GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a PC
peekay
googlegroups :
Are these processors suitable for real-time video encoding/decoding
and mobile application like phones / netbook?
question was in response to this post :
--- SVFIG Meeting July 25, 2009 ---
AGENDA
10:00 --- COFFEE AND A CHAT
[SNIP]
14:23 --- A New Generation of Forth Arrays and Tools --- Jeff Fox
Jeff will review the details of the Green Arrays GA4, GA32, and
GA144 and give an example of software using the hardware changes
that Chuck has listed at his site. He will show an interactive Forth
on
GA32 talking to ColorForth on a PC
peekay
> Are these processors suitable for real-time video encoding/decoding
> and mobile application like phones / netbook?
Because the amount of RAM on-chip is so limited, and because
it is distributed these chips are not best for problems that
require random access to large data sets like multiple
video frames. GA4 has only 256 words of RAM on-chip and
would not be suitable for common video compression
or decompression. IntellaSys did demonstrate H264 video
format decoding and it can be done with some of a GA144
but it was also demonstrated that there was not sufficient
memory bandwidth for doing the higher video resolutions.
Phones have been discussed here before. Much external hardware
would be needed as the 100MHz A/D and D/A are not sufficient
for multi-gigahertz cell frequencies. A later generation chip
with faster analog technology and more memory might be
sufficient for a one-chip cell phone but that's not going
to happen any time soon.
Netbook is another application that needs high speed wireless
network access and thus external hardware. It could not
be very compatible with netbooks that are more closely
modeled on PC technology. I think 'netbook' tends to
imply a certain amount of PCness. Chuck is generating
1024x768 video and running parts of the OKAD VLSI suite
on the Haypress Creek board at the present time which
means it is a sort of colorforth PC for cad. Driving an
LCD is easy but unless you own the company that makes the
display there is no profit margin in a device that would
use it, it is a hard market to enter.
> Jeff will review the details of the Green Arrays GA4, GA32, and
> GA144 and give an example of software using the hardware changes
> that Chuck has listed at his site. He will show an interactive Forth
I reviewed information that Chuck has published on his
website about the chips and changes to the designs. I
gave some examples of new ROM code and explained how
the changes that Chuck lists reduce the complexity
of the packet protocol boot code and make it smaller,
faster, and simpler than it was previously.
The Green Arrays web site lists the chips. GA4 is a
2x2 array, GA32 is an 8x4 array, GA40 a 10x4 array
and GA144 and 18x8 array of cores. As I had said
before people should think family of chips. All are
based on the multiple copies of Forth core running at a
maxium of 700 Forth MIPs and at better than 10uw
per Forth MIP. GA4 were put into a couple of different
packages and in a bag look a lot like smart dust.
The main thing covered was how pin wake-up circuits
have changed and now put something useful, not garbage,
on the bus when a pin triggered wake-up event happens
as it makes the ports executable. External event
response is still a couple hundred picoseconds.
Making the pin wake-up circuit work more like
neighbor wake-up circuits simplifies software
by making pin wake-up ports executable and thus
more like other ports. This is pretty
obvious to anyone who has written much code
for this architecture because you had to
deal with pin ports and neighbor ports with
different code.
I reviewed the differences between packet boot
from on-chip or off-chip neighbor connections,
and the ROM boot routes including the SPI flash
code, the asynchronous autobaud code, and the
synchronous serial protocol code. All had been
reviewed before but not the new pulse width
modulated serial protocol using the new
executable pin port hardware. The details
were covered and various code examples were
shown to demonstrate how changes allow for
dramatic reduction in program size and timing.
I also reviewed the history of how some code
has been reduced over the years from a
paragraph, to a line, to a single word of
ROM, to in some cases no code. When attempting
to make the technology as small, simple, low
power and fast as possible this has been the
history. In one example I showed how much code
was needed on previous models of the design
compared to the latest and compared to what
a PC processor needs to do the same thing
but more slowly and with a lot more code.
Bill Muench and John Rible were present and
also talked a little about recent changes.
Bill talked a little about the eForth
port to GA32. It was a small meeting but
there were people who were interested. I
mentioned some work going on at IntellaSys,
like things that distinguish their hearing
recovery application, and the organizational
and technical approach differences between
IntellaSys a TPL Group, and Green Arrays.
