5 May
2001
5 May
'01
10:01 a.m.
From: Richard Erlacher <edick(a)idcomm.com>
No question the 6502 was fewer gates than z80/6800 as that was the guys
to chase.
The 8080 was hard to hit as the 8080 required so much more external
support
that any of the 5V only cpus. It would be more fair to compare 6502 to
8085
in the gate count and die size derby.
believe the 650x core with only four internal
software-accessible
registers
(A,X,Y, SP) would have a substantially lower gate count
than a Z80,
which has
lots of register resources, (A,B,C,D,E,H,L,IX,IY,SP,
plus a second set
of the
same) and it does, but nowhere near the ratio that
these registers
suggest. Of
course there are several ways of looking at the
definition of "gate" but
it's
Think of registers as memory bits... Z80 those number about 208 where
6502 has far fewer.
However this is relative as a register can be one transistor and cap
(dynamic)
or a lot more for static. I believe the 6602 was a dynamic machine like
many
of the time.
odd that the ratio of gates consumed by each of these
cores doesn't
approach the
>2:1 that this estimate reflects. The production
level pricing,
basically a
cost based on silicon by the pound, seems to reflect
this same ratio as
do the
comparisons of the era when the 6502 was current.
Whats not relected is the associated gating and silicon busses. Those
eat
logic and realestate as they are often less regular.
What would be interesting, since it's conspicuously
absent from the list
quoted
below, is the transistor count in the 6802 (a 6800 with
internal clock
generator) and 6809.
Yes, those should be there.
Allison