On Nov 1, 2022, at 7:59 PM, Chris Zach via cctalk
<cctalk(a)classiccmp.org> wrote:
On 11/1/2022 11:49 AM, Ethan Dicks via cctalk wrote:
I don't think the 11/750 could have handled
that many users on DZ-11s.
I think this is why COMM-IO-P systems were sold. They basically acted like front end
processors for a stack of DZ11's to interface to a Unibus system.
Problem was those communications processor boards only had enough microcode to handle
either block lines or async lines but not both. When the fruitcakes at HKJC demanded that
they support the betting systems (polled block oriented devices) and async terminals (for
betting odds and displays and such) for the mighty Sha Tin Totalizer (multiple 11/74 type
systems and a boat-ton of 11/04 front ends connected by those weird IPC connections) all
hell broke loose.
Literally.
You'd think the obvious answer would be to use two KMC11s rather than just one...
The big win from those coprocessors is that they could offload not just simple DMA, but
also protocol processing like DDCMP or even BISYNC.
HKJC, now there's a designation I haven't seen in many years.
paul