If you run both the Z80 and the 6502 at the same clock rate, the 6502 will be
quite a bit faster out of the chute and leave the z80 in the dust. There's long
been argument as to which CPU is faster at a nominally 4:1 Z80/6502 clock ratio.
There's never even been a debate over which CPU is faster at the same clock
rate. I believe there's no comparison at all in that case. It wouldn't matter
what percentage of16-bit operations you use as a comparison basis if that's your
basis. The 6502 will always outperform the Z80 if they both run at the same
clock rate, simply because of the difference in how the cycles are constructed
around the clock.
If you think you can come up with a procedure to serve as a basis for comparison
in which a 4 MHz Z80 will outperform a 4 MHz 6502, wait-states or no, then
please present it here for us to evaluate. That would be VERY interesting!
Dick
----- Original Message -----
From: "Allison" <ajp166(a)bellatlantic.net>
To: <classiccmp(a)classiccmp.org>
Sent: Thursday, December 20, 2001 6:58 AM
Subject: Re: MITS 2SIO serial chip?
None of that means much. At the same sped which one
will do
a long list of benchmarks faster is the only one that counts in
the speed derby and which one is cheapest to implement for
the specific case is likely number one.
Allison
-----Original Message-----
From: Richard Erlacher <edick(a)idcomm.com>
To: classiccmp(a)classiccmp.org <classiccmp(a)classiccmp.org>
Date: Thursday, December 20, 2001 4:21 AM
Subject: Re: MITS 2SIO serial chip?
If you run a 20 MHz Z80 against a 20 MHz 6502,
you'll find the 6502
performs WAY
(3x-5x) faster than the Z80. It's difficult
to base a comparison on clock
rate
alone. All 6502 memory cycles take 1 clock tick.
Z80 memory cycles, aside
from
M1, take 3 clock ticks, with no wait states in
use. If you run the 6502 at
a
rate that fully utilizes the memory bandwidth at a
rate such that the Z80
can
perform an M1 in the same amount of time, without
wait states, the 6502
will
always be faster, because it's running faster.
The Z80 uses 1-1/2 clock
ticks
to execute its instruction fetch. If that's
to be memory access window for
each
processor, and you run them both from static
memory, and you allow minimal
recovery time, e.g. use 10-15 ns memory, (just for the comparison) then you
can
use a 20 MHz Z80 and a 20 (actually 14) MHz 6502,
and clock the 6502 with a
25ns
low, 75ns high clock, and drive the Z80 with a
square 20 MHz. That will be
a 10
MHz clock for the 6502 and a 20 MHz clock for the
Z80. I'd submit that the
6502
will still run rings around the Z80, since it is
still going to be cycling
memory at an average of 200 ns per cycle, while the 6502 does it a 100 ns
rate.
Surely a better comparison can be arranged.
Dick
----- Original Message -----
From: "ajp166" <ajp166(a)bellatlantic.net>
To: <classiccmp(a)classiccmp.org>
Sent: Wednesday, December 19, 2001 4:46 PM
Subject: Re: MITS 2SIO serial chip?
Ben
Franchuk <bfranchuk(a)jetnet.ab.ca>ca>:
> what is the faster CPU -- A 6502 or Z80 style processor like
> the rabbit.
Depends...
For the same instruction execution rate (ignores clocks and cycles)
the z80 is likely faster. However... if you have a 20mhz 6502 and
a 10mhz z80 it gets muddier with the 6502 being the faster. And
if you know one better than the other you can certainly exploit it
all the more. In the end it's not which one does a task faster, it's
what one you can code the task for faster.
Sorta like asking apples or oranges.
Allison