PiDP-10 Initial Software Feedback / Issues

[johntk...@gmail.com]

I am working on a 4Gb Raspberry Pi 5. Latest firmware and OS updates.

I have replaced the SD card with a NvME card.

I do not have the PiDP-10 hardware yet.

I've not installed anything on the Pi, other than settings up VNC and SSH.

1. Installation

cd /opt

sudo git clone https://github.com/obsolescence/pidp10

sudo /opt/pidp10/install/install.sh

The last line immediately return "do not run as root". So I dropped the sudo and the installation continued until the end. I said Y to everything.

2.  First use

I connected via VNC, opened a terminal. I'm running at 1920 x 1080.

I enter pdp and see the Blinky view (I presume) of Console Data Switch definitions.

I CTRL+A and D to stop that view.

I try pdpcontrol stop but that leads to command not found.

I see my path is /opt/pidp10/bin

I see ./pidp10 and run it, and odd things happen but when I ctrl C out my pwd is /home/pi and pdp and pcpcontrol are recognized.

I do:

$ pdpcontrol stop

$ pdpcontrol start 1

in order to get ITS running as per guide. The screen says

Starting PiDP-10 pidp10


screen -dmS pidp10 ./pidp10 bootscript.simh 1
0

So I skip to the How to Use section (aka the Using the 10 section?)

I ignore the part about the hardware for now, and enter. 

$ pdp con

A terminal view flashes on the screen for the briefest time and then shrinks and vanishes. No terminal for me! My experiment ends.

[johntk...@gmail.com]

I repeated installation on a different Raspberry Pi 5, this one using an SD card.

Again, all the latest updates.

Again, the error about running as root, so I dropped the sudo.

This time I did not install all the source code - much, much quicker.

I launched pdp and saw the Blinky panel. I did a control A-D, and this time back to /home/pi

I was able to do a 

pdpcontrol stop

pdpcontrol start 1

but again, 

pdp con

flashed up a terminal for a microsecond and then it vanished.

[Patrick Giagnocavo]

What is the TERM environment variable set to, when you run "pdp con"?  Might be worth it to try running "export TERM=vt100" to see if there is a change in behavior.

Regards,

Patrick

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/2ea60812-0594-4f9b-b80b-6ffeef836c0an%40googlegroups.com.

[johntk...@gmail.com]

Initially it was xterm-256color; setting it to vt100 made no difference.

I've used two clean installs, so I'm guessing something on the PiDP-10 software authors' Pis is subtly different.  

[Mitchell Wolrich]

my results are the same..  no matter what I do, pdp does show ITS running, but I am unable to get any response out of any of the emulation methods..  Was going to try a reinstall, but seems like theirs an issue.

Mitch

[Oscar Vermeulen]

John,

On Wed, 6 Mar 2024 at 19:28, johntk...@gmail.com <johntk...@gmail.com> wrote:

I am working on a 4Gb Raspberry Pi 5. Latest firmware and OS updates.

Great! At the moment, I have performance constraints with the Pi 4, but the Pi 5 runs everything without showing any signs of stress.

I'm sure I will get the Pi 4 to run smoothly too - even with  all the fancy hardware enabled, but I need to explore a bit more what to tweak.

1. Installation

cd /opt

sudo git clone https://github.com/obsolescence/pidp10

sudo /opt/pidp10/install/install.sh

The last line immediately return "do not run as root". So I dropped the sudo and the installation continued until the end.

Yes, that's the safety catch. As opposed to the PiDP-11, I wanted nothing to run as root, as it just causes trouble at some point somewhere.

 

2.  First use

I connected via VNC, opened a terminal. I'm running at 1920 x 1080.

I enter pdp and see the Blinky view (I presume) of Console Data Switch definitions.

I CTRL+A and D to stop that view.

So far, so good!

But then --

 

I try pdpcontrol stop but that leads to command not found.

pdpcontrol is a link in /usr/local/bin ; it is made in the install script, line 143:

     sudo ln -i -s /opt/pidp10/bin/pdpcontrol.sh /usr/local/bin/pdpcontrol

The funny thing is that the symlink to pdp did work, but the one to pdpcontrol does not. That is very confusing, I've not heard of anyone with this problem yet.

 

I see ./pidp10 and run it, and odd things happen but when I ctrl C out my pwd is /home/pi and pdp and pdpcontrol are recognized.

./pidp10 is just the simh simulator, you'd not normally run it manually. pdpcontrol is supposed to handle it.

 

$ pdp con

A terminal view flashes on the screen for the briefest time and then shrinks and vanishes. No terminal for me! My experiment ends.

I'll add a message to the script. The issue: you'd run two pidp10 instances, it clogs up the telnet ports.

 

My experiment ends.

Would you mind rebooting and see it everything goes smoothly now? I think I learn two things from your feedback:

1) somehow, the pdpcontrol symlink was not immediately activated. Or you made a typo? The link is made in the install script and the line that does it, 

     sudo ln -i -s /opt/pidp10/bin/pdpcontrol.sh /usr/local/bin/pdpcontrol

...should not cause any trouble! So this is mysterious. 

2) I should prevent two copies of the simulator running at the same time. They will battle for contention to access the GPIO front panel, but worse, clog up telnet ports. 

3) the 'pdp con' program (and others) should be polite and tell you that they find their telnet ports blocked. It is a reasonable thing to expect.

Kind regards,

Oscar.

[johntk...@gmail.com]

Rebooted multiple times with no change.

>pdpcontrol symlink was not immediately activated. Or you made a typo

yes, I think somehow I fat-fingered something. that error has not happened again.

[Mike Kostersitz]

I get here using 1 as the parameter for the switch state since I don't have the HW (yet)

pi@rpi-its:/opt/pidp10/bin $ telnet localhost 1025

Trying ::1...

Connected to localhost.

Escape character is '^]'.

Connected to the KA-10 simulator CON-TELNET device

No input ^Z or typing does anything though

In SIMH 

READ IN boots ITS from paper tape

/opt/pidp10/systems/its/boot.ini-28> at -u tk 10000 speed=300

%SIM-INFO: Listening on port 10000

/opt/pidp10/systems/its/boot.ini-31> at -u dpk 10002 speed=4800

%SIM-INFO: Listening on port 10002

/opt/pidp10/systems/its/boot.ini-32> at -u dpk line=11,10019 speed=4800

%SIM-INFO: Line 11 Listening on port 10019

/opt/pidp10/systems/its/boot.ini-33> at -u dpk line=15,10020 speed=4800

%SIM-INFO: Line 15 Listening on port 10020

/opt/pidp10/systems/its/boot.ini-35> at -u mty 10003 speed=50000

%SIM-INFO: Listening on port 10003

/opt/pidp10/systems/its/boot.ini-36> at -u mty line=9,10018 speed=9600

%SIM-INFO: Line 9 Listening on port 10018

/opt/pidp10/systems/its/boot.ini-37> at -u mty line=8,10017 speed=9600

%SIM-INFO: Line 8 Listening on port 10017

/opt/pidp10/systems/its/boot.ini-38> at -u mty line=7,10016;notelnet speed=50000

%SIM-INFO: Line 7 Listening on port 10016

/opt/pidp10/systems/its/boot.ini-39> at -u mty line=6,10015 speed=9600

%SIM-INFO: Line 6 Listening on port 10015

/opt/pidp10/systems/its/boot.ini-41> at ten11 10011

%SIM-INFO: Listening on port 10011

/opt/pidp10/systems/its/boot.ini-43> at auxcpu 10006

%SIM-INFO: Listening on port 10006

NAT args: tcp=2123:10.0.2.15:23,tcp=2121:10.0.2.15:21,tcp=2195:10.0.2.15:95

NAT network setup:

        gateway       =10.0.2.2/24(255.255.255.0)

