27 Feb
2009
27 Feb
'09
7:45 a.m.
>>>> "Chuck" == Chuck Guzis
<cclist at sydex.com> writes:
Chuck> On 26 Feb 2009 at 22:22, Dennis Boone wrote:
> I always assumed the 12-60 relationship existed
because the 12-bit
> architecture derived from a previous CDC design, and because it
> was somewhere between helpful and important that they be a
> multiple.
>
> But I haven't run into anything that talks about it. Anyone have
> a reference I should have RTFMed by now?
Chuck> Well, the PPU design is basically a reworked and regurgitated
Chuck> CDC 160A, a 12-bit design, shared among 10 sets of memory and
Chuck> registers (A,P,Q and K) so each PPU got its slot in the barrel
Chuck> every 10 minor cycles (or every 1 usec., which was also the
Chuck> speed of a core bank).
Rather loosely based on the 160A... while both are 12 bit machines,
and some of the instructions are somewhat similar, there are lots of
differences. Different addressing modes, different shifts, and
completely different I/O.
Chuck> But did the PP word size determine the CPU word size? I don't
Chuck> know.
It has to in some way, because the PP manipulates CPU words. So an
integer ratio is the sensible thing to do. Why 5 rather than some
other ratio is an interesting question.
Chuck> Maybe a 72 bit word length might have made more sense, as that
Chuck> would have matched the 7090 double-precision word size. On
Chuck> the other hand, everything about the 6000 architecture does
Chuck> "fit together" nicely.
Right. For example, 15 and 30 bit instructions pack well, while a
72-bit word would not have produced that nice packing (and 15 bits is
the right basic instruction size). Also, 60 bits is the single
precision float size, while double is 120 bits -- so you get a lot
more precision than the 7090.
Chuck> Maybe there's something in the CHM oral history archives.
I was hoping to find some in Thornton's book but I didn't see anything
there.
paul