My PiDP8i likes to sleep

[ggrinton]

I am more puzzled than worried by the following observations.

My PiDP8i works beautifuly and (except as below) utterly reliably. Running with a Pi-W, I am connecting over WiFi using either SSH through PuTTY, or VNC, depending on what I want to do.

When I run OS/8 the system runs constantly and without interruption waiting at the console prompt. I have had it up for a period of over two weeks without interruption.

However, if I run Deeper Thought, the system runs for a few hours then appears to close down. All lights go off on the front panel, and the SSH connection drops out. VNC also says it cannot find the system. To all intents and appearances, it seems "dead". However, after a couple of hours everything comes to life again, and I can reconnect with SSH or VNC.

I have also written myself a 'blinkenlights' program, running in PDP8 mode, and when I  run that it demonstrates the same behaviour. On coming back to life it just picks up where it went dark.

So, I am assuming there is some sort of hibrnation going on. Is it possible it is tripping on temperature?  I think I remember reading that SimH is smart enough to know when the system is in a 'wait for keyboard' loop, and does not drive the CPU as hard, but the other programs I mentioned are running at full load all the time.

A quick search about RPi hibernation did not suggest anything relevant, so I wonder if others can shed light on this?

Thanks,

Geoffrey

[Alen Shapiro]

Hi Geoffrey,

I have a PiDP8 and a PiDP-11/70. They both run blinkenlight patterns, the PiDP 11/70 shows when it is in an OS wait-loop (BSD 2.11) and the PiDP-8 runs a modified assembler-code program. Both halt their lightshow after several weeks, the PiDP 11/70 does not crash/reboot the OS, but the blinkenlights operation stop after 50+ days. It might be running out of some resource (memory leak?), while the PiDP-8 stops in its tracks after a couple of weeks. Sometimes the PiDP-8 reboots, other times it halts and needs to be power cycled. I know these symptoms are not exactly what you’re seeing but I’m trying to list the odd symptoms in case they have any overlap and will help figuring out what’s going on.

My PiDP-8 is a few feet above a steam radiator. The driving Raspberry Pi-Zero 2W overheats and may be my issue with that box. I’ll add heat-sinks to that processor and see if “htop” (running in the Linux OS that hosts the PiDP-8 emulator) reports lower core temperatures. Does Linux wind down the CPU clocks when the temperature gets too high, I’ve not seen that behavior, just a halt? In my case, the PiPD-11/70 running BSD 2.11 is accessible via ssh/telnet when the lights stop, but the PiDP-8 Linux-host stops responding to ssh.

The most recent change I made to the hosting Linux was to make sure the SD card is not used for swapping, (changing the swap priority to zero stops the swap filling up over time (as it really is not needed)). I also turned off the desktop (which chewed through RAM over time for some reason). “htop” running in a ssh terminal was very useful showing the dynamic RAM/CPU use over time. “htop” might show you some clues when you get into your freeze state (last screen refresh it displayed could show you the most recent core clock speeds and temperatures).

Please let me know what you find out.

Regards,

—Alen