New subject: bit addressing, alignment, bus errors, and VM (was Re: PDP-10 arch question)

1 Nov 1998

...
 > Are you certain of that? Cray designed the CDC
160, and had a hand in the
> CDC 6000s, but I do not think the STAR. 
Seymour Cray had more than a hand in the design of the 6000 series.
Although he worked with some fine engineers (Thornton (sp?) is one name I
remember), I believe that the instruction set, architecture and packaging
scheme all primarily came from Seymour.

...
 After more than 25 years, I am not certain of anything.
 But if Cray had a
hand in the CDC 6000s, and he was still around CDC at the time the
STAR 100 was designed and built, I would certainly think that he was
also involved in the STAR 100 as well. 
As I remember (I was working on CDC's mainframe products at the time, not
the STAR), the STAR's were done by the people "left behind" after Seymour
left.  As such, there was an unspoken agenda to the effort:  The engineers
were trying to prove that they were just as good as Seymour, and the
executives were trying to prove that letting him leave wasn't a mistake.

...
 I thought I saw an article about
the STAR 100 recently which described the circular nature of the physical
rack which was required in order to reduce to length of the wires which
interconnected the different parts of the system. 
Again, I'm not positive about this, but I think that the STAR's that I
walked past on the test floor were in conventional rectangular boxes.  
----
John Dykstra                                jdykstra(a)nortel.com
Principal Software Architect                voice:  +1 651 415-1604
Nortel Networks                             fax:    +1 612 932-8549