Also boots Linux.

An IMSAI 8080 (sort of like the altair, a little newer) that suposedly 
worked when I bought it. I've never tried.

Front pannel off a General Automation computer.

Various odd bits and pieces, including a delay line memory and an 
original distribution copy of PDP-15 DOS on dectape. A bunch of old 
manuals. I should do an inventory some time.

I didn't realy mean to write so much, I thank you for your time if you 
got this far.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.trailing-edge.com/pipermail/simh/attachments/20060202/41b9cdc7/attachment-0003.html>