I showed how the Interactive Development
Environment works in colorforth and how it
requires a minimal amount of code on most
target nodes. And as Chuck likes to say the
minimum is zero. I showed how it uses
an edited block with some blue words as a
template to simplify the design.
I also covered the use of blue words.
Chuck is very happy with the new color.
Blue words denote execution of a Forth
word at block list or edit time. I
recounted the details of the video I made
of Wil Baden's presentation about what he
had done to integrate the execution of
Forth words in his foreign file editor as
it was a rather involved process while
what Chuck has done is very simple
partly due to everything being in Forth
and partly by design.
I also brought in about 50 DVD and CD
with videos of old presentations at SVFIG
some of which had been available on my
website in the past. I let people make
copies. There were some unpublished
ones that I made available too. I did
not shoot a video of the presentation
or make a transcript thought some
video was shot. I tried to answer the
questions people had.
Best Wishes
peekay
> 14:23 --- A New Generation of Forth Arrays and Tools --- Jeff Fox
> Jeff will review the details of the Green Arrays GA4, GA32, and
> GA144 and give an example of software using the hardware changes
> that Chuck has listed at his site. He will show an interactive Forth
peekay
---------- Forwarded message ----------
From: Howerd <howe...@yahoo.co.uk>
Date: Aug 9, 2:58 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
Hi Jeff,
As always, thanks for the info...
> I also covered the use of blue words.
> Chuck is very happy with the new color.
> Blue words denote execution of a Forth
> word at block list or edit time.
I like this concept.
Are there any plans to publish the code for blue tokens?
Best Regards
Howerd
On 8 Aug, 18:09, Jeff Fox <f...@ultratechnology.com> wrote:
> On Aug 5, 11:27 am, peekay <pksharmakolk...@gmail.com> wrote:
> > Are these processors suitable for real-time video encoding/decoding
> > and mobile application like phones / netbook?
> Because the amount of RAM on-chip is so limited, and because
> > 14:23 --- A New Generation of Forth Arrays and Tools --- Jeff Fox
> > Jeff will review the details of the Green Arrays GA4, GA32, and
> > GA144 and give an example of software using the hardware changes
> > that Chuck has listed at his site. He will show an interactive Forth
peekay
---------- Forwarded message ----------
From: Dennis Ruffer <druf...@worldnet.att.net>
Date: Aug 9, 5:18 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Aug 8, 12:09 pm, Jeff Fox <f...@ultratechnology.com> wrote:
> I showed how the Interactive Development
> Environment works in colorforth and how it
> requires a minimal amount of code on most
> target nodes. And as Chuck likes to say the
> minimum is zero. I showed how it uses
> an edited block with some blue words as a
> template to simplify the design.
Is the colorforth IDE going to be released?
DaR
peekay
---------- Forwarded message ----------
From: Jeff Fox <f...@ultratechnology.com>
Date: Aug 9, 8:21 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Aug 9, 2:58 am, Howerd <howe...@yahoo.co.uk> wrote:
> I like this concept.
> Are there any plans to publish the code for blue tokens?
Blue words were present in the last public release but
were not integrated into the editor control panel with a
key to select blue and the word to toggle blue word
display in the editor, seeb. In addition to adding that
to the editor extensions compiled after boot I understand
that there have been other changes such as making it
safe to have an undefined blue word on block rather
than have it crash the system then the block is edited.
The editor command extension is needed in the same
form as the one to cycle the word under the cursor
through yellow/green/white with the color cycle key.
One adds the key to the command table for the editor
and adds the display to the menu displayed in the
editor. I see no reason why the new block 68
can't be released but I haven't tried all of it
with the year and half old release.
I don't recall exactly what was in that release and
would have to go back and see if the current block 68
will work as it is with that release.
Best Wishes
peekay
---------- Forwarded message ----------
From: Jeff Fox <f...@ultratechnology.com>
Date: Aug 10, 4:48 am
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Aug 9, 5:18 am, Dennis Ruffer <druf...@worldnet.att.net> wrote:
> On Aug 8, 12:09 pm, Jeff Fox <f...@ultratechnology.com> wrote:
> Is the colorforth IDE going to be released?
Probably. But I don't know when.
