21 Mar
2007
21 Mar
'07
7:28 a.m.
Brent Hilpert wrote:
Scott Quinn wrote:
CMU did a parallel machine like that using PDP-11/40s. They used 16
machines and 16 memory banks. It was called C.mmp. I think for me,
reconstructing that would be the ultimate in classic geek.
Also, while I haven't watched all that carefully, I'm sure there are others
who have also used crossbars to route inter-processor communication.
I even messed around with that approach a little.
BLS
(1) What is the history of the crossbar in computer architectures as a
mainbus replacement?
I recall reading an article in Annals of the History of Computing about a
late 60's era mainframe in which memory and multiple processors were
connected via some sort of crossbar/matrix switching arrangement.
I think it was a Burroughs machine but my memory could be wrong on that even. 
21 Mar
21 Mar
1:03 p.m.
"Brian L. Stuart" wrote:
Brent Hilpert wrote:
Not sure which way you are referring to the 11/40s/C.mmp being constructed, but
to clarify a bit, perhaps I should have said execution units instead of
processors. The machine I recall being described in the article was not
constructed by connecting up more of an existing or off-the-shelf processor,
rather the multiple execution units and crossbar matrix were inherent in
the design of the one main processor.
Scott Quinn wrote:
CMU did a parallel machine like that using PDP-11/40s. They used 16
machines and 16 memory banks. It was called C.mmp. I think for me,
reconstructing that would be the ultimate in classic geek.
Also, while I haven't watched all that carefully, I'm sure there are others
who have also used crossbars to route inter-processor communication.
I even messed around with that approach a little.
BLS
(1) What is the history of the crossbar in computer architectures as a
mainbus replacement?
I recall reading an article in Annals of the History of Computing about a
late 60's era mainframe in which memory and multiple processors were
connected via some sort of crossbar/matrix switching arrangement.
I think it was a Burroughs machine but my memory could be wrong on that even.
6486
days inactive
6486
days old
1 comments
2 participants
participants (2)
-
blstuart＠bellsouth.net -
hilpert＠cs.ubc.ca