TOPS-20AN

[Lars Brinkhoff]

I wrote this elsewhere, but it should be its own thread.

DEC picked up TENEX and made TOPS-20 from that.  There was a DEC product called TOPS-20AN which retained the ARPANET code.  So far, no one has tried running that on an emulator.

Here are some resources: tapes, documentation, etc.

https://github.com/obsolescence/arpanet/wiki/TOPS%E2%80%9020AN

DEC made their own IMP interface called AN10 or AN20.  This would have to be added to the KL10 emulator(s).

[Noah Smith]

Thanks for starting this thread Lars!:) Will be learning and watching here. And will also discuss with Clive Dawson in Austin. Clive was a core person for the UTCC DEC-10 until it was shutdown in 1982, and then was manager of the UTCC DEC-20, named UTEXAS-20. BTW there are reports (need to check this with Clive in upcoming Oral History Interview) that the Dual KL at UT was actually a DEC-20, not DEC-10 (?despite the mention of TOPS-10 in the table below?). And there are reports that Decwar was also on the Dual KL. Will be finding out more about all this soon!:)

690 TOPS-10 University of Texas Austin, TX KI10. (HRC DEC-10 on which Decwar was created) 
1114 TOPS-10 Texas St. Purc. Austin, Texas KL10B CPU1? (Painter Hall Dual KL?)
1360 TOPS-10 Texas St. Purc., Austin, TX KL1099D CPU0? (Painter Hall Dual KL?)

Pure speculation here, it could be that the best way to get Decwar directly onto the ARPANET is to get it running on a DEC-20 that is on the ARPANET?

Simpler is probably just some machine that is on the ARPANET and provides telnet access to a DEC-10 running Decwars? But who needs simple hahaha:)

Cheers,

Noah

[Lars Brinkhoff]

Noah Smith wrote:

BTW there are reports (need to check this with Clive in upcoming Oral History Interview) that the Dual KL at UT was actually a DEC-20, not DEC-10

Note that TOPS-20 didn't support multiprocessing, but TOPS-10 did.  If it was a dual DEC-20, it would have been two separate systems.

 

1360 TOPS-10 Texas St. Purc., Austin, TX KL1099D CPU0? (Painter Hall Dual KL?)

1099 implies a dual CPU DEC-10 running TOPS-10.

Simpler is probably just some machine that is on the ARPANET and provides telnet access to a DEC-10 running Decwars? But who needs simple hahaha:)

The quick cheat for all systems that doesn't have an NCP, is to run the "Linux NCP" as a front end and have that serve TELNET to the ARPANET, and then direct incoming logins to a terminal line.

For Decwars specifically, possibly have it running on TOPS-20AN.  Or snarf the NCP from TOPS-20 or WAITS and put that in TOPS-10; a big job, I know.

[Noah Smith]

Fantastic info, thanks Lars!:) This supports another picture that have definitely had in mind, will find out more about this from Clive soon!

Alternate picture, that actually explains mannnnnny things, there were actually more than three PDP-10 CPUs at UT

KI in HRC that Decwar was created on

Dual KL DEC-10 in ???, maybe even the old UTCC machine room? Completely explains reports of playing Decwar on a TOPS-10 Dual KL!

DEC-20 in Painter Hall, named UTEXAS-20 (and not listed in that list of CPUs up above)

More as we find out, Cheers!

Noah

[Noah Smith]

Quick update around machines in Austin. We received some fantastic new info and things seem muuuuuch clearer now!:) Here's a link
https://decwarorg.blogspot.com/2026/04/machines-in-austin-and-san-marcos.html  

[Noah Smith]

Hi Lars, am starting to dig a bit deeper into this. So for Austin, there are three historically appropriate host machines: a DEC-20 and two PDP-11s.

My inclination is to focus on the DEC-20, but one question is maybe it makes more sense to focus on the 11s at first? Not sure at all:)

And then another question, when there isn't yet an NCP (so the 20 and the 11s) does it make sense to get going first with the Linux NCP as an interim bridge?

Hmmm, so maybe one way to ask is - if you had these two choices, would one be clearly preferable?

1. DEC-20, get something working with Linux NCP, then try for TOPS-20AN

2. PDP-11, get something working with Linux NCP, then try for historical NCP
Cheers!

Noah

[Lars Brinkhoff]

Noah wrote:

And then another question, when there isn't yet an NCP (so the 20 and the 11s) does it make sense to get going first with the Linux NCP as an interim bridge?

Hmmm, so maybe one way to ask is - if you had these two choices, would one be clearly preferable?

1. DEC-20, get something working with Linux NCP, then try for TOPS-20AN

2. PDP-11, get something working with Linux NCP, then try for historical NCP

The interim bridge will definitely make things easier.

The problem with PDP-11 is that we currently don't have any NCP that is good to go.  There is the Illinois Unix, but there are some files missing or broken.  There is ELF which I have managed to assemble, but not link.

So I'd say TOPS-20AN is your best bet.

[Noah Smith]

Thanks Lars! This makes complete sense and fits with what was thinking:) 

So here's current rough sketch for PUA (Project UTEXAS ARPANET)

• CC IMP
• UTEXAS-20, Painter Hall Research DEC-20, get something working with Linux NCP, then try for TOPS-20AN
• HRC DEC-10

Note - Could be a reasonable minimal setup that's at least somewhat realistic for September/October 1982? Understood that HRC DEC-10 was not an actual ARPANET host, but guessing it was reachable from UTEXAS-20? Twisted-pair through the steam tunnels maybe, hahaha?:)

--
You received this message because you are subscribed to a topic in the Google Groups "PiDP-10" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/pidp-10/tB0FXa4syWw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to pidp-10+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pidp-10/34ab81bd-c745-41c1-8451-c5bf257fe1can%40googlegroups.com.

[Noah Smith]

Hi all, quick update here with a new blog post of where things are at

https://decwarorg.blogspot.com/2026/05/the-first-internet-computers-in-texas.html 

With help from Lars and Rich am hands on trying to get up to speed on TOPS-20 and getting TOPS-20AN going, any and all help welcome! Definitely learning new things myself hahaha - trying to get a KS10 to boot from a t20v4 tape, sometimes getting an MTBOOT prompt, sometime nothing, what? hahaha:)

Does anybody happen to have a working simh ini script to boot a KS10 from tap, maybe even t20v4? Below is what I'm trying right now.

Cheers,

Noah
set cpu 1024k
set rpa0 rp06
set rpa1 rp06
set dz lines=8
at dz 2020
at rpa0 -n ks0.rp06
at rpa1 -n ks1.rp06
at tua0 bb-d867c-bm_tops20_v4_2020_instl.tap
boot tua0

[Michael Thompson]

Noah,

As far as I know only the ADP modified KS10 systems had 1024kw of RAM. The rest were limited to 512kw of RAM.

On May 2, 2026, at 7:17 PM, Noah Smith <no...@statespace.dev> wrote:

Hi all, quick update here with a new blog post of where things are at

You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pidp-10/37465038-c015-445e-aeb2-a7045ae02279n%40googlegroups.com.