        DNS           =10.0.2.3

        dhcp_start    =10.0.2.15

        redir TCP     =2195:10.0.2.15:95

        redir TCP     =2121:10.0.2.15:21

        redir TCP     =2123:10.0.2.15:23

  Protocol[State]    FD  Source Address  Port   Dest. Address  Port RecvQ SendQ

  TCP[HOST_FORWARD]  23               *  2195       10.0.2.15    95     0     0

  TCP[HOST_FORWARD]  22               *  2121       10.0.2.15    21     0     0

  TCP[HOST_FORWARD]  21               *  2123       10.0.2.15    23     0     0

/opt/pidp10/systems/its/boot.ini-51> at imp nat:tcp=2123:10.0.2.15:23,tcp=2121:10.0.2.15:21,tcp=2195:10.0.2.15:95

%SIM-INFO: Eth: opened OS device nat:tcp=2123:10.0.2.15:23,tcp=2121:10.0.2.15:21,tcp=2195:10.0.2.15:95

/opt/pidp10/systems/its/boot.ini-108> go 1000

                                             %SIM-INFO: Running with Buffered Console

This is what I see...

---

From: pid...@googlegroups.com <pid...@googlegroups.com> on behalf of johntk...@gmail.com <johntk...@gmail.com>
Sent: Thursday, March 7, 2024 1:15 PM
To: PiDP-10 <pid...@googlegroups.com>
Subject: Re: [pidp-10] PiDP-10 Initial Software Feedback / Issues

 

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/e49d60e9-c301-435d-b030-f485565d00c2n%40googlegroups.com.

[Mike Kostersitz]

Also realized that pdp con​ only works in a local session on the pi device directly not over SSH so expected to only use telnet localhost 1025 in that case.

Still not getting to login to its

Mike

---

From: Mike Kostersitz <mikekos...@gmail.com>
Sent: Thursday, March 7, 2024 1:50 PM
To: johntk...@gmail.com <johntk...@gmail.com>; PiDP-10 <pid...@googlegroups.com>

[johntk...@gmail.com]

I'm logged in via VNC if that makes a difference.

[Mike Kostersitz]

Same here now 🙂

If I uncomment b ptr in the /opt/pidp10/systems/its/boot.ini I get to boot to ITS as expected. Then do a pdp vt52 and ^Z then login with mike$u

Works. Seems to be be the load from paper tape in pdp.sh  that is borking out

Mike

---

From: pid...@googlegroups.com <pid...@googlegroups.com> on behalf of johntk...@gmail.com <johntk...@gmail.com>

Sent: Thursday, March 7, 2024 3:13 PM

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/b2746d6d-36ae-40ce-8f79-7589f1d01509n%40googlegroups.com.

[johntk...@gmail.com]

sudo /opt/pidp10/install/install.sh

The last line immediately return "do not run as root". So I dropped the sudo and the installation continued until the end.

>Yes, that's the safety catch. As opposed to the PiDP-11, I wanted nothing to run as root, as it just causes trouble at some point somewhere.

With regard to this point, my suggestion is to remove the sudo from the instructions:

/opt/pidp10/install/install.sh

 

[Heinz-Bernd Eggenstein]

Should we report problems here or on github?
Like there is a typo in the install.sh,
 sudo apt install -y libx11-dev libxt-libxft-dev
should probably be
 sudo apt install -y libx11-dev libxt-dev

HBE

[oscarv]

Yes, you're right, reporting issues on github is more efficient! 

I'll work in all the suggestions over the weekend.

Also, FYI, the install package now works on a regular Linux machine. Just overwrite the binaries in /opt/pidp10/bin with the files from the zip archive in that directory.

The install script does nothing nasty to your (Pi or Linux) laptop setup. If you let it when asked, it will add 'pdpcontrol start' to your .profile to autostart. And it copies 'pdpcontrol' and 'pdp' to /usr/local/bin. No messing with anything else.

Kind regards,

Oscar.

[Heinz-Bernd Eggenstein]

> Also, FYI, the install package now works on a regular Linux machine. Just overwrite the binaries in /opt/pidp10/bin with the files from the zip archive in that directory.

I'm confused: the set of files in the tar.gz is different from what is there for the Raspberry Pi build, e.g. there is a "pdp11" executable for x86-64, but a "pdp6" for the raspi, a "pdp-10-ka" for x86-64 but a "pdp10" for the Raspi?

Cheers

HBE 

[oscarv]

I will clean up the selection of binaries!

[Michael Thompson]

I just set up my Pi-5 in anticipation of attaching it to the PDP-10 front panel. The setup process was painless and everything seemed to work. I bought an 8GB Pi-5. My colleagues advised me to update the firmware to fix a memory performance problem with the 8GB versions. That process was also painless. Hopefully my kit will arrive in a few days and I can get started on the assembly. Lars and Oscar will be visiting the Rhode Island Computer Museum on April 2, so I should have plenty of time to assemble and debug the kit. Maybe we can get two of the PiDP-10s to talk to each other through emulated IMPS.

On Sat, Mar 9, 2024 at 1:25 AM oscarv <vermeul...@gmail.com> wrote:

I will clean up the selection of binaries!

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/50a1fc0a-c302-4ca9-b8de-eff2e4d99840n%40googlegroups.com.

--

Michael Thompson

[johntk...@gmail.com]

When you say "everything seemed to work", can you launch ITS and use it with a virtual terminal?

[oscarv]

I forgot I am the only one here with a switch called 'READ IN'... it is fixed now, also on a 'naked Pi', ITS should boot after a 'git pull' in /opt/pidp10 .

Do please let me know if not!

Kind regards,

Oscar.

[Lars Brinkhoff]

It may be noted an 8G SD card seems to be the very minimum for installing the Raspberry Pi OS (Desktop) plus the PiDP-10 software.

It was all I had at home when the PiDP-10 arrived.  I'll get a larger card soon.

[Mitchell Wolrich]

Oscar,

Another issue is, if you're running "headless" (no HDMI monitor connected, VNC only), "pdp con" flashes on the screen and disapears... it doesn't like that theew is no audio device.  When I connect up an HDMI monitor, it comes up fine.

Mitch

[johntk...@gmail.com]

The fact that was using VNC without the Pi connected to a display was definitely causing me issues, thanks Mitch.

I've now got as far as running the SHRDLU demo, which as been a long-time ambition! :-) :-) :-)

[johntk...@gmail.com]

BTW, how can I quit these programs? Entering :KILL doesn't seem to do much. I see CTRL-Z and then :KILL but again, no dice.

p.s. this is awesome. 
p.p.s. Ah, pressing F1 and then :KILL?

[Ethan Dicks]

On Sun, Mar 10, 2024 at 5:58 AM Lars Brinkhoff <lars.br...@gmail.com> wrote:
> It may be noted an 8G SD card seems to be the very minimum for installing the Raspberry Pi OS (Desktop) plus the PiDP-10 software.
>
> It was all I had at home when the PiDP-10 arrived. I'll get a larger card soon.

You can't even buy an SD card close to that small at the Micro Center,
and haven't been able to for years. 32G is the smallest they carry.

I went for high performance so that meant 128G.

-ethan

[oscarv]

Yes. 'pdp con' is an awful program, it is a very quick hack. 

