[Simh] Klh10 vs Simh

[Timothe Litt]

On 25-Feb-16 12:32, Zachary Kline wrote:
> Hi All,
>
> I’ve begun trying out Tops-20 on the KLH10 emulator, and also have an interest in ITS for the sheer novelty of the operating environment.
> I don’t know a lot about PDP10 CPU families, but was wondering how the two emulators compare.
> I gather DEC didn’t release newer versions of Tops-20 for KS10, and for some reason KLH10 seems to be preferred for ITS as well.
> I notice that the latter doesn’t seem to be maintained anymore.
>
> What are the chances of a KL10 emulator for SIMH?
> Apologies for the vagueness of the question. :)
> Thanks,
> Zack.
>
I suppose this is as good an excuse as any to provide a bit of history.

Ken was interested in the possibilities for a PDP-10 emulator product
after LCG engineering closed down, starting with the KS and eventually
the KL.

When he started, there were still a few PDP-10s around the company for
customer support - meaning that the hardware and software maintenance
contracts were profitable.  I had a large personal collection of related
materials.

He approached the DEC licensing office, and got permission to implement
the architecture and some arrangement for his customers to use the OS. 
He also got OS distributions and other materials from us to support his
development in exchange for DEC's ability to OEM what he produced. 
(Which we never did, as far as I know.  He was too late to market.)

He did his implementation work on Solaris.

I was his contact for architecture questions.  Dick Greeley was the CLO
person.  Larry Sendlosky (at one time in LCG, later worked in my group,
but at the time in the Alpha Migration Tools group) did a lot of 
research and packaging of materials for the effort.

In the process, Services finally gave up on the 10/20 line, but I was
able to provide a home for the last KL10 and what remained of the
software engineering archives.  It was a midnight project... many winks
and nods.

I don't think KLH10 was commercially successful for Ken - but the
exercise led directly to the TOPS-10/20 hobbyist license.  That was my
price for helping with the effort - and Dick was the right person to
make it happen.  And that was the precedent for the VAX (and PDP-11)
hobbyist licenses.

KLH was the first very complete implementation of the KL, complete with
the pdp-11 FE functions and ethernet.

At some point around 2001, he decided to release it as open source. 
It's popular mostly because it's a KL, and so supports the most
functional versions of TOPS-20 and ethernet networking.  

TOPS-20 stopped development for the KS at version 4.1 due to resource
constraints - and the technical reason that it just doesn't fit in the
single section, limited physical memory machine.  (Some people did  make
5.0 sort of fit - but it required recompiling the monitor and it
couldn't run many jobs.)  Of course, long-term the group had to support
both monitor versions, but the people who made the decision weren't in
it for the long term.  Sigh.

TOPS-10 was intended to suffer the same fate because it used KI paging
and the monitor group couldn't afford to maintain both KI and KL
paging.  So without telling anyone, I spent a long evening converting
TOPS-10 support for the KS to work with KL paging & patched the KS
microcode to make it work.  When I told JMF at the 4AM monitor meeting
after he removed KI paging that the KS monitor still built, he helped by
implementing a pseudo extended section for DECnet.  That saved us the
rather large cost of supporting two versions for the rest of the product
life.  As a result, you can run the last TOPS-10 monitor on SimH.

SimH has ethernet support, but KS10 support for ethernet was never
released by DEC.  There are traces of it - I had that as another
midnight project, but never completed it due to my short, 31 hour work
day.   It's still on my list :-)

Bob had been working on SimH, and wanted to complete his set of
simulators for the DEC product line.  And he likes the -10.  At that
point, Ken hadn't decided to open-source KLH-10.  And Bob wanted to use
his universal simulator framework for a PDP-10.

He thought that the KS would be adequate for most software -- and it's a
LOT easier to implement.  KL10 extended addressing is not easy to get
right.  And the CI, NI & DTE are non-trivial.  Having been thru the
effort of supporting Ken, supporting Bob was relatively easy.

SimH supports the ITS microcode for the KS10.

Tim Stark attempted to implement the KL10 with his TS10 emulator.  I
think he got quite far, but I haven't tried running his implementation.

SimH is actively maintained.  Because it shares code with the VAX and
PDP-11 (peripherals), it's easy to reconfigure for odd configurations.

Last I knew, Ken was working at Google and has moved on to other
pursuits.  I think KHL10 is fairly stable.

> What are the chances of a KL10 emulator for SIMH?

Great.  I expect that it would be accepted as soon as you write it,
demonstrate that it works & that it meets the project standards.

FWIW, I'm the last member of the PDP-10 architecture committee still
involved.



-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4994 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://mailman.trailing-edge.com/pipermail/simh/attachments/20160225/e03fb892/attachment.bin>