super computer possibility SEA40, Haypress2 and beyond with ColorForth or Color(Less)Forth
peekay
Sent: Monday, November 02, 2009 7:53 PM
Subject: super computer possibility SEA40, Haypress2 and beyond with
ColorForth or Color(Less)Forth
here's the picture from that link :
suddenly jeff fox's speculation about many many many processors on
one single wafer of silicon for phenomenal parallel processing power
doesn't look like speculation any more !
sadly the forth community doesn't feel that it has a world changing
technology in hand and a linux like world changing opportunity !
even though forth's inventer himself affirms this all the time :-((
the progress emanating from silicon valley is based more on whom
you know rather than what you have .. and small changes rather
than radical changes
no wonder people make tons of money by peddling music playing
toys rather than serious devices which could have a nicer world
don't we need a proper system rather than what-we-have ?
(bernd pysan's .pdf on internet2.0 just touches the tip of the
iceberg ..
these and other far reaching changes are needed and WILL happen !)
John Passaniti
> suddenly jeff fox's speculation about many many many processors on
> one single wafer of silicon for phenomenal parallel processing power
> doesn't look like speculation any more !
Chips with multiple processors have been around for a few years now.
It didn't exactly require a visionary to think that continued
miniaturization would lead to yet more cores on a chip. So if you're
impressed by that prediction, here's one that will blow your mind: I
predict that in the future, there will be even *more* processors on a
chip. I know-- sounds crazy and I'm really sticking my neck out on
this one, but there, I wrote it.
> sadly the forth community doesn't feel that it has a world changing
> technology in hand and a linux like world changing opportunity !
> even though forth's inventer himself affirms this all the time :-((
That's odd. I see no shortage of people who are excited by systems
with tens or hundreds or thousands of processors on a single chip.
But that excitement is tempered by reality. Programming these kinds
of systems will require new tools, new techniques, and new ways of
thinking. Moving applications to such systems is hard. If you've
been following Charles Moore's "Haypress Creek Blog" you can see that
even the designer of such a system is finding challenges.
> the progress emanating from silicon valley is based more on whom
> you know rather than what you have .. and small changes rather
> than radical changes
Which is completely understandable. I don't know what you do
professionally, but my work involves creating the hardware and
software for actual, real-world, sits-on-a-shelf-and-is-sold
products. Radical solutions to problems are fine for skunkworks
operations and visionaries who can spend years pursuing their ideals.
Such efforts often later forms the basis for entire industries. Great
stuff-- but it doesn't work so well when there is a factory waiting
for your hardware and software.
In other words, sometimes the small incremental changes are what pay
for the radical changes.
Mux
about this for a while now..
The following is based on the assumption that GA's business model is
to go as wide as they possibly can. If they don't care about
evangalizing GA's and / or Forth and just get it in the hands of major
corporations, all this below is moot.
First and foremost. Without:
1. Affordable development hardware ( $99 and up )
2. Reasonable suite of development tools ( compiler, simulator,
debugger )
3. Community support
I feel GA's is doomed to succeed... One of the biggest hurdles I
personally faced was getting hold of a devboard. Even if you *could*
track one down, forking over $500 felt wrong. For the average
hobbyist, that's not an insignificant investment. Add to that that the
tools are somewhat 'odd' (i.e. Forth) and it'll be really hard for
somebody to want to check it out. An open-source solution, perhaps
based on Eclipse, dev board schematics, examples and support would go
a long way to entice people to at least give it a try. Good example of
this is XMOS, who seem to be targeting the microcontroller- and FPGA
sector.
Secondly, the fact that you can stick a whole bunch of chips on a
single chip doesn't mean it's a fantasic piece of kit. From a
hardware / spec performance point of view, sure.. But from a
programming point of view, it becomes somewhat of a nightmare. I think
even Chuck can attest to that.. People are lazy and, if history has
told us anything, gravitate to whatever is the easiest, main-stream
solution.
Last but not least, there's only so much a single person can crank
out. All the money in the world won't buy the output of thousands of
people trying to make The Next Cool Thing(tm). Want parallelism?
Here's your prime example :-)
As I stated earlier, if this isn't in GA's business plan, this all
doesn't matter.. I'd still like a dev-board though :-)
-Mux
Elizabeth D Rather
> I'll chip in here ( pun not intended ) as I've been wanting to talk
> about this for a while now..
>
> The following is based on the assumption that GA's business model is
> to go as wide as they possibly can. If they don't care about
> evangalizing GA's and / or Forth and just get it in the hands of major
> corporations, all this below is moot.
>
> First and foremost. Without:
>
> 1. Affordable development hardware ( $99 and up )
> 2. Reasonable suite of development tools ( compiler, simulator,
> debugger )
> 3. Community support
>
> I feel GA's is doomed to succeed... One of the biggest hurdles I
> personally faced was getting hold of a devboard. Even if you *could*
> track one down, forking over $500 felt wrong. For the average
> hobbyist, that's not an insignificant investment. Add to that that the
> tools are somewhat 'odd' (i.e. Forth) and it'll be really hard for
> somebody to want to check it out. An open-source solution, perhaps
> based on Eclipse, dev board schematics, examples and support would go
> a long way to entice people to at least give it a try. Good example of
> this is XMOS, who seem to be targeting the microcontroller- and FPGA
> sector.