Best Wishes
peekay
---------- Forwarded message ----------
From: Dennis Ruffer <druf...@worldnet.att.net>
Date: Aug 10, 5:25 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Aug 9, 6:48 pm, Jeff Fox <f...@ultratechnology.com> wrote:
> On Aug 9, 5:18 am, Dennis Ruffer <druf...@worldnet.att.net> wrote:
> > On Aug 8, 12:09 pm, Jeff Fox <f...@ultratechnology.com> wrote:
> > Is the colorforth IDE going to be released?
> Probably. But I don't know when.
> Best Wishes
I hope you can convey to Chuck and his crew that this would be
desirable. I realize that the community is pretty small right now,
and I'm not sure what any of us can do to help. However, I do know
that there is at least one (me) who would like to see what the
colorforth interface can do.
Thanks for the update.
DaR
peekay
---------- Forwarded message ----------
From: Howerd <howe...@yahoo.co.uk>
Date: Aug 10, 6:28 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
Hi Jeff,
Maybe I'm missing something here, but I haven't seen any editor source
code in any release (apart from my own incomplete version in
CFDOS4.blk).
Block 68 is labled "USB boot" in the colorForth 2a 2008 release...
I think your comments would be easier for me to understand when
combined with the source code and a working program ;)
I am reluctant to finish my own colorForth editor, knowing on the one
hand that it is already out of date (no blue tokens), and on the other
hand that fully working code already exists - it offends my sense of
efficiency!
All of that notwithstanding, your comments are like precious drops of
rain in the arid desert of sourceless colorForth...
Best regards
Howerd
peekay
---------- Forwarded message ----------
From: Howerd <howe...@yahoo.co.uk>
Date: Aug 10, 6:30 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
Hi Dennis and All,
> However, I do know that there is at least one (me)
> who would like to see what the colorforth interface can do.
Make that two of us!
Best regards,
Howerd
peekay
---------- Forwarded message ----------
From: Howerd <howe...@yahoo.co.uk>
Date: Aug 10, 6:30 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
peekay
---------- Forwarded message ----------
From: Pavel Klinkovsky <pavel.klinkov...@gmail.com>
Date: Aug 10, 11:07 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On 10 srp, 15:30, Howerd <howe...@yahoo.co.uk> wrote:
> Hi Dennis and All,
> > However, I do know that there is at least one (me)
> > who would like to see what the colorforth interface can do.
> Make that two of us!
> Best regards,
> Howerd
> On 10 Aug, 13:25, Dennis Ruffer <druf...@worldnet.att.net> wrote:
> > On Aug 9, 6:48 pm, Jeff Fox <f...@ultratechnology.com> wrote:
> > > On Aug 9, 5:18 am, Dennis Ruffer <druf...@worldnet.att.net> wrote:
> > > > On Aug 8, 12:09 pm, Jeff Fox <f...@ultratechnology.com> wrote:
> > > > Is the colorforth IDE going to be released?
> > > Probably. But I don't know when.
> > > Best Wishes
> > I hope you can convey to Chuck and his crew that this would be
> > desirable. I realize that the community is pretty small right now,
> > and I'm not sure what any of us can do to help. However, I do know
> > that there is at least one (me) who would like to see what the
> > colorforth interface can do.
> > Thanks for the update.
> > DaR
Can you accept me to your group of interesed? ;-)
Pavel
peekay
---------- Forwarded message ----------
From: Jeff Fox <f...@ultratechnology.com>
Date: Aug 11, 12:50 am
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Aug 10, 6:28 am, Howerd <howe...@yahoo.co.uk> wrote:
> Hi Jeff,
> Maybe I'm missing something here, but I haven't seen any editor source
> code in any release (apart from my own incomplete version in
> CFDOS4.blk).
> Block 68 is labled "USB boot" in the colorForth 2a 2008 release...
> I think your comments would be easier for me to understand when
> combined with the source code and a working program ;)
Block 92 is labeled "word search." The f command is added to
the extended key table, ekt there. You type "find" on the command
line and the word to find and you jump to the next occurance of the
word in the source with the "f" command. "f" and find are editor
extensions and find a word in any color in the source. It was the
first published editor extension I think. "def" finds a red
word with a name, the definition of the word.
Block 94 is labeled "editor additions recolor" and is where the
color cycle key is defined, inserted in the key table and put
on the editor menu. It cycles the color of green/yellow/white
words under the cursor.