I only use it for demonstrations to others, not for 'real work'. And please do not look at the source code. I'm not telling where it is either. Well, in pidp10/src somewhere. Will clean it up!

Use 'pdp telcon' instead of 'pdp con' to get a well-behaved, normal console port. No clatter, no slowness.

In fact, you can use 'telnet localhost 1025' instead of 'pdp telcon', because they're all the same telnet program in a different wrapper.

There are three bugs that I am fighting right now:

1. If you do a 'pdpcontrol stop' with the console window still open ('pdp con' or 'pdp telcon' or regular 'telnet localhost 1025', they're all the same telnet really), then the telnet daemon gets confused, waits, and blocks the telnet port for half a minute or so. And therefore, if you do 'pdpcontrol start' quickly after the stop, the console will just open and close straight away. 

If you want to inspect why, after the problem manifests itself, run 'pdp' to go into the simh simulator, type Ctrl A [ and scroll up all the way to the top. You'll see that simh was unable to open the telnet port when it restarted.

There are two workarounds:

- you can do 'pdp' and type 'ITS-esc-G' straight into the simh window.

- or you do 'pdpcontrol stop', wait 20 seconds and try again.

Unsatisfactory but temporary.

2. I know of no way that a SIGTERM will make simh close down graciously. So it's handled by an expect script (stop.expect) which works fine except:

-if you shutdown the Pi with the PDP10 simulator still running, you get to wait 90 seconds for Linux waiting on simh. That is very, very annoying. But I feel I'm not 'entitled' to edit the simh sources to solve this unless I can find no other option?

3. the type340 window does not always want to open when you try. I think it is due to a bug in the original 340 code in simh, that I only half fixed.

If anyone has ideas or suggestions, please use the Issues page on the PiDP-10 repository: 

https://github.com/obsolescence/pidp10/issues

I'll get the bugs fixed, but they are probably irrelevant:

Richard Cornwell will be quicker to release his brand new PDP-10 simulator version. He's working on it right now, it will be a drop-in replacement for the simh core we are using at present.

Don't forget to 'git pull' regularly, there are daily changes at the moment.

I'm grateful for any feedback, although https://github.com/obsolescence/pidp10/issues is the better place for it, soas not to clutter the Google Group.

Kind regards,

Oscar.

[Lars Brinkhoff]

ohntkennedy wrote:

BTW, how can I quit these programs? Entering :KILL doesn't seem to do much. I see CTRL-Z and then :KILL but again, no dice.

p.s. this is awesome. 
p.p.s. Ah, pressing F1 and then :KILL?

No most terminals, type ^Z to stop a program (just like Unix).  And then :KILL or $^X. ($ is altmode (altmode is ESC)) (include the period)

On a Knight TV, F1 takes the role of ^Z everywhere.

[johntk...@gmail.com]

Thanks for the improvements and information, Oscar. Things are improving every time I look. So much fun learning the ways of the PDP-10 _and_ some more Pi/Linux information.

I heart the pop con program. Hearing it clatter away when I log in or something changes is just magical.

[johntk...@gmail.com]

Ah, thanks Lars. I guess I've been trying this within the Knight TV (named after Tom Knight?) so the CTRL-Z needs a little help with the F1.

BTW, how I can run ELIZA? I see in the list but don't know how to launch it..?

[Lars Brinkhoff]

> Ah, thanks Lars. I guess I've been trying this within the Knight TV (named after Tom Knight?) so the CTRL-Z needs a little help with the F1.

Yes, Tom Knight designed the hardware, and the first version of the software.

> BTW, how I can run ELIZA? I see in the list but don't know how to launch it..?

Something like this:

lisp^K
Alloc? n
(load "liblsp;doctor")
(worker)

Type return twice to enter your message.

[johntk...@gmail.com]

That worked great! (Note: your discussions need to be in ALL CAPS)

[oscarv]

Progress report:

I added 'pdp view' to the command list, intended to be the default view on the PiDP-10.

pdp view pops up the left half of this screen shot. It very flexibly adapts to whatever the screen size is, and gives a status view of the PDP-10 simulator and PDP-11 (Knight TV system) at the bottom, and the console Teletype (sans noise) in the middle. The upper part is just Linux terminal, but do a 'telnet localhost 10018' and you're in the PDP-10. 

The bottom status views are maybe not useful for later, but certainly for now. 'pdp view' can be opened and closed without interfering with anything.

It also allows you to quickly access the simh emulators CTRL-E command line environment. From there, 'set dpy ena' will enable the type340 for instance, if you type 'set scale' right before that you get a reduced-size type340. 'cont' continues the simulator.

...the right half is still manual dragging and resizing of windows. Unbelievably, Wayland does not come with xdotools that you'd normally use. But not to complain.

Also, from tomorrow the main 'pdp' commands will live in the Pi's system menu. Nobody will care but I can show the icons...

Which is all minor news compared to the fact that Rich Cornwell is almost finished with the new version of the actual simulator. It will give some useful 'magic' commands through the front panel.

Because, one switch on the panel has no meaning in the simulator: PAR STOP. We can be quite sure there will be no parity errors in the PiDP-10, so this switch will now do things if set in combination with other switches:

As Rich wrote:

Special functions are accessible via parity stop switch. No
      matter what happens there will never be a parity error with
      simulator.

      Pressing "readin" will copy the lower 9 bits of address to the
      read in device. This will also be displayed in the lower 9 bits.

      Pressing "continue" will toggle MI indicator disable. It will
      also clear the MI indicator. This is indicated in the address
      display.

      Pressing "Stop" will simulate a power off quit the simulator.

      Pressing "Examine This" will set the top 4 bits of address
      switches into the repeat rate. This gives you 16 repeat speeds.

It is time to write the manual for the PiDP-10!

But the above special functions will not be activated until Rich releases the updated simulator. These features are actually in the current simulator as well, but not in a user-friendly manner.

Kind regards,

Oscar.

[oscarv]

I forgot to add: if you use the top window as the terminal into the PDP-10, by typing

telnet localhost 10018

then set the terminal to VT52 by entering:

:tctyp vt52

and it is good to see how much online help there is:

:info tctyp

<ctrl-Z>

:info

We need a manual, but not necessarily for ITS. It's pretty wel documented.

Kind regards,

Oscar.

[Lars Brinkhoff]

Oscar wrote:
> We need a manual, but not necessarily for ITS. It's pretty wel documented.

It is, but there are still many things that are not explicitly
explained in INFO or other files. Most of the information in this
wiki is either not part of the online ITS documentation, or is better
explained for a modern user.

https://github.com/larsbrinkhoff/its-manual/wiki

[Mike Ross]

Just installed everything on a brand new Pi 5 8GB. Install went as smoothly as... a very smooth thing!

Brought up the emulation, all worked exactly as described, full marks!

Started tvcon and type340. Ran :LORENZ to the great edification of the other scientists in the room. Wonderful. 

After around 5 minutes I noticed the attractor was no longer moving or redrawing... went to tvcon to :KILL it. No response from tvcon either. Within a few seconds, the Pi stopped responding completely - keyboard and mouse locked. 

At this point, I noticed a suspicious amount of heat coming from the Pi. Poked gently at the SOC - and burned my fingertip! Even the metal housings for USB/ethernet were uncomfortably hot. 

Utterly unusable without a fan!

Mike 

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/81773111-32ef-4f7c-bcf4-6864981a3647n%40googlegroups.com.

