The PiDP-10 is finally done! And to celebrate I put up a web site covering the 'family': PiDP-1/8/10 and 11

[oscarv]

Hi,

I hope it’s appropriate to post this here but after some seven years, the PiDP-10 replica is finally done! We’re going to send out the first kits next week (if US Customs will not delay us further).

With a PiDP-1 coming up later this year, I feel we’re covering a pretty good chunk of DEC history now:

The PiDP-10 took many years to complete, but it is also the most ambitious of the range. It ‘captures in hardware’ the ITS Reconstruction Project of Lars Brinkhoff et al. So it not only boots up TOPS-10 and ITS, but the hard disk images contain more than 400 programs; PDP-11 and PDP-6 simulators are part of the simulation package. The PDP-10 was the only mainframe ever that became a playground for hackers. So the hardware mix is delicious (the PDP-6 and a PDP-11 share the Ten’s memory!); and there are lots of video games (spacewar, Maze War) to provide entertainment between bouts of PDP-10 assembly, Lisp. This was a networked machine, and thanks to an emulated IMP it’s possible to recreate the ARPANET if the modern internet is too boring.

Here is the new website: www.obsolescence.dev.

I can’t resist two glamour shots of the Ten, now it is finally, finally done:

We’re working on a PiDP-1 next, and a Whirlwind will complete the range after that. The PiDPs have become a real team effort, with Lars Brinkhoff and Richard Cornwell to thank for ITS and the PDP-10/KA10 simulator, Angelo Papenhoff for the PDP-1 simulation, and Guy Fedorkowski for the ongoing Whirlwind recovery project. We’re having fun!

Kind regards,

Oscar.

[oldben]

Why do you need BIG FANS? A cool design would be much cooler.
This version I like,One can touch the hardware, or get good vibes from the
microcode.

http://www.fpgaretrocomputing.org/pdp10x/
It is good to have fun.
I have fun small stuff.

[oscarv]

Ben,

Yes! A FPGA plug-in to replace the Pi is definitely something on the menu. It should not be too hard - Sytse van Slooten did it for the PiDP-11 to, adapt his PDP2011 FPGA. The PiDP hardware can be driven through the FPGA without any resistors, buffer chips or whatever.

As regards to the fans, there are two reasons.

First, the new Raspberry Pi 5 runs hot - it actually comes with a little clip-on fan that is quite good. But I still hate the whiny noise of tiny fans and it does come on now and then.

Second, more importantly, the fans are an optional add-on to provide - er, soundscaping. It just adds to the experience if a mainframe has (soft) fan whirr of large 12cm PC fans.

Same reason is why I want to add Teletype clatter in a future update of the software. It adds to the experience (if it can be muted once it becomes annoying)

There’s quite a bit of thermal design in the PiDP-10 actually - the PCB has air vents cut into it to help air flow around the Pi, and the big fans will suck out air that can only come in by flowing past the Raspberry Pi…

But as said, the two big fans are an option.

Kind regards,

Oscar.

[oldben]

Saving the really the BIG fans for the PDP 11 (emulated) to boot the system.
Do you have a filter for dust?
A PDP 4 is here https://www.fpgaretrocomputing.org/pdp4x/
With the PDP 4 , I think most of DEC is covered.
Thank you for the feedback.
Ben.

[EdS]

Great to see the product, @oscarv! And very interesting to hear of the next steps.

I’m pretty sure the pi5 will run without a fan attached, and will throttle if it needs to. For a single-core application I personally would try that first. But I know people get worried about hot chips. And yes, there’s something to the soundscaping argument!

[scruss]

@EdS

I’m pretty sure the pi5 will run without a fan attached, and will throttle if it needs to.

They will, but if they’re anything like earlier SoCs used in Raspberry Pis, thermal throttling is an enter-only flag. Once the CPU has gone over the limit temperature, it will remain throttled until it is power-cycled. So you want to avoid thermal throttling, unless you want to run at a reduced capacity most of the time.

[EdS]

I’m really surprised - do you have a reference for that?

[scruss]

Sure! Bits 16-19 of vcgencmd get_throttled output remain set if their individual throttling conditions ever get hit. They are cleared on reboot.

The CPU frequency and voltage do seem to recover after a throttle event, but there must be some reason/effect of the persistent flags

[EdS]

The CPU frequency and voltage do seem to recover after a throttle event

Ah thanks that’s what I would have expected. I’m not so worried about the sticky bits - I’d assume they are intended to be helpful for the OS.

[larsbrinkhoff]

I think we should also mention it runs WAITS, the Stanford AI lab system. And TOPS-20, although it’s anachronistic to run it with a KA10 console. There has also been some work on bringing up TENEX.

[oldben]

Is the emulation engine, just for the PI or will it run on other platforms?