Block 98 is labeled "blue words." It defines the blue function,
but does not add to the key table, or put "b" on the menu like
the systems that are newer than a year and half old.
seeb can be put on the boot screen since 90 92 94 and 98 are
compiled at boot. However there are no blue words to see and
98 did not put the blue function into the key table or on
the menu so it isn't quite setup to add blue words by just
selecting blue and typing in the editor. seeb has to be
invoked anyway so you can see blue words on blocks.
I never had to make any changes to the editor myself. I
just turn off qwerty on the boot screen and use the system.
Versions produced internally since then add b to the key table
and menu just below the d key under the ring finger of the left hand.
So I never had to do that myself. But the code has been then
and labeled 'blue words' for over a year.
Capitals were changed and other changes were made to the editor
since a year and a half ago too. 66 and 68 have the recolor and
blue word source today. recolor has less code than 2008 but
blue words have more. I see no reason the new 68 can't be
released but I haven't tested it against the old kernel since
internals have changed.
> I am reluctant to finish my own colorForth editor, knowing on the one
> hand that it is already out of date (no blue tokens), and on the other
> hand that fully working code already exists - it offends my sense of
> efficiency!
> All of that notwithstanding, your comments are like precious drops of
> rain in the arid desert of sourceless colorForth...
They didn't want to release colorforth 2.0a because the source
had been moved to a proprietary Forth. They wanted to wait until
the source to a new kernel was defined in colorforth. I lobbied
to have a release before that happened even if it was just the
kernel, editor, utilities sources, documentation, and Windows
wrapper. For one thing people had complained that they were
blocked because they needed qwerty and could not use dvorak, had
no documentation, or did not have the required hardware to run
native and needed a windows wrapper or some way to run hosted.
So they got all that stuff but not other stuff from inside of
okad II.
I didn't know how many years it would be before porting kernel
source to colorforth would become a priority and hoped that
people would want what was released and not complain too much.
It is not sourceless however like in Chuck's sourceless
programming days in the early 90s (threeForth and OK) it is
just that the cross compiler and kernel sources were not
publicly released. I could have let it all remain proprietary
and internal for a few more years.
As an application programmer I don't bother much with the
details of the internals of the kernel at all. I just boot
it up, use colorforth and the okad app and extend the okad
application suite. I am not one of the people who has been
making modificaitons to the kernel, the editor, the boot code,
editor extensions, etc. as that has been mostly Chuck, Greg,
and Charley. I have talked Chuck and Charley into doing some
talks about all the internal changes to SVFIG. Every chip
spin is in effect a new colorforth release so I see them
almost weekly and over the years changes accumulate. People
working with the seven, five, or year and a half old versions
will always be more than a little out of sync with people
working with today's release or last week's release.
There is not much I can do about that. Once in a while I
have taken the time to get other people to agree to and
release a public version and to do work on it myself to
help strip out all the proprietary stuff that is not be
released, put it on the web and to take time to answer
some of the questions that people have.
Best Wishes
peekay
---------- Forwarded message ----------
From: Wayne <news_putmynamehere...@optusnet.com.au>
Date: Aug 11, 6:03 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Mon, 10 Aug 2009 22:25:31 +1000, Dennis Ruffer
<druf...@worldnet.att.net> wrote:
> On Aug 9, 6:48 pm, Jeff Fox <f...@ultratechnology.com> wrote:
>> On Aug 9, 5:18 am, Dennis Ruffer <druf...@worldnet.att.net> wrote:
>> > On Aug 8, 12:09 pm, Jeff Fox <f...@ultratechnology.com> wrote:
>> > Is the colorforth IDE going to be released?
>> Probably. But I don't know when.
>> Best Wishes
> I hope you can convey to Chuck and his crew that this would be
> desirable. I realize that the community is pretty small right now,
> and I'm not sure what any of us can do to help. However, I do know
> that there is at least one (me) who would like to see what the
> colorforth interface can do.
Jeff,
How many targets will the IDE support? It would be good if it could
support the chips as well as the conventional PC's. What I am going
to
say might sound a bit silly, but it would work, that on the computer
versions the misc cores could be replaced by tasks (but otherwise
emulated
to speed up, if any brain deadening OS's are involved) depending on
word
sets involved, or interpretation of codec sequences messaging could be
sped up in chunks (however, the regular bit banging technique is out
of
the question obviously). External interfaces replaced by word sets
(most
likely custom word sets for each program on computers to emulate). In
this way people can run their misc code programs on a PC with minimum
rewrite and maximum speed (through some things will be slower). Using
standardised code sets to implement execution for external memory,
messaging and IO functions, synchronisation can be generated and
locally
efficient versions compiled. This is some of the work I am hoping to
do
on my vos for it. The benefit, very little, except, people can
commercially create all the code they want for the chips but also
commercially use it for PC's (not so useful for many things past
drivers
and IO at the moment, but i am aiming for regular consumer electronics
applications, and I hope Misc will move back in this direction
eventually)
but not running applications created for the PC on Misc, of course.