--

http://www.corestore.org
'No greater love hath a man than he lay down his life for his brother.
Not for millions, not for glory, not for fame.
For one person, in the dark, where no one will ever know or see.'

[oscarv]

Mike,

At this point, I noticed a suspicious amount of heat coming from the Pi. Poked gently at the SOC - and burned my fingertip! Even the metal housings for USB/ethernet were uncomfortably hot. 

Hmm - that is the sort of load I put on the Pi 12 hours per day. The Pi 5 does get hot, but the fan never whirrs up to more than its lowest level.

Did you put insulation between Pi and PiDP board?? Check the Building instructions, this is majorly important.

Kind regards,

Oscar.

[oscarv]

Added fix for the 'pdp view', time to do a 'git pull' to update things and run through the '/opt/install/install.sh' dependencies section.

Rich Cornwell just completed his new KA10 simulator, so that will be in very soon. It fixes quite a few rough edges.

[Mike Ross]

My PiDP doesn't arrive until next week - this was standalone Pi ITS testing, without the PiDP. 

And I discovered some more. 

I don't have a fan hat or CPU cooler for the Pi 5 - so I improvised, pointing a desk fan at it. This kept temperatures reasonable. 

I could run the emulation, run PEEK on tvcon continuously for hours; that was all stable. Same with ADVENT. But if I ran LORENZ using tvcons and 340, it would still totally lock the Pi up within a few minutes, same as last time - no response to mouse or keyboard, totally locked up, could only power off. And heat wasn't the cause - the temperatures remained reasonable this time, because of the fan. It seems to be specific to the LORENZ program. 

Not sure what's going on but I suspect something like a memory leak, either in the program code or in the -10 / tvcons / type340 emulation. Can you run LORENZ for an extended period with no problems??? 

Cheers

Mike 

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/8ce33a1a-2aa6-4edb-954c-cd429a8e61fen%40googlegroups.com.

[johntk...@gmail.com]

I have a heatsink and fan (the Raspberry Pi ones) on my 5, and Lorenz works fine. 

[Lars Brinkhoff]

Must be that weather simulations are inherently unstable and prone to the butterfly effect.

[Oscar Vermeulen]

Mike,

On Tue, 12 Mar 2024 at 14:52, Mike Ross <tmfd...@gmail.com> wrote:

Not sure what's going on but I suspect something like a memory leak, either in the program code or in the -10 / tvcons / type340 emulation. Can you run LORENZ for an extended period with no problems??? 

It runs a long time - at least, I used it at Vintage Computer Festivals as my demo. (Ah, but that was not under Wayland - suspicion)

But this might not be a bug hunt we need to chase in to, Richard Cornwell has just released a new version of the simulator. And (pointer to the fact you may be on to something) my previous fix-one-bug-with-another to get the type340 more pliable is gone from his version. The type340 window is now also resizable.

So let's see if you still have this problem in a day or two! 

Once you have the actual PiDP-10 hardware, consider adding a 12cm PC fan to the back instead of the little fan that they have for the Pi 5. It sucks the air out of the machine, and the only place air can come in is around the Pi. So it seems very effective. And adds a mainframe soundscape.

One program to avoid is 'pdp con'. Use 'pdp telcon' or much much nicer, 'pdp view' instead. 

'pdp view' is basically the left-hand side of this screen shot. You get to keep an eye on simh (bottom right), on the PDP-11 simulator for the Knight TV system, and above a telnet connection with Teletype font. Above that, either a linux terminal or just telnet localhost 10018 to get into the PDP-10.

Haven't had so much fun in a long time :-)

'pdp con' was a quick hack for VCF events, but under Wayland it is stunningly power hungry for what it does. It needs a rewrite.

Kind regards,

Oscar.

[oscarv]

New update, time for a 'git pull', and run the install script again (running it multiple times does not hurt anything BTW) although you can skip most parts, not the "DEC flavour to the desktop' thing.

This is a big one, it adds Richard Cornwell's new PDP-10 simulator. You can revert to the old one through the install script, but there is no need for that.

One change: when you power up with ITS, you have to:

press STOP

press READ IN

then follow through in the Teletype with 'ITS<return><esc>G

The type 340 display window is now resizable. But as it has a 1024*1024 resolution, you lose pixels when sizing down.

And when you power down the Pi and the PDP-10 is still running, it used to be that you get stuck waiting for 90 seconds. Now, press PAR STOP and STOP together and the PDP-10 gets powered off.

Kind regards,

Oscar.

[Steve Pitcher]

I'm getting excited.. I've just swapped my Pi 5 and Pi 4, so now the 4 is running my PiDP-11, and the 5 is for the 10.  I'm playing with the 10 simulation, but so far, struggling.

I'm waiting eagerly (patiently ???) for the PiDP-10 kit.  Hopefully, the 2nd batch.

- stp

[Mike Kostersitz]

Getting an error when installing clean

/opt/pidp10/install/install.sh: line 117: unzip: command not found

-----------------------

Use currently installed PDP-10 simulator (yes makes sense)? y

/opt/pidp10/install/install.sh: line 132: Leaving: Permission denied

Download PDP-10 emulator source code? y

Cloning into 'pidp10'...

remote: Enumerating objects: 707, done.

remote: Counting objects: 100% (99/99), done.

remote: Compressing objects: 100% (66/66), done.

Mike

---

From: pid...@googlegroups.com <pid...@googlegroups.com> on behalf of Steve Pitcher <stpit...@gmail.com>
Sent: Wednesday, March 13, 2024 11:23 AM
To: PiDP-10 <pid...@googlegroups.com>
Subject: Re: [pidp-10] PiDP-10 Initial Software Feedback / Issues

 

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/68c63c06-6da6-4c75-a714-c5b325b35246n%40googlegroups.com.

[Mitchell Wolrich]

Oscar,

Anyway to get the date correct in ITS?  It seems to be 2 years and 1 day behind, mine is showing 3/14/22 14:44

(today is 3/13/2024 2:44PM, so time is correct)

FYI with "chess2" running on the knight Terminal, and  ':LORENZ' running on the Type340, the Pi5 all 4 cores are nearly maxed out, see picture.

Also, I am seeing the emulatore crash, but it seems to recover..  I did a "git pull" just before loading all of this up, and re-ran the install.sh script.

Mitch

[Mike Ross]

Replies inline... 

On Wed, Mar 13, 2024 at 8:25 PM Oscar Vermeulen <vermeul...@gmail.com> wrote:

Mike,

On Tue, 12 Mar 2024 at 14:52, Mike Ross <tmfd...@gmail.com> wrote:

Not sure what's going on but I suspect something like a memory leak, either in the program code or in the -10 / tvcons / type340 emulation. Can you run LORENZ for an extended period with no problems??? 

It runs a long time - at least, I used it at Vintage Computer Festivals as my demo. (Ah, but that was not under Wayland - suspicion)

But this might not be a bug hunt we need to chase in to, Richard Cornwell has just released a new version of the simulator. And (pointer to the fact you may be on to something) my previous fix-one-bug-with-another to get the type340 more pliable is gone from his version. The type340 window is now also resizable.

So let's see if you still have this problem in a day or two! 

OK I'll do a pull in the next day or so and see where we get to. 
 

Once you have the actual PiDP-10 hardware, consider adding a 12cm PC fan to the back instead of the little fan that they have for the Pi 5. It sucks the air out of the machine, and the only place air can come in is around the Pi. So it seems very effective. And adds a mainframe soundscape.

