🚀 go-pugleaf

In article <951@bbx.UUCP> usenet@bbx.UUCP (USENET manager it's really me: Russ) writes:
>Anyone else out there ever dink with the lights on the PDP-10?
>Not the console ones, they were a bit too easy; we played with
>the ALU register display.  Seems you could feed the ALU using 
>the right series of instructions.  Take all of the values, 
>shift 'em left a bit and run it through again... moving marqee!
>
>Memories of running diagnostics on the '10 and turning off the
>CPU room lights - better than Laserium and free to boot.

One of my jobs at C-MU was to help the engineering lab keep the KA10s in
running order and that sometimes included writing diagnositics and scope
loops.  Once we had a problem with our Ampex memories that only showed up
when the system was running, but never with DEC's diagnostics, which only
handle one peripheral per diagnostic then.

So I wrote a diagnostic that was really three processes:

1) Tested memory on the processor by writing and checking large chunks of
   memory with various patterns.
2) Tested memory via IO to the disk drives by transferring patterns to one
   of the spare cylinders on an RP02.  (15Mb disk pack)  It ran at the
   disk's interrupt level and could interrupt the CPU memory port test.
3) Tested memory via IO to the swap space on the RM10B. (Bryant drum. No, you
   couldn't execute from it.)  This could interrupt the other two.

No program in existance at the time made the machine look busier.  The DF10s
(memory channels) and the RP10 and RC10 (disk controllers) they were hooked
up to ran through patterns that you could follow, except the RM10 was a little
too fast to follow well.  The memories would sometimes idle, sometimes had a
single pattern, sometimes did two or three things at once.  The processor
looked like it could never make up its mind.  While I didn't exercise most
of the interrupt lights or relocation registers, the rest of the lights on the
three bays made the machine look busier than anything else I'd ever seen run
on the machine.

And of course, the effect was about three times better with the computer room
lights off.

By the time I got the program working, we had figured out the problem was a
intermittant short in a core plane and moved that bit to a spare plane.
However, my diagnostic was very useful as a quick system test we could run
without damaging anything the OS might need.

While I work on much bigger and faster (and better?) systems these days, I
miss the days I could control a room full of equipment and see it do my
bidding.
-- 

| A pride of lions              | Eric J Werme                |
| A gaggle of geese             | uucp: decvax!linus!alliant  |
| An odd lot of programmers     | Phone: 603-673-3993         |