These
wiordsets can emulate a future more advanced version of the misc
processors and communications sub systems, allowing more seamless
upgrade
from the represent bit banging interprocess communications (I know,
much
loved interprocess communications scheme)). Pretty much what I also
was
planning if I could generate a successful enough first round product
that
justified, and capable of, funding it.
> Thanks for the update.
Yes too, as most always, thanks for the updates, better than the
announcements sections of many less than average tech companies.
Wayne.
peekay
---------- Forwarded message ----------
From: Jeff Fox <f...@ultratechnology.com>
Date: Aug 11, 10:54 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Aug 11, 6:03 am, Wayne <news_putmynamehere...@optusnet.com.au>
wrote:
> How many targets will the IDE support?
I don't understand the question. As many as are made?
> It would be good if it could support the chips as well
> as the conventional PC's.
If a conventional PC has a software layer to make it
able to execute incoming instruction streams in the same
format the a conventional PC could do it but it would
probably not make a lot of sense except in the sense
of following a standard.
> What I am going to
> say might sound a bit silly, but it would work, that on the computer
> versions the misc cores could be replaced by tasks (but otherwise emulated
> to speed up, if any brain deadening OS's are involved) depending on word
> sets involved, or interpretation of codec sequences messaging could be
> sped up in chunks (however, the regular bit banging technique is out of
> the question obviously). External interfaces replaced by word sets (most
> likely custom word sets for each program on computers to emulate). In
> this way people can run their misc code programs on a PC with minimum
> rewrite and maximum speed (through some things will be slower). Using
> standardised code sets to implement execution for external memory,
> messaging and IO functions, synchronisation can be generated and locally
> efficient versions compiled. This is some of the work I am hoping to do
> on my vos for it.
Yes it could be done.
> The benefit, very little, except, people can
> commercially create all the code they want for the chips but also
> commercially use it for PC's (not so useful for many things past drivers
> and IO at the moment, but i am aiming for regular consumer electronics
> applications, and I hope Misc will move back in this direction eventually)
> but not running applications created for the PC on Misc, of course. These
> wiordsets can emulate a future more advanced version of the misc
> processors and communications sub systems, allowing more seamless upgrade
> from the represent bit banging interprocess communications (I know, much
> loved interprocess communications scheme)). Pretty much what I also was
> planning if I could generate a successful enough first round product that
> justified, and capable of, funding it.
Yes. It is similar to the tack I took of writing simulators and
emulators
that ran MISC chip target code on a PC at up to about 25 Forth MIPS a
decade ago when the PC was clocked at couple of hundred megahertz.
The main advantage in my mind was that it made it possible for people
to learn the details and develop complete applications before the
target hardware existed. I wasn't doing it to port the same apps
to the PC but in fact one could run the iTV browser/email app
on a PC and telnet into it and run the internal Forth script
compiler etc. It is certainly possible to think of it as just
one of many implementations of Forth using a virtual machine
interface.
I thought that the byte-tokenized standard came and went with Sh-Boom.
Chuck and I thought it made sense in the 80s to use it as a standard
on the PC in the eighties for compatibility reasons. (Also because so
much stuff on a PC is byte centric.) But almost twenty years ago
things moved to five-bit tokens which do not map as well to Pentium
hardware. On a PC with lots of memory one could emulate all word
wide combinations of five-bit opcodes to get a more efficient
implementation (in terms of speed not size of the implemenation
given that you had had enough cache memory) instead of emulating
each five-bit opcode. This would also nicely map the way that
"dup ..." maps to single opcodes in register designs.
Most of the people who did this stuff in the past much prefer
to just work with real misc hardware rather than compatible
virtual machine implementations on a PC. It is why Chuck said
last week that softsim is so unsatisfyingly slow.