Unfortunately it has to run in a quiet environment, so the last thing I need is a "mainframe soundscape"! I'll try a passive heatsink or small quiet fan on the Pi and see how I make out!
 

One program to avoid is 'pdp con'. Use 'pdp telcon' or much much nicer, 'pdp view' instead. 

I did try pdp view a few times - it opened, but always just gave me four windows of Linux shells. No PiDP display or activity. Maybe I'm running it at the wrong time or something? Are you supposed to start it before you start ITS, or after? 
 

'pdp view' is basically the left-hand side of this screen shot. You get to keep an eye on simh (bottom right), on the PDP-11 simulator for the Knight TV system, and above a telnet connection with Teletype font. Above that, either a linux terminal or just telnet localhost 10018 to get into the PDP-10.

Haven't had so much fun in a long time :-)

'pdp con' was a quick hack for VCF events, but under Wayland it is stunningly power hungry for what it does. It needs a rewrite.

Kind regards,

Oscar.

[johntk...@gmail.com]

Upgrades, excellent.

press STOP

press READ IN

Is this supported on a hardware-less just-a-pi system?

My package is still days away :-(

[Lars Brinkhoff]

Mitchell Wolrich wrote:

Anyway to get the date correct in ITS?

Run PDSET, use the D command to enter the date, and finally !. to set the system clock.  It should look like this:

*pdset^K

[chatter]

*240313D

*!.

*Q

:KILL

[Michael Thompson]

Updating the ITS source from install.sh didn't work. A "git pull" in the ITS src directory did work.

Download ITS project source code? yes
fatal: destination path 'its' already exists and is not an empty directory.

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/CANrY3pYQ3_ZCxmjwy7dg6DJ1690ADV9wdFFBeGQKuJrZOpuhJg%40mail.gmail.com.

--

Michael Thompson

[oscarv]

Mike,

On Wednesday, March 13, 2024 at 1:34:28 PM UTC-5 mikekos...@gmail.com wrote:

Getting an error when installing clean

/opt/pidp10/install/install.sh: line 117: unzip: command not found

That will then not be on a Pi - it has unzip by default?

I will check how the install on X86 fares on Saturday again, it was perhaps a bit ambitious to claim 'it just works' without checking every day with all the changes. I'm travelling for 24 hrs, so Saturday is the next day to work on this.

Kind regards,

Oscar.

[oscarv]

Mitch,

Yes, the type340 adds quite a lot to the load. I did the same as you, here's my screen shot (just for fun):

One core fully taxed, the three others around or below 50%.

But that's only fair to the Pi, it has to work for its electricity? The system remains nicely responsive. 

file:///home/x/Documents/x2/github/20240314_00h54m51s_grim.png

(I hope that image loads, Google Groups decided today to not let me see the pictures I load up, odd.

>> Also, I am seeing the emulatore crash, but it seems to recover..

You refer to the output on the Teletype? No, that's not the emulator crashing, it is ITS daemons doing maintenance. (Lars will correct me here for choice of words?)

Anyway to get the date correct in ITS?  It seems to be 2 years and 1 day behind, mine is showing 3/14/22 14:44

We got the century correct! But yes... the year.

 

[oscarv]

John,

On a Blinkenless Pi you're supposed to do 

pdpcontrol stop (if it was running)

I recommend a brief pause, telnet sometimes needs some time to unblock. I fixed that but it is not in the repository yet.

pdpcontrol start 1

The '1' tells the script that there shall be no looking at front panels. It just runs systems/its/bootscript.pi instead of bootscript.pidp, and skips the read in switch.

On a PiDP with front panel, pdpcontrol start (sans 1) would be preferred. That is in the ~/.profile for autostart!

Kind regards,

Oscar.

[Mitchell Wolrich]

Well, I did a "git pull", re-ran the install script and rebooted.. I see the icons on the "other" tab, but the "telcon" part of PDP-10 View now refuses to connect,..  I am impressed, as an ancient "Luser" of this very ITS system at the MIT-LCS lab, 9th floor, 545 Tech Square in Cambridge back in 1976/77 when I was in high school (dating myself now).  All these old ":" commands are coming back :-)

Here is the message in "telcon", from the PDP-10 View section

Trying 127.0.0.1...
telnet: Unable to connect to remote host: Connection refused

Then of course at the top in grey "The child process exited normally with status 1", hitting "relaunch" nada

Mitch

[oscarv]

Mitch,

On Thursday, March 14, 2024 at 1:19:48 AM UTC-5 mwol...@gmail.com wrote:

Trying 127.0.0.1...
telnet: Unable to connect to remote host: Connection refused

I see this when do pdpcontrol stop and pdpcontrol start very quickly after each other, and then, occasionally.

--> Is that what you describing? It only happens when you quickly restart?

If it is, then:

Yes, I see this occasionally. That is, on the Pi, when I restart within a few seconds. 

On X86 ubuntu (which is not really the intended target, it just came more or less for free!), the pause between stop and start needs to be a bit longer. 

1) Another new git pull might improve matters. I forgot a -u in a telnet connection, a simh boot script.

2) A more radical solution is:

     sudo sysctl -w net.ipv4.tcp_fin_timeout=1

     It is a fix to the issue, but the issue should not be there. And doing this on a Pi is fine, but not on a work laptop.

--> or is this 'always', ie, you just don't get a telnet connection ever?

Then, you probably don't run on a Pi but on X86, you have no telnet. Could that be?

March is beta-month. But I'm receiving lots of encouraging screen shots, it seems the system is pretty stable now on a Pi(DP).

Kind regards,

Oscar.

[johntk...@gmail.com]

Thanks - all working great!

Definitely better than FedEx who are still sending my kit on a tour of the US, trying to get to every State at least once.

[Mike Ross]

Hi Oscar, 

PIDP-10 up and running!



Couple of wrinkles... the pdp view windows seem quite user-hostile; I can't seem to resize them vertically and they're taller than the display I'm using, some is lost off the bottom, and...

OK, I get tvcon, I do stuff, I run LORENZ, it draws nicely on the 340 as above... but, stupid question, how in heck do you stop it?! Going back to the tvcon shows the last command you entered - :LORENZ - but there's nowhere to type another command, typing :KILL should exit the program according to the instructions, but nothing happens, no characters are echoed back, it's like the terminal is locked up? 

So I just closed (X) the window completely. I then started a new instance of tvcon - except it wasn't new, it just brought me back to the same 'locked' instance with my last command - :LORENZ - still sitting there exactly as before. 

Oh and... I did a git pull and install 3 hours ago so I should have latest code, I thought the 340 window was supposed to be resizeable now, but I can only adjust the width, again it fills the entire screen from top to bottom and you can't get to the top or bottom edge to get a vertical resize cursor - very similar to the pdp view issue; both applications seem to 'force' the windows to be an inappropriate size for the screen resolution.  

Can confirm 'PAR STOP - STOP' works as described  :-) 

Mike 

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/97ce3323-25fb-4ecb-b52d-e204fd32096dn%40googlegroups.com.

[Mike Ross]

Followup: I got a message giving me the clue about stopping LORENZ: in the tvcon window, you first have to hit F1 to get a prompt, *then* you can issue commands, like :KILL 

Very much the ITS newbie here, you can tell  :-)

I got a workaround to the 'window size' issues I mentioned by switching to a higher res display, *and* flipping it into portrait mode:


Mike

[DR]

I know most of the time computers burn empty cycles and with the pi5
being fast, wonder if I can run my WeeWx weather stuff on the same
machine while playing with my 10 kit when I get it.


