ND-10 (was Re: DMA is obsolete)

MitchAlsup1 wrote:

It is more like the Peripheral Processors of CDC 6600 that run
ISA of a CDC 6600 without as much fancy execution in periphery.

I seem to remember the CDC 6000 series as 60 bit ISA, with the PPUs
being 12-bit. Is that wrong?

On 2025-05-03, Terje Mathisen <terje.mathisen@tmsw.no> wrote:

Similar timeframe: The ND10 minis were popular in process control, CERN
bought a brace of them.
>
When they later came out with the larger ND100 and then ND500 machines,
the latter had a 100 (or 10?) as a front-end IO processor, partially
required because the original ND10 came with a very early version of
SINTRAN os which didn't have proper/complete IO support, so customers
had written machine code to handle it.
>
The 500 wasn't machine code compatible, so all such IO routines then had
to run on the front-end processor.

My first job after I finished my apprentice years at the Copenhagen
University Computer Center, was at a bespoke engineering house mostly
interfacing specialty equipment at research labs. My first project there
(fall 1985) was writing a SINTRAN device driver for an A/D scanner with
a hundred or so channels. Debugging was fun. SINTARN was written in
PL-10, a programming language invented for the purpose. The command
language was unexpectedly smart, allowing many abbreviations for the
commands. I think the same abbreviation rules also applied to filenames.

The most memorable thing about the project, was that this was a very
early use of solid-state memory. The SRAM chips used, had a bug that
sometimes reverted bits to a state that they had earlier stored for a
prolonged amount of time. Really fun when tha affected memory word was
an allocation bitmap for memory pages or for disk sectors!!

I enjoyed the weeks I spent in Oslo learning the OS operation and system
build procedure.