The point I was trying to establish was that it's difficult to develop a basis
for comparison that treats both processors fairly. The Z80 has cycles of
different lengths that make it difficult to find a basis for comparing the two
based on similar memory bandwidth, while the 6502 tends to be used with half its
bus-time idle. That's not efficient use of the most costly element in the
system (based on 1980 parameters). If you compare oscillator frequency, the
6502 has a great advantage, and if you base the comparison on no-wait-state
memory bandwidth requirements, it has an unfair advantage again. My contention,
of course, is that even if you do level the playing field, the Z80 is still
slower overall, but especially at 8-bit microcontroller things, than the 6502.
What remains to be seen is whether the Z80 can muster any sort of win over the
6502 in terms of code that takes advantage of its rich internal register
resources.
One can't proceed with that, however, until a fair basis for comparison is
established.
Dick
----- Original Message -----
From: "Sipke de Wal" <sipke_at_wxs.nl>
To: <classiccmp_at_classiccmp.org>
Sent: Thursday, December 20, 2001 2:21 PM
Subject: Re: 6502/Z80 speed comparison (was MITS 2SIO serial chip?)
> Again clockfrequencies and cycletimings are not a good way to
> establish computingpower. You should look at the performance of
> both processors at their maximum memory-throughput or bus-bandwidth
> with the memory-technology of those days.
>
> Although I like the Z80 and worked with it a lot, I still think that the
> 6502 was more powerfull from an userapplication perspective, all
> other things being equal of course ..........
>
> Anyway, both CPU's served their purpose and their downfall was
> probably more related to their limited 16-bit performance and the
> limit of the amount of directly accessable memory than their
> respective 8-bit performance. After all, if they would have evolved
> with the advance in technology they could both have achieved
> more glorious days.
>
> Remember that the M68000 family was designwise far superior
> to the 8086/8088. But unfortunatly, we all know that (more often than not)
> design-qualilty has no real relation to the succes of a CPU-family...........
>
> Sipke de Wal
> ----------------------------------------------------------------
> http://xgistor.ath.cx
> ----------------------------------------------------------------
>
>
>
> ----- Original Message -----
> From: Bryan Pope <bpope_at_wordstock.com>
> To: <classiccmp_at_classiccmp.org>
> Sent: Thursday, December 20, 2001 7:31 PM
> Subject: Re: 6502/Z80 speed comparison (was MITS 2SIO serial chip?)
>
>
> > >
> > > Also, if you realy want 'ultimate speed' you have to look at actual
> Versions.
> > > close Z80 relatives (like 180) are available up to 32 MHz (I have seen
only
> > > 20s, but well) while actual versions of the classic 6502 and 65816 is
able
> > > to do 14 MHz. Now take in account that a clock ratio of 1:3 to 1:4 for
6502
> >
> > The 65816 can go up to 20Mhz. It is used in the SuperCPU for the 64 and
> > 128.
> >
> > > vs. Z80 (*) is a good rule of thump, so a 14 MHz 6502 equals a 42 to 56
MHz
> > > Z80. Quite a racer ...
> >
> > Cheers,
> >
> > Bryan Pope
>
>
Received on Fri Dec 21 2001 - 00:28:53 GMT