Please see embedded comments below.
Dick
-----Original Message-----
From: Allison J Parent <allisonp_at_world.std.com>
To: Discussion re-collecting of classic computers
<classiccmp_at_u.washington.edu>
Date: Sunday, April 11, 1999 7:24 AM
Subject: Re: stepping machanism of Apple Disk ][ drive (was Re: Heatkit 51/4
floppies)
><I believe (guessing because I've learned memory doesn't serve as it once
><did) The load is two clock ticks and the indirect, indexed jump is five, s
><that's 3.5 microseconds, give or take a tick. it's less at 4 MHz, which
i
><what the 65C02C is rated, though it readily will run at 4.9152 (24.576
><MHz/5) over a wide temperature and voltage range provided the clock is
><phased correctly. the divice-by five yields a 40/60 h/l which must be
><inverted to give a little longer phase-2 than phase-1.
>
>The point was apparently missed. Of course I can take a cmos z80 and blow
>that out of the water using a 6 or 8 mhz clock. Heck using a 1989 version
>of the z80, the Z280 at 12.5mhz I can get the execution time way down. In
>the time frame before 1982 (as a marker) there werent any 4mhz 650c02s and
>there were 4mhz z80s and pdp-8s were still produced. In that context the
>the example represent programming style rather that absolute speed as they
>didn't vary that much over all to represent a great diffferece unless you
>needed a characteristic that was specific to a given CPU.
No, the point wasn't missed. I remember what was going on back then
(1979-1984) because it was at a critical juncture in the course of my life.
In 1979, my "favorite" CPU was the 6502, the fastest
MOS-Technology-comptible version of which was the 4 MHz NMOS part from
SYNERTEK. I was VERY involved in making things run faster than most folks
thought they could/should at that time and also had occasion to attempt
comparison and contrast on the basis of a number of parameters, including
performance. In '79, ZILOG and MOSTEK put out the 6-MHz "B"-series of their
Z-80, which was enjoying almost universal acceptance as the most widely
applicable and easiest-to-use microprocessor available. Most of the popular
statements about it were pretty much on the money. Of course, the evolution
of the 64K DRAM made its refresh counter more or less useless, but the
impact of that wasn't to be felt for a couple of years yet, as commercial
production of the 3-voltage 16K DRAMS was just getting into full swing.
>I'm not slamming the 6502 or it heirs as it's also a very popular embedded
>CPU still. For that fact so are the Z8 and Z80 heirs. Just from that it's
>possible to conclude they all had desirable enough characteristics to keep
>them in the running.
I agree with you there. The Zilog boys had the CP/M crowd to maintain the
low-end of their development system market, so nobody could complain it was
too expensive to develop. The MOS-Technology folks had merely to point at
the Apple to accomplish the same thing. Meanwhile, Motorola was making a
BIG mistake, abandoning the amateur and "small" users.
>As a CPU the 8051 is ok, I use it. As a controller it's without question
>a popular part still. But as a general purpose cpu, it's a really bad
>C or Pascal compiler host/target.
Since the evolution of the now-popular 'C' and PASCAL compilers for the
8051-core micro's, I believe the popularity of this 25-year-old model has
actually increased. The HLL's and the development of high-speed versions of
this processor family by DALLAS and Philips, among others have definitely
extended the life of this family. The simple migration path to "bigger"
parts of more or less the same architecture, e.g. '251, has also made many a
'51-core user. I believe that it's as a consequence of that, that there are
now compilers for several truly "ugly" architectures, e.g. the PIC/SCENIX
class of processors. There are also VHDL and VERILOG cores for several of
the older architectures, e.g. 650x, available for those who prefer to
"roll-their-own" which are also, though less well, supported with compilers
and other tools.
Again, Motorola seems to have been left behind at least with their smaller
MCU's. I guess that's because of their reputation for spurning applications
which consume fewer than 100K parts per week.
>
>Allison
>
Received on Sun Apr 11 1999 - 10:00:55 BST