12 May
2019
12 May
'19
8:34 a.m.
On Sat, May 11, 2019 at 10:17 PM Charles Dickman via cctalk <
cctalk at classiccmp.org> wrote:
On Sat, May 11, 2019 at 8:50 PM Steve Malikoff via
cctalk <
cctalk at classiccmp.org> wrote:
The SEL (later Gould, later Encore) line was designated in two ways
27/nn, 57/nn, 87/nn for their machines which ran their real time OS, MPX
30nn, 60nn, 90nn for their machines which ran their Unix variant, UTX.
IIRC, the nn was 20, 50, or 80. Larger numbers designating increasing
capacity.
The 3000, 6000, 9000 series were virtually identical to the real time
versions.
They just upped the first digit and took out the "/".
All were true 32 bit. ISTR you could update a 20 to a 50 with a board swap,
since the backplane bus was the same. The 80s were dual cabinet machines,
more boards for the CPU and more space for memory.
The 87/9000 series were ECL based. I can't remember what technology
was in the 2x and 5x series.
I could be remembering incorrectly but I think
the Gould PN6080 mini we
had exclusively for third year
comp sci at Macquarie Uni in the mid/late 80s was 32-bit made up of
AMD2900 family logic (2901 ALU's).
Purdue EE had I think 2 Gould PowerNode 9080s. I don't know anything about
the internals, but Purdue EE was doing development or testing for them. It
might have been a port of 4.3 BSD. As an undergraduate you could get an
account on en.ecn.purdue.edu for the asking and it was significantly
faster
than the overloaded Dual VAXen. It also crashed from time to time.
Perdue also had Gorge Goble. https://en.wikipedia.org/wiki/George_H._Goble,
who was a character.
-chuck