[Simh] PDP-8: The possibilities?

[Johnny Billquist]

Bob, you are right in that the PDP-5 used address 0 for the PC, and put 
the saved PC at address 1 at interrupts, which is not compatible with 
the PDP-8, and means any interrupt driven code will not work across the 
PDP-5 and PDP-8.

Not sure what you mean by 0/1 for interrupts. Maybe you mean that 0 is 
where the pre-interrupt PC is saved, and execution starts at 1?

Anyway... As far as the early PDP-8 models go, the 8/S is the odd ball 
out. As far as I can remember, a bunch of OPR combinations did not work 
the same (or at all) on the 8/S, compared to any other PDP-8 model. So 
special care needs to be taken when you write something for an 8/S. 
Apart from that, the machines are mostly upward compatible, indeed. The 
Omnibus machines added a few new things, but yes, you normally use 
various undocumented opcodes to tell the machines apart. RAR RTR is the 
one I know the best, but there are probably others too.

Kermit-12 is a good source if people want to check how to tell which 
model it is running on, since that program does a pretty decent job of 
identifying pretty much all machines.

Ray Jewhurst mentioned that the serial bus of the 8/S was the basis for 
the Omnibus - that is backwards and wrong in several ways. First of all, 
I'm not sure the bus was serial on the 8/S. The CPU was serial.
Second, the Omnibus is most definitely not serial, and I also seriously 
doubt there are any relationship at all between the Omnibus and anything 
on the 8/S. Third, I have some vague memory that the Negibus was used on 
8/S, but I should probably look that up.

	Johnny


On 2016-09-09 03:53, Bob Supnik wrote:
> The PDP-5 is, in fact, not all that compatible, because it used memory
> location 0 as the PC, pushing the interrupt locations to 1/2, instead of
> 0/1. So any program requiring interrupts will not work on a -5 vs an -8.
> The PDP-5 had an IO halt/restart facility, modeled on the PDP-1 and
> dropped from the PDP-8, which allowed an IOT to "wait" for completion
> without looping and testing a flag. It does not seem to have supported
> an EAE or extended memory.
>
> The PDP-8 family (8, 8/S, 8/I and variants, 8/E and variants, 8/A) are
> superset compatible for defined operations. It's possible to tell them
> apart based on their behavior on undefined operations. The code for
> identifying a PDP-8 is out there, but I don't have it at hand. I
> remember that the behavior of RAL RAR and RTL RTR was one way of telling
> the 8, 8/S, and 8/I apart.
>
> Most of the work for supporting models would be in the peripherals,
> particularly the ones that are 'compatible' across the line (reader,
> punch, terminals, clock). The pre-Omnibus machines used the older style
> IOP1, IOP2, IOP4 pulse methodology; the Omnibus machines can decode all
> 8 possible combinations. Beyond that, peripherals tended to be distinct:
> the RK8 for the 8/I vs the RK8E for the Omnibus machines; the Type 552
> DECtape controller for the -5 and -8 vs the TC01/TC08 for the later
> machines.
>
> The "CMOS 8s" are a whole different kettle of fish. They were only used
> in word processing/DECmate systems and had many unique features.
>
> /Bob
>
> On 9/8/2016 9:10 PM, simh-request at trailing-edge.com wrote:
>> Message: 1
>> Date: Thu, 8 Sep 2016 18:57:52 -0400
>> From: Ray Jewhurst<raywjewhurst at gmail.com>
>> To: simh<simH at trailing-edge.com>
>> Subject: [Simh] PDP-8: The possibilities?
>> Message-ID:
>>     <CAMFEAABLe-s+qSZmm4AXyR8Pqhx3dPkiaDJb_aUXQo5hAHEt1g at mail.gmail.com>
>> Content-Type: text/plain; charset="utf-8"
>>
>> After both reading and participating in some recent discussions, I got to
>> thinking that maybe the array of PDP-8 models could be better
>> represented.
>> I say this because from what I have read very early PDP-8 code is not
>> 100%
>> compatible with later models conversely the PDP-5 is compatible with the
>> early code and likewise uses a negibus like the Straight-8. I thank this
>> could be a rewarding experience for some of us and since I can't work I
>> would be able to help coordinate, write pseudo code and beta test. If
>> anyone is interested in this let the discussion begin.
>>
>> Thanks
>> Ray
>
> _______________________________________________
> Simh mailing list
> Simh at trailing-edge.com
> http://mailman.trailing-edge.com/mailman/listinfo/simh


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