The WeeWx folks recommend now that their program be running in something
that is called a venv sort of like a protected mode and for ease of
updating, fixing, etc.


It sounds to me like the venv has the possibility of many instances of
it and wonder if it is something that can be or even should be done with
the SIMH software emulator for the 10 and all the toys that come with it?

Dale

[johntk...@gmail.com]

My type340 display opens up in a default state that too large for the DEC Mini terminal I've made, as it only has a screen resolution 1152 x 864.  There is no way to grab the titlebar or resize it: that I know of.

Please make it possible to resize this window, or make it start smaller or launch?

Also, is it possible to change the F1 mapping on other terminals? I can't get the F1 key on this Pi terminal to work :-/

[johntk...@gmail.com]

Phew: CTL ALT (up, down, left, right) will resize windows on the Pi. Good to know!

[Patrick Giagnocavo]

The "venv" part is something specific to the  Python programming language. 

Should be easy to set up on any version of Linux that will run on the Pi.  

It is not specific to Pi but venv is instead something that Python does on all CPU architectures.

Cordially

Patrick

--
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 on the web visit https://groups.google.com/d/msgid/pidp-10/b3ab319c-dba6-4c0a-b6aa-9e74248044b1%40gmail.com.

[Mitchell Wolrich]

Oscar,

Yes, delaying a minute or two fixed that issue.

Since the launch of the new emulator, when I launch "pdp type340" from a VT52 window, the emulator/ITS semi crashes, ITS is still running, but is states "ITS Revived", doesn't seem normal.. here's some text from when it crashes

SIM-INFO: Eth: opened OS device nat:tcp=2123:10.0.2.15:23,tcp=2121:10.0.2.15:21,tcp=|
2195:10.0.2.15:95                                                                    |
%SIM-ERROR: Eth: Can't operate asynchronously, must poll.                            |
*** Build with USE_READER_THREAD defined and link with pthreads for asynchronous ope|
ration. ***                                                                          |
/opt/pidp10/systems/its/boot.pi-113> b ptr                                           |
%SIM-INFO: Running with Buffered Console   |
|
Simulation stopped, PC: 000017 (AOJA 0,17)
^Eset dpy ena
sim> set dpy ena
co
sim> co
pi@PiDP10-ITS:~ $ 

[oscarv]

Hi,

On Friday, March 15, 2024 at 9:33:38 PM UTC mwol...@gmail.com wrote:

Yes, delaying a minute or two fixed that issue.

This will reduce the waiting time to a second or two. But only do it on a Pi, not on a production Linux machine; shortening the telnet wait delay is considered adding a vulnerabilty:

sudo sysctl -w net.ipv4.tcp_fin_timeout=1

Since the launch of the new emulator, when I launch "pdp type340" from a VT52 window, the emulator/ITS semi crashes, ITS is still running, but is states "ITS Revived", doesn't seem normal.. here's some text from when it crashes

No, it's not ITS crashing and recovering, it's ITS noticing that the time has mysteriously stood still whilst the simulator was stopped to enabe the type 340 simulator.

Kind regards,

Oscar.

[Mitchell Wolrich]

Is their a fix for that?  What about changing ITS to use the host machines date/time function?  that would insure that the time keeps moving forward, and fix the TYPE340 "ITS Revived!" issue as well, since time would keep moving on..

The other issue is that when ITS is started each time, the date/time is off by 2 years, 1 day..  we saw the commands to adjust it, but of course it isn't saved..  using the hosts date/time would solve that issue as well.

Thanks,

Mitch

[Heinz-Bernd Eggenstein]

> The other issue is that when ITS is started each time, the date/time is off by 2 years, 1 day..  we saw the commands to adjust it, but of course it isn't saved..  

It *is* saved, kind of...

>using the hosts date/time would solve that issue as well.

The sim  kind of does this already. 

So when I stop the simulation and later restart it, it does remember the correct year, so no need to re-set it again..unless you re-install the disk image in between. Have you tried it?

Here is why:

As far as I understand the code, the way the original clock device is simulated in the simulation is that it indeed is synchronized with the time from the host OS, but the problem is that that vintage device could only represent time as ticks starting from the beginning of the current year. That is its internal representation of the real-time-clock. ITS had to figure out separately what the current year was, and while booting it would try to query its fellow computers in the network of the MIT AI lab ... but we are not in Cambridge anymore, so no go. The last falback is to look at the date a particular file was last written to, assuming that must have happened very recently (==> leave ITS running at New Years eve). And this fallback should still work. 

It fails initially when it sees a file date of 2022 when booting for the first time from a freshly installed disk image, and that is also the reason why it is off by one year and one day: 2022 was not a leap year, 2024 is, and internally the clock counts ticks since midnight Jan. 1st. But once this magic file is touched in 2024 , all is well.

HB

[Mitchell Wolrich]

Thanks, that worked..  I created my login MITCH, put a file in with emacs and saved it, even filled out the :INQUIR from here:

https://github.com/larsbrinkhoff/its-manual/wiki/Adding-a-new-ITS-user

Then did a proper ITS shutdown, per instructions:

Type ':LOCK' and hit return. Then, '5down', Y, CTRL-C. Hit return, and type ':LOGOUT' in case you had indeed properly logged in at the beginning. Wait until you are in the DDT/Z debugger. Now you are done. On the linux command line, you can power down the PDP-10 with 'pdpcontrol stop'. No surprise, 'pdpcontrol start' will bring you back.

waited until I was in DDT, then did a "pdpcontrol stop"

brought it back up, and it now ist' all good :-) thanks.  now who is this "TARAKA" dude, and how is he getting into my system, guessing a "DAEMON" :-)

Mitch

LOG bringing it backup below:

Trying ::1...
Connected to localhost.
Escape character is '^]'.

Connected to the KA-10 simulator CON-TELNET device

 DSKDMP
ITS
$G
SALVAGER.317
.MSGS.; NO FILES, USER DIRECTORY DELETED
.TEMP.; NO FILES, USER DIRECTORY DELETED

IT IS NOW  5:41:19 PM EST, SATURDAY, MAR 16, 2024

KA ITS 1651 IN OPERATION AT 17:41:19
KA ITS 1651 SYSTEM JOB USING THIS CONSOLE.
TV 11 WENT DOWN -- 17:41:19
 LOGIN  TARAKA 0 17:41:20
TOP LEVEL INTERRUPT 200 DETACHED JOB # 3, USR:TARAKA CNAVRL 17:41:21
172660/ 0       1       TARAKA 17:41:24
TV 11 IS UP - 17:41:24
IT IS NOW  5:41:48 PM EST, SATURDAY, MAR 16, 2024

[Lars Brinkhoff]

Heinz-Bernd Eggenstein wrote:
> ITS had to figure out separately what the current year was, and while booting it would try to query its fellow computers in the network of the MIT AI lab ... but we are not in Cambridge anymore, so no go.

We should be able to have a time server running on the host. The
latest version from https://github.com/PDP-10/its comes with a local
Chaosnet set up out of the box, so when the PiDP-10 software is
updated, that can be implemented.

[Lars Brinkhoff]

Mitchell Wolrich wrote:
> now who is this "TARAKA" dude, and how is he getting into my system, guessing a "DAEMON" :-)

That is indeed the name of a demon, which comes from the book Lord of
Light that was popular in the 70s.

[dab]

On Sunday, March 17, 2024 at 1:50:55 AM UTC-4 Lars Brinkhoff wrote:

We should be able to have a time server running on the host.

