<gets on soapbox>
It's true that the '186 and '188 were VERY widely used. The 68000 was really
not nearly so popular, but it looks as though they were similarly popular in
the PC world. That market, in the '80's represents a completely insignificant
share of the market, however, since many products using processors of the
'186/88 class were made in quantities each week equalling the entire PC
market. Intel may, if its dominance on the PC market continues, eventually
sell as many pentium chips as they'd sold '186's by the end of 1989, but only
if they stop selling '186's right now, which they won't. The '186 was so
significant that the Soviets even counterfeited the thing for use in their
cruise missiles, among other things.
The 68000 was a disappointment on the market, not living up to its promise, or
even its claims, badly marketed, and too costly for the applications that it
did suit. The 68010 was an improvement in some respects, but its main
performance hit came when MOT made the decision to punt high performance, as
promised by the presence of that "DTACK" signal that ended a bus cycle, in
order to make applications engineering easier. The result was that this
processor, which could have run quite fast if outfitted with a small amount of
cache, had a shortest-cycle of about 500 ns, which was as slow as nearly any
other CPU. Further, because of its long opcodes, it required two strokes into
memory for an opcode, at least two for the effective address, and two for the
data. There were other CPU's at the time that would easily outrun it even
though they were only 8-bit types, in many tasks. Because of its architecture
and its inability to run fast, on short addresses and data, it ran 8-bit tasks
very badly, and Intel's processors looked pretty good by comparison. In fact,
the 68K was inherently a better-designed architecture, but the abandonment of
the DTACK-enabled fast bus cycle which the early masks could use, hurt them in
the market. The 68020 finally seemed to hit the mark, but was quite costly to
use. In a fair test of computing power, the 68K would outperform the '186,
but most tasks for microprocessors aren't really computing tasks, they're
simple signal processing tasks, and not in the sense of "Digital Signal
Processing" which is VERY numerically oriented, but in the sense of simple
ANDs and ORs, and the like.
Had MOT gotten behind the CP/M 68K costing a few hundred dollars, rather than
the then (1980) very costly UNIX types, not to mention their own rather weak
operating systems, they'd have had a tiger by the tail, because the computer
market was ready. MOT wasn't, however. They thought a 6809 was a better
choice for home computing.
<falls off soapbox>
Dick
----- Original Message -----
From: "Peter C. Wallace" <pcw_at_mesanet.com>
To: <classiccmp_at_classiccmp.org>
Sent: Tuesday, January 15, 2002 5:33 PM
Subject: Re: 68010 & 80186 (was: Re: Mac IIci)
> On Tue, 15 Jan 2002, Ken Seefried wrote:
>
> > >Yes, and the 80186 -- 68010.
> > >Both existed but were not popular in many systems. Both equally
> > >quite rare in that regard.
> >
> > Not entirely true.
> >
> > Clearly the 68010 was quickly eclipsed by the 68020 and thus showed up in
> > relatively few systems. This is different, BTW, than not being popular.
> >
> > OTOH, the 80186 (including the AMD Am186 line) was a wildly successful
chip
> > in the embedded systems world. Vast numbers of devices based on the '186
> > are out there, often because one could use the IBM PC and follow-ons,
> > hardware & software, as development platforms on the (relative) cheap.
> >
> > Ken
> >
>
> And dont forget that the MAD 1 Computer used a '186...
>
> (saw one at Comdex '84 and thought it looked cool)
>
> Peter Wallace
>
>
Received on Tue Jan 15 2002 - 19:22:29 GMT