home .. forth .. misc mail list archive ..

Re:chips


Dear Miscers

Jeff Fox Wrote:
>I picked up "More on Forth Engines" Volume 20 from Dr. Ting last weekend.
>This issue includes a description of the 44 pin MuP21h PLCC, a chapter
>on Tom Zimmer and Andrew McKewan's F95 (Win32For), new OK and rectangle
>drawing code for MuP21, and Chuck's patents on Forth, Sh-Boom, and MuP21.

Interesting, finally Chuck has got around to patenting the Sh-Boom, and
Mup21, this should be a great legal future, though Forth is to public
domain, Idon't think such a patent would go through in Australia, even if we
had software patentting (gratefully). Why should we have the more general
form of protection (rather than you can't copy my work, or protend to be my
product (look like etc), obvously the value of a language is in it's words
and interface, and thus there should be rulings on anything that is 100% (or
largerly) compatible with it.  In the end I would argue that it offers More
than Reasonable protection to those who can afford it.  For those who can't
afford it (smaller startups) it makes life even harder, and constricts what
they can do, but of course I could be wrong here, is a large sum required
like normal patents or is it cheap like copyright, does the new proposed
software patent system replace copyright etc, this is Australia over here
and I have been to busy to follow it?  Anyway I'll have to get this issue,
and have a look at that new PLCC.


>Also Mosis delivered P8 and F21 a couple of weeks before expected on 
>Friday.  Chips are being tested now and results will be posted.


Hooray, I'll look forward to finding out about these.


Now the more important stuff.

I have come to the incredible realisation that my completion date for my own
project that I want to put on PC's 21's and 32's, is slipping (again), and
it now looks like I won't be able to complete it before the end of the Year,
so hopefully I'll be able to complete it before the end of next year, before
the 10th anniversery of the start of the project.  I was going to keep quite
until completion, but now so many chips are near completion I want to raise a
few application possibilities.

The current market for micro-controllers, so I have read, is moving toward
PC compatability (maybe a bit of a waste), allready we have embeded, 8086,
286 and 386, they (people in industry) are likely to adopt this despite what
we may do (they will probably trust the compatability for business/project
security, and cheap hardware).  I propose that a usefull project for some
people with the PC know how would be for them to implement a chip on a plug
in board that fitted the socket of 8086 (or other), and program it to
emulate (through rom/ram), the 8086 instruction set.  This is not impossible
and recent advances in emulation technology, such as emulating PC's on Mac's
etc, have brought them up to a reasonable speed (not that many of these
techniques are usefull for this).  Once the technology for this has been
demonstrated it can be implemented on chip level for different processors,
allowing direct drop in replacements for variouse controller designs.  What
advantages does this offer:

- With a programmable clock rate, a faster, or a
lower powered compatible cpu.

- A software selectable native mode (Misc),
that offers the designer 80mips+ performance aswell.

- Pc compatability

- option of having connectors running off the top of chip or etc for the
extra functionality of these processors (video, sound etc), for custom
designs.

A Big project I don't have the expertise to do myself, or the time (must try
to finish this other project first), but for somebody with the expertise, a
way to get the Mup & F designs in many PC's, Micro-contollers and consumer
applications, what do you think?

Even though there is a problem with implementing the Rom/Ram for the above
in Chucks present design, multiple chip modules (so I have notice) have been
taking on quiet a spurt of growth in the press (Electronic Design, EDN,
Information Week (I think), etc) recently (due to cost
reductions).  It might be possible to put the processor on a die of an
existing design of support chip or alternately, stack the chips on top of
one another or put them on a small surface mount module, to archeive the
above.  I beleive that a reason previously for using this has been costs,
I'm just saying cost are appearring to come down.

Indeed it might be possible, in future, to implement a Super Computer in a
multiple stacked module (I saw one of these modules in a press release
recently, had at least 4 chip stacks on it). Just Imagine 10 cpu's, Memory,
support chips, graphics accelerator etc on one module (this should be
possible with the Right technology), maybe fantasy now, but in future?  At
the moment single chip stacks are capable of upto 128Mb+ of flash (I think
it might be a Giga Byte now).

One last thing, at the moment PCI buss is currently a popular trend, if it
will remain so (look at it this way, it's speed, and data transfer rate is
being greatly outstripped by new technology, and there are alternatives in
the pipeline, as was the case with 8-bit ISA) then it would probably pay to
make a PCI bridge chip and software to the Mup32 (when complete), it should
be able to use any PCI cards then in it's designs.  I understand, that the
PC-Card (PCMCIA 32-Bit), uses PCI as the basis for the 32-bit extensions,
though I could be wrong again and there is a planned embedded (small form
factor version).  PCI offers a large extent of processor independent
compatability for cards, and uses a forth type code to help archeive this. 

Out of the three suggestions here, PCI offers the most pratical help for the
processors, while the plug in replacement for micro-controllers might be the
most lucative.

Only a few suggestions for a possible future for the processors. There are 
others but they might have to wait until my project is ready.

By the way, apart from evaluation boards, what about a ISA based plug in
Card with the Chips on it?

Look after yourselves Miscers.

Bye.

Wayne Morellini
waynem1@cq-pan.cqu.edu.au 
--