Seems to me that it's time that ITS had its own implementation of NTP. 

[oscarv]

Since just now, Lars Brinkhoff helped me to implement the Chaosnet-bridge. Git Pull will update for you. And do a '/opt/pidp10/bin/supdup its' to see if it works.

First step in PDP-10 networking!

Kind regards,

Oscar.

[Heinz-Bernd Eggenstein]

Nope, 

after 

git pull

and

cd /opt/pidp10/bin

./supdup its

I just get:

===

Trying its ...supdup: connect: Connection refused
its: unknown host.

===

Do I have to repeat parts of install.sh?

I don't think the iTS image has changed, right? I'm beginning to create new files in there and would have to back them up first.

Do I have to use any of the other new exec like gateway.sh or change any of the new *.conf files in there?

Cheers

HB

Cheers

HB

[Lars Brinkhoff]

Heinz-Bernd Eggenstein wrote:
> Trying its ...supdup: connect: Connection refused

> I don't think the iTS image has changed, right? I'm beginning to create new files in there and would have to back them up first.

Yes, the image has changed. The old one didn't have Chaosnet enabled.

It's a good idea to back up your files. Once Chaosnet is there,
backing up and transferring files is even easier.

[Heinz-Bernd Eggenstein]

So I updated the image,  restarted ITS (nice! it's clock is now in 2024).

I put on a vinyl of "Let it bleed" by the Rolling Stones (ca 1968) , just to make sure ...

Started supdup (which I think is short for "Super Duper", isn't it?  )

an got this:

=========

KA ITS.1651. DDT.1548.
TTY 53
2. Lusers, Fair Share = 99%
Welcome to ITS!

For brief information, type ?
For a list of colon commands, type :? and press Enter.
For the full info system, type :INFO and Enter.

Happy hacking!
:TIME
(Please Log In)
The time is 15:51:29 EST.
Today is Tuesday, the 19th of March, 2024.
KA ITS 1651 has run for 5 minutes, 40 seconds.

:KILL
*

===============

Wonderful!! Thanks a lot!

HBE

[oscarv]

HB,

On Tuesday, March 19, 2024 at 7:56:35 PM UTC hbep...@googlemail.com wrote:

I just get:

===

Trying its ...supdup: connect: Connection refused
its: unknown host.

===

Do I have to repeat parts of install.sh?

Fixed now. I forgot to update the ITS system disks. Download them again and you get a working Chaosnet (bold claim, but quite sure).

So you can use 'bin/supdup its' but also mlftp. Which was new to me, but Lars told me about it yesterday.

bin/mlftp is a sweet file exchange program between Linux and ITS, so you exchange files like so:

bin/mlftp -w its test ~/test

...will store the file in the .TEMP. directory of ITS. Run mlftp without arguments to see its options.

or

mlftp -w its -- "-READ- -THIS- GAMES;" read-this.txt

... note: GAMES; , the colon denotes the directory name

I don't think the iTS image has changed, right? I'm beginning to create new files in there and would have to back them up first.

mlftp! Wow, just in time!

 

Do I have to use any of the other new exec like gateway.sh or change any of the new *.conf files in there?

I'm just exploring gateway.sh, any tips welcome!

Oscar.

[Heinz-Bernd Eggenstein]

Wow, being  able to ftp into ITS from Linux will be very useful now that people actually create stuff in their new PiDP-10s! 

Cheers

HB

[Richard C]

I had some trouble with FTP on KA10 ITS. It would not start the server.

[Lars Brinkhoff]

Heinz-Bernd Eggenstein wrote:
> I put on a vinyl of "Let it bleed" by the Rolling Stones (ca 1968) , just to make sure ...

I think the AI lab was more of a Beatles shop. And Bach of course.

> Started supdup (which I think is short for "Super Duper", isn't it? )

The full name is something like "super duper image mode telnet". To
fully understand this, I have to launch into a lecture about ITS
programming. Yay!

ITS allows a user space program to have 16 channels for I/O. If you
are familiar with Unix, you can think of them as file descriptors.
Each channel has a mode. "Image mode" means to treat data mostly
literally, as opposed to "ASCII mode" which does some character
interpretations. The exception in image mode are a few things like ^Z
which is still an escape character. To treat all data literally
without exception, you have to engage "super image mode".

The other piece of the puzzle is that ITS has an internal buffer for
each user terminal. The buffer stores text and internal cursor
movement codes, which are terminal-independent. The SUPDUP protocol
uses these internal codes directly, sending them to the client side
as-is. So in a sense this is a level of literalness even above "super
image mode", hence the super-duper adjective.

[Oscar Vermeulen]

Richard,

On Wed, 20 Mar 2024, 03:07 Richard C, <rco...@gmail.com> wrote:

I had some trouble with FTP on KA10 ITS. It would not start the server.

My first suspicion is, did you let install download the ITS disk images again?

Kind regards,

Oscar

[Michael Thompson]

I have been playing with the console switches while the emulator is running, but without an OS running.

There is a typo in the emulator text window when the EXAMINE switch is pressed:

Examime 000000 000000000000

Pressing EXAMINE displays the correct address from the switches on the text console, but displays the data switches not the memory data.

It displays the correct values on the LEDs.

Pressing EXAMINE NEXT displays nothing on the text console, but displays the correct values on the LEDs.

Pressing DEPOSIT displays the correct address and data on the text console, and on the LEDs.

Pressing DEPOSIT NEXT displays the correct address and data on the LEDs, but nothing on the text console.

Pressing SINGLE INST and then START causes the RUN LED to go on and put the emulator in the RUN state instead of executing a single instruction.

Pressing SINGLE CYCLE and then START causes the RUN LED to go on and put the emulator in the RUN state instead of executing a single cycle of an instruction.

--

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 on the web visit https://groups.google.com/d/msgid/pidp-10/CAJAwMc2DJPwbXDF%3DSRRfpK8KFLf%3DhyCGC3fc%3D5c_qSmYF6ujVA%40mail.gmail.com.

--

Michael Thompson

[Lars Brinkhoff]

Michael Thompson wrote:
> There is a typo in the emulator text window when the EXAMINE switch is pressed:
> Examime 000000 000000000000

That, and the corresponding message for deposit, look like debug
printfs in the code. I guess it can be removed now.

> Pressing SINGLE INST and then START causes the RUN LED to go on and put the emulator in the RUN state instead of executing a single instruction.

That was the case in an earlier version of the emulator, but Rich
fixed that. Please try updating to the latest version and see if
SINGLE INST works better. I have tested almost all switch functions,
and I think they mostly work ok. I wrote up the testing steps here:
https://github.com/obsolescence/pidp10/wiki/PiDP%E2%80%9010-testing

For this case, I do wonder if PC should display the instruction about
to be executed, or the next.

> Pressing SINGLE CYCLE and then START causes the RUN LED to go on and put the emulator in the RUN state instead of executing a single cycle of an instruction.

I believe SINGLE CYCLE has not been implemented at all.

[Richard C]

Which emulator are you using?

Single Cycle does not work since the emulator does not emulate cycles, but instructions as a whole. Eventually I might do a cycle accurate emulator for a KA10.

Examine and Deposit should not print out anything on the simH console. Readin, Continue, Start will print something.

Rich

[Michael Thompson]

I updated everything and the SING INST switch now works as expected. I recently updated everything, so it must have been fixed in the last few days.

You could make the SING CYCLE instruction work just like the SING INST switch. It would not be correct, but it would be better than not being implemented.