[wyan]

I believe the emulation engine is Open-SIMH, which runs on many platforms

[larsbrinkhoff]

That’s essentially correct.

It’s Richard Cornwell’s KA10 emulator, which is very closely related to Open SIMH. And also, Open SIMH has a frequently updated copy of Cornwell’s suite of emulators. Richard is adding support for the PiDP-10 panel, so it’s as upstream as it can be.

The emulator runs on Raspberry Pi, most Linuxes and Unices, maybe MacOS, and Windows.

[larsbrinkhoff]

Here’s how you can toggle in a small program.

Toggling in a small program on the PiDP-10

[Coderjon]

You and your team must have a great sense of satisfaction and relief after completing this challenging project.

What development environments are available on the PiDP10: assembler, compilers & linkers, interpreters?

[EdS]

Welcome, @Coderjon!

There’s an active community on Google Groups (which also functions as a mail list) - see
https://groups.google.com/g/pidp-10

As the PiDP-10 works very much like the historical PDP-10, the choices of operating systems and tools are (almost all) from that era. ITS appears to be a popular choice, as does TOPS-10, according to my reading of the list. These are sophisticated operating systems running on a complex (simulated) computer - you should expect to need to read at least some tip sheets and probably some manuals, to get yourself trained as a sysadmin and as a user.

Perhaps start your research at Oscar’s site

PiDP-10 | obsolescence

[Jim_F]

The view from half a century ago:

Ted Nelson, “Dream Machines”
(from “Computer Lib/Dream Machines”, first edition, 1974;
Special Supplement to the third printing, August 1975, p. X)

This is obviously the place tell you about Alan Kay and the Dynabook.

The hottest project at Xerox PARC is Alan Kay’s Dynabook, formerly
the Kiddy Computer. As lots of people will tell you, it’s going to
cost five hundred dollars, be small enough to carry around on a
shoulder strap, have a built-in screen, run on batteries, and have
all the books a kid wants to read from the screen stored on a cassette.

And the demos: They’ll knock you out. On a color TV screen, they’ll
show you a wildly changing pageant of toy soldiers, photographs,
beautiful patterns, all generated by the computer in real time
(see 'Bit Maps," p. Z). And if you’re into computers, they’ll show
you how all this is run by the beautiful SMALLTALK language (it was
previously called the Kiddy Computer, remember), which any bright
child can learn and which has some awfully powerful features.

Now let’s sort this all out.

There have been a lot of cons in the computer field, but this is not
one of them. It’s marvelously real.

So how come Xerox is leaving the computer field?

Answer: they’re not exactly leaving: they’re taking a break until
they can sell thls beauty for five hundred dollars.

What’s the delay?

The Dynabook, or Kiddy Computer, is actually a PDP-10.

You’re supposed to laugh. A PDP-10 is a big computer, the best.
(See page 41.) A PDP-10 system costs hundreds of thousands of dollars.

But the last laugh will be Xerox’s. The way computer prices are
coming down, through integrated circuits ever more powerful and cheap,
that PDP-10 can be sold for $500 in… (check your choice)
__1978 __1979 __1980 __1981 __1982.

(Interesting anecdote: the guys at Xerox PARC asked to buy a PDP-10,
but management bridled, seeing as how Xerox was in the computer
business and made competitive machines. So the fellas, nothing daunted,
built their own. They modestly say the parts only cost a few
thousand.)

(Note: the above predictions are based, of course, on the assumption
of Xerox management knowing what it’s doing. Assumptions of this type
in the computer field all too often turn out to be without basis.
But we can hope.)

[oldben]

I think they made too much profit on the PDP-11 chip sets, that you would never get
a budget PDP-10 had DEC not developed the VAX. The PC really killed all the other
main frame systems as everybody moved to 386’s so one never got cheap other guy
chips for BIG IRON. DEC most likely would have moved to 32 bit RISC design instead,

Did the dynabook need 36 bit data word data?

The Pdp-11 and other 8/16 bit cpu’s pay big price to have byte addressible memory,
both in busing complexity and loss of address space and instruction complexity.
This made computers expensive, compared to say a 6502, with BASIC in rom
and 4K of ram, untill the mid 1980’s. Even then PDP stuff cost a ARM and a LEG
but one could never get a CPU chip set. Heathkit had the H11 from 1977 to 1982.
$1,300 for the basic system with 4K of ram.

.

[NoLand]

And the PiDP-10 goes mainstream. See this article in the Guardian:

SpaceWar is back! Rebuilding the world’s first gaming computer

[falk]

This is so amazing. I’ll be ordering one as soon as I can figure out where to put it.

Dare I ask: how is its performance compared to the genuine hardware?

Speaking as one who used to haunt the old AI lab back in the day, I am totally jazzed that you’ve recreated it here.