Your list of 3 items is appropriate. Unfortunately, they can be
difficult and expensive to build. Intellasys was working on this, but
had a way to go before the curtain came down.
GA is operating on a shoestring, pending developing financing from one
or more sources. However, they do have a major asset in the form of a
team of engineers who now have several years' experience working with
the existing (ColorForth-based) toolset and hardware. It would make
good sense for a company looking to gain a quick advantage in their
market to make use of this *existing* expert team to actually develop
their products, and in my opinion attracting one or more such customers
is a sounder approach at this time.
If a succession of successful projects done on a custom basis attracts
sufficient funding to take the necessary steps to "go wide" they may
choose to do so. That would ultimately be up to the funding source.
Cheers,
Elizabeth
--
==================================================
Elizabeth D. Rather (US & Canada) 800-55-FORTH
FORTH Inc. +1 310.999.6784
5959 West Century Blvd. Suite 700
Los Angeles, CA 90045
http://www.forth.com
"Forth-based products and Services for real-time
applications since 1973."
==================================================
Mux
without the financial resources to get a fully development system up
and running. I also agree that leveraging existing expertise is one of
the fastest ways to get things off the ground. It's a bit of a chicken
& egg situation unfortunately.
That said, putting out a GA32 or even GA4 with some very rudimentary
support hardware / software ( i.e. Haypress-lite ) should entice many
a hobbyist to experiment and perhaps contribute to the company's
success.
Just a thought..
-Mux
> Los Angeles, CA 90045http://www.forth.com
>
> "Forth-based products and Services for real-time
> applications since 1973."
> ==================================================- Hide quoted text -
>
> - Show quoted text -
peekay
>
> "Forth-based products and Services for real-time
> applications since 1973."
> ==================================================
"It would make
good sense for a company looking to gain a quick advantage in their
market to make use of this *existing* expert team to actually develop
their products"
what 'products' can be made from the SEA40 or Haypress2 ?
especially products with mass appeal/use ? (like ipod ? ugh ! etc)
actually, it is better if the *existing* expert team suggests
what use the processors can be put to
for a venture fund or financer or maker to come up with a need
is a not-happening thing !
Brad
> Your list of 3 items is appropriate. Unfortunately, they can be
> difficult and expensive to build. Intellasys was working on this, but
> had a way to go before the curtain came down.
>
can they just buy a few chips? In fact, if a hobby user is going to
the effort of making a little 2-layer PCB to test-drive the part, it
would make sense to work out a trade.
And it's not as if the parts are hard to solder. The GA32 sounds like
a QFN package, which is easy to solder by hand with hot air and solder
paste. Any excess solder blobs up around the outer edge of the package
and cleans up easily by running a soldering iron tip around the
package.
-Brad
peekay
> To: Color...@googlegroups.com
> Sent: Monday, November 02, 2009 7:53 PM
> Subject: super computer possibility SEA40, Haypress2 and beyond with
> ColorForth or Color(Less)Forth
>
> technology in hand and a linux like world changing opportunity !
> even though forth's inventer himself affirms this all the time :-((
more on this -
chuck's comments on
http://www.simple-talk.com/content/article.aspx?article=775
"All I can do is provide existence proofs:
Forth is a simple language;
OKAD is a simple design tool;
GreenArrays offers simple computer chips.
No one is paying any attention."
Elizabeth D Rather
Yes, they do have a list of possible product areas they've researched
that they're presenting, but the ideal investment arrangement would be
with a partner who has an appropriate technology need (with the
application domain expertise) *and* money, rather than a traditional
venture capitalist who has *only* money and expects a partner to provide
all the expertise.
peekay
ideal is the right word .. yet, i wonder :-(
i am doubt that such an ideal partner can be easily found
so let's dig into history and try to pinpoint some such
partnerships
this would probably be similarly repeated (OR similarly NOT happen)
Wayne
wrote:
> On Nov 5, 12:01 am, Elizabeth D Rather <erat...@forth.com> wrote:
>> peekay wrote:
>> > On Nov 3, 5:24 am, Elizabeth D Rather <erat...@forth.com> wrote:
>> >> Mux wrote:
>> Cheers,
>> Elizabeth
> "the ideal investment arrangement would be
> with a partner who has an appropriate technology need (with the
> application domain expertise) *and* money"
>
> ideal is the right word .. yet, i wonder :-(
>
> i am doubt that such an ideal partner can be easily found
>
> so let's dig into history and try to pinpoint some such
> partnerships
>
> this would probably be similarly repeated (OR similarly NOT happen)
Ideal, versus venture capitalist that basically want to buy the company
for next to nothing, but still get you to do all the work. It will
continue to be this way as long as people regard ideas as worthless. In
Japan, employees even sue companies if they fail to give them public
recognition for things they come up with.
An idea extends to the implementation, before that is more of an notion
than an implementation (solution).
Wayne, again.