The debugging information on the simulator console is OK for DEP, START, and CONT. It is nice to see the disassembly of the instructions on the simulator console when single-stepping instead of just LEDs. I think that improving the debugging information on the simulator console would be better than removing it.

The XCT switch works once. If you press it again it executes the instruction at PC+1 instead of from the switches.

--

Michael Thompson

[Richard C]

I guess I could make Single Cycle do same as Single Instruction.

I will look into the XCT function.

Rich

[steve...@gmail.com]

I'm trying to install the software on a Pi 3B+ without the PiDP-10 hardware just to test run the software. I installed the 64 bit OS (Full). On following the installation instructions on the PiDP-10 Build Instructions page I said "yes" to everything. After rebooting the pi, none of the pdp commands work. they all return 0:

pi@raspberrypi:~ $ pdp
0
pi@raspberrypi:~ $ pdpcontrol stop
PiDP-10 is already stopped.
pi@raspberrypi:~ $ pdpcontrol start 0
Starting PiDP-10
screen -dmS pidp10 ./pidp10 /opt/pidp10/systems/hills-blinky/boot.pi
0
pi@raspberrypi:~ $ 

If I cd into /opt/pidp10/bin and try ./pidp10, I get:

pi@raspberrypi:/opt/pidp10/bin $ ./pidp10
./pidp10: /lib/aarch64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by ./pidp10)
./pidp10: /lib/aarch64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found (required by ./pidp10)
pi@raspberrypi:/opt/pidp10/bin $ 

It seems that somewhere along the way these libraries aren't being installed. Is there a fix or a workaround?

Thanks,

-- steve

  

[Mitchell Wolrich]

change pdpcontrol start 0 to pdpcontrol start 1

1 is the option to start ITS, 0 just runs Blinky, 2 is for TOPS10

Mitch

[steve...@gmail.com]

Mitch,

Thanks but the issue isn’t the number parameter. I’ve tried all three. The issue is the missing libraries. How are they supposed,to get installed because they aren’t.

— steve 

[Lars Brinkhoff]

pi@raspberrypi:/opt/pidp10/bin $ ./pidp10
./pidp10: /lib/aarch64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by ./pidp10)
./pidp10: /lib/aarch64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found (required by ./pidp10)
pi@raspberrypi:/opt/pidp10/bin $

It seems that somewhere along the way these libraries aren't being installed. Is there a fix or a workaround?

 I would suggest trying to rebuild the emulator from source code.  You'll need to install all dependencies, but the install script can do that for you.  Once you have the source code downloaded, you can just type "make pdp10-ka PIDP10=1".

[Heinz-Bernd Eggenstein]

Which version of 64 bit Raspberry Pi OS did you install?

cat /etc/os-version

will tell you, and if it's not "Bookworm", that might be the problem. Bookworm runs on all Raspberry Pi Models, including your 3B+, while the reverse is not true: the Raspi 5 needs Bookworm. At the start, the new "Bookworm" was still a bit rough around the edges (some problems with popular VNC clients), but the latest update fixed many problems. So currently I see no reason why people should not use Bookworm with their PiDP-10, with or without the blinkenlights panel.

HB

[steve...@gmail.com]

HB,

Thanks for the hint. I used the "recommended" 64-bit Pi 3 OS from the MicroSD writer program on MacOS. First, it turns out to be:

cat /etc/os-release

Nonetheless, it returns:

pi@raspberrypi:/opt/pidp10/bin $ cat /etc/os-release
PRETTY_NAME="Debian GNU/Linux 11 (bullseye)"
NAME="Debian GNU/Linux"
VERSION_ID="11"
VERSION="11 (bullseye)"
VERSION_CODENAME=bullseye
ID=debian
HOME_URL="https://www.debian.org/"
SUPPORT_URL="https://www.debian.org/support"
BUG_REPORT_URL="https://bugs.debian.org/"
pi@raspberrypi:/opt/pidp10/bin $ 

I went ahead and rebuilt the code per Lars Brinkhoff's mail and it still fails in exactly the same way. I will re-install with Bookworm this time. Stay tuned . . . 

Cheers,

-- steve

[johntk...@gmail.com]

Make sure to use the latest version of the Raspberry Pi Imager: https://www.raspberrypi.com/software/

That said, when setting up older devices such as the 3B, the software defaults to installing Bullseye. To me hints that the newer OS isn't supported and if you can install it at all, you're on your own. Perhaps it is time for a Pi 4 or 5 upgrade, to make the most of the raw power of a PiDP-10.

[steve...@gmail.com]

John,

I am using the latest version of the Raspberry Pi Imager. And yes, it does default to Bullseye for Pi 3's. I told it to image the microSD card for a Pi 5 and got the right version. I am able to bring up the KA10 simulator now. A Pi 5 is on order, it should arrive this Thursday but I did want to try playing with the OSs before the kit and the Pi 5 arrive.

-- steve

[Richard C]

I checked the XCT function and it execute the switch register without problems. I will be pushing a version that makes single cycle behave like single instruction.

Rich

On Friday, March 22, 2024 at 10:42:28 AM UTC-4 michael.9...@gmail.com wrote:

[Michael Thompson]

I am using pidp10.

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/72184d84-2a85-43e7-aa58-2faa1c029c53n%40googlegroups.com.

--

Michael Thompson

[Michael Thompson]

The KA10 Maintenance Manual says that the PC will not increment when XCT is used. Every time XCT is pressed on the front panel the simulator increments the PC.

--

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 on the web visit https://groups.google.com/d/msgid/pidp-10/35307e68-90fb-4709-9bce-7775b3029f04n%40googlegroups.com.

--

Michael Thompson

[Michael Thompson]

>>> Are you using Panama 5 version or my version? 

 pdp10-ka dated 19-MAR-24

On Wed, Mar 27, 2024 at 9:29 AM Richard C <rco...@gmail.com> wrote:

Are you using Panama 5 version or my version?

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/37497e4e-147e-4465-aae7-b72f6660e85bn%40googlegroups.com.

--

Michael Thompson

[Michael Thompson]

I just updated the software on my PiDP-10. TOPS-10 gets part way through boot and then hangs. Usually at this point all of the PIs are enabled, but this time only channel 4 is enabled. There is a request for channel 7, which I believe is the console, but since it is not enabled it is not being processed. I reran install.sh, but it didn't make a difference.

Prior to this update TOPS-10 had been running OK for several weeks.

Connected to the KA-10 simulator CON-TELNET device

KA603 04-21-78
WHY RELOAD: sched
DATE: 28-mar-94
TIME: 1904

STARTUP OPTION: no initia

%DPA4 IS OFF-LINE

%DPA5 IS OFF-LINE

%DPA6 IS OFF-LINE

%DPA7 IS OFF-LINE

--

Michael Thompson

[Michael Thompson]

I tried the simulator pdp10-ka instead of pidp10 and TOPS-10 booted OK.

I tried the simulator pidp10 again and TOPS-10 would not finish booting.

--

Michael Thompson

[Richard C]

I only support the version that I have pushed to Oscars repository. You should have a link in the src directory to my repository.

Also panama5 code is about 2 to 3 years old and lacks many patches and functions.

[Michael Thompson]

That’s nice to know. I am running the normal software from Oscar’s repository with no changes and no modifications. My PiDP-10 ran TOPS-10 OK until I updated it from Oscar’s GitHub site tonight. Obviously the problem is mine.

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/23329492-8945-46f6-b946-3dad0e5acd5fn%40googlegroups.com.