Re: micro processor vs controller

[Andrew Sieber <asieber@xxxxxxx>]

Jeff Fox wrote:
> Regularly timed interrupts can be provided, but they won't come from
> the RTC circuit.
For a periodic interrupt such as is necessary for preemptive
multitasking, and is available from a watchdog timer, one can simply
connect an external timer to the interrupt pin on the processor.  This
of course requires extra external components, but on the F21 at least it
can indeed be done, whereas on the MuP21 there was no way (that I know
of) to get a periodic interrupt.  Which means that only cooperative
multitasking can be done on MuP21.

> F21 is not really a microcontroller because it does not have any on chip
> RAM and 32 words of ROM is not enough for a complete application.
I read on iTV's web site (during the brief period that it was available
a couple months ago, and has since been taken down for some reason) that
the i21 processor will have a full megabyte of on-chip flash ROM to hold
the entire software repertoire for the Pegasus box, and it will be
fabricated at the same size (.8 micron) and speed as the F21.  Rather
impressive.  I hope that was not a typo on the web page.  Thus it seems
that much more memory could be fit on F21 than currently exists.  I am
assuming that the die size for F21 and i21 are the same.  And as far as
F21 being a microcontroller, that should be doable, considering what the
current standard for microcontrollers is.  Microchip makes little 8-pin
chips with 1kB of on-chip eprom and 48 bytes of on-chip RAM or something
pitiful like that, and if that's all it takes to qualify as a
microcontroller, then F21 could qualify.  Even if it takes ten times as
much space for a bit of RAM than a bit of flash ROM, far more RAM and
ROM could be fit onto F21 than exists on Microchip's microcontrollers. 
F21 could then be operated as a complete, stand-alone chip just as
Microchip's chips, except 100 times as fast.

> When I look at a single chip solution using a microcontroller I don't
> see much of a match to F21.  With F21 you need external memory.  You
> could do some things with only a single 8 bit ROM, but you could not
> make use of the I/O coprocessors that use 20 bit memory.
With only minimal on-chip RAM and flash ROM, F21 can be a standalone
microcontroller that competes with stuff like Microchip sells.  Of
course, without attaching external DRAM, most of F21's coprocessors
would be pretty useless (especially the video coprocessor), but it could
still function as a microcontroller that does pretty much everything a
Micochip chip will do.

> The upper limit is quite high,
> but so may be the price (it is hard to beat a $3000 microprocessor on raw
> computation).
Since those $3000 microprocessors (and I assume you are referring to the
likes of Alpha, which I think is now up around 500MHz) are fabricated on
.25 micron processes, and F21 is on .8 micron, it shouldn't be hard to
beat top-notch processors on raw computation.  I think Chuck has
predicted beyond 1GHz for .5 micron chips, so _if_ he had access to .25
micron fabrication facilities, I have little doubt he could produce
literally the fastest processors on the planet.
I have read that you sometimes have trouble getting investors to
continue funding iTV Corporation.  I find that rather curious.  Chuck
has already demonstrated his ability to make chips far faster and better
than other companies.  He simply needs some time and money for fab
runs.  To me, iTV sounds like an investor's dream come true.  Just pump
in a bunch of money for a little while and soon you have not only the
world's fastest processor, but one that costs almost a couple orders of
magnitude less than the competition, leaving folks like Intel in the
dust.  MuP21 already works, and it sounds like i21 and F21 will also
soon work, and there's no reason to think that .25 micron could not
yield chips that ran at over 2GHz.  Perhaps I'm being really naive, but
iTV sounds like a very low-risk investment opportunity with the
potential for huge payoffs within just a few years at most.  I would
have expected investors to flock to it.

> Bill Gates thinks that your toaster should be running windows.  While
> a Pentium may be able to actually cook the toast for you a PC running
> Windows is going to be a pretty expensive toaster controller.  F21
> was not designed for your average toaster, but it would be a better
> choice than a Pentium PC running MS Windows. :-)
I don't know what you're talking about; there is no way an MuP21 would
get hot enough to cook your toast.  I have a feeling a F21 wouldn't,
either.  Not unless you wired up the 5V backwards.
I usually agree with you, but this time I really must take Bill's side. 
I think a Pentium PC running MS Windows is an excellent idea for cooking
toast.  Cooking toast is about all a Pentium PC running MS Windows is
good for anyway.  I challenge anybody to disagree.  Anybody who notices
the header to this message that betrays the fact that I am running
Windows 95 right now, please ignore it.  I freely admit my hypocricy.

Just in case anybody cares, I have designed from scratch a double sided
PC board to hold an MuP21, a 72-pin SIMM, boot EPROM, some support
chips, etc.  I printed out the design and photolithographed (is that a
valid word?) it to a circuit board, drilled the board, and tried to
solder some sockets into the board.  That's when I noticed that, like an
idiot, I had designed the board with the chips too close together and I
couldn't physically fit the sockets into the board next to each other
where they needed to go because they were too big.  Don't laugh.  These
things happen.  By sort-of melting the sides of the sockets away with
the soldering iron (vapor plastic really stinks) and bending the pins
quite a bit, I can sort-of get the sockets to go in.  I also have other
problems, like accidental solder bridges that I can't eliminate between
the sockets and the circuit board (I had to solder the top as well as
the bottom because I can't do plated-through holes).  I shouldn't have
tried this at home.  Hmmm...  I spy a wire-wrap tool...

--Andrew Sieber
asieber@usa.net