The waveguide synthesis code was a good example. It is much
more satisfying to be able to play and hear sound in realtime
than to wait for softsim to generate enough samples to hear
anything. Even if you have an efficient virtual machine
implementation unless your PC has hundreds of Pentium core...
Best Wishes
peekay
---------- Forwarded message ----------
From: Jeff Fox <f...@ultratechnology.com>
Date: Aug 11, 11:10 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Aug 10, 11:07 am, Pavel Klinkovsky <pavel.klinkov...@gmail.com>
wrote:
> Can you accept me to your group of interesed? ;-)
I am not sure what you mean. If you mean I am the only person
to try to answer some questions in c.l.f, yes. If you mean
you are interested in blogs or tutorials, yes.
I am not the person who hires people to work with Chuck. I
helped some of the people who have been hired to learn what
they needed to know to be qualified to do that sort of work.
I think it remains up to individuals to qualify themselves
unless Chuck institutes a program to official rate and
rank Forth programmer's skills so they could know who
best maps to which jobs as they come up. When Chuck
commented last year that this might be the way to go I
thought it would be funny if he put me in charge of that
project. I can think of a few people who would love
that. ;-)
Just as IntellaSys gave away chips, boards, development
systems, and jobs to people who demonstrated that they
had skills that could benefit the company I think it is
up to individuals to demonstrate their qualifications
by doing things. Chuck is always thrilled to have
qualified people contribute to his efforts.
It starts with interest but people also realize
that some approriate skills are needed. I recall
IntellaSys offering to give systems away at SVFIG
and getting no takers at the time. When there
was some interest expressed and arrangements
were made to offer chips through FIG IntellaSys
laid off most employees and put most project,
including that one, on hold days before the
chips were to be distributed. It is always
a gamble.
Best Wishes
peekay
---------- Forwarded message ----------
From: Howerd <howe...@yahoo.co.uk>
Date: Aug 12, 12:53 am
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
Hi peekay,
> it is ominous that 'precious drops of rain in the arid desert'
> are the only thing to look forward to
I didn't mean to sound pessimistic - just the opposite.
Jeff has done a great job publishing and explaining what he is allowed
to.
> otherwise it will just pass into oblivion
I doubt it - not so long as there are people who are interested in
computer science.
I think that colorForth is too cute to make it big in the commercial
world, but it can be used to make Forth chips which do some things
much better that the alternatives - these chips might make it big some
day...
Best Regards
Howerd
On 11 Aug, 15:00, peekay <pksharmakolk...@gmail.com> wrote:
> On Aug 10, 6:28 pm, Howerd <howe...@yahoo.co.uk> wrote:
> <snipped>
> > All of that notwithstanding, your comments are like precious drops of
> > rain in the arid desert of sourceless colorForth...
> i guess all ColorForthers are looking forward to use the
> facility of colored coding (perhaps on multi processor
> chips .. but even on the pc !)
> it is ominous that 'precious drops of rain in the arid desert'
> are the only thing to look forward to
> much more contents need to be tossed around if ColorForth is
> to reach its enourmous potential .. otherwise it will just
> pass into oblivion as a very good thingy .. the end !!
peekay
---------- Forwarded message ----------
From: peekay <pksharmakolk...@gmail.com>
Date: Aug 12, 1:35 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Aug 12, 12:53 am, Howerd <howe...@yahoo.co.uk> wrote:
> Hi peekay,
> > it is ominous that 'precious drops of rain in the arid desert'
> > are the only thing to look forward to
> I didn't mean to sound pessimistic - just the opposite.
> Jeff has done a great job publishing and explaining what he is allowed
> to.
> > otherwise it will just pass into oblivion
> I doubt it - not so long as there are people who are interested in
> computer science.
> I think that colorForth is too cute to make it big in the commercial
> world, but it can be used to make Forth chips which do some things
> much better that the alternatives - these chips might make it big some
> day...
> Best Regards
> Howerd
hi
it's very heartening indeed .. the clearing up of what i \
misunderstood .. and that in reality being very optimistic
is what should be understood from the comments
yes, mr. fox has been almost like a bedrock upon which
the ford citadel has been able to stand/survive .. he
definitely has been very very helpful with his very
detailed replies to forthers' queries over several
decades now
in addition, making several resources available on
webpages for viewing/downloading are very important
to keep the information flow fresh and interesting
it would help, of course, if mr. fox could be helped
by some more 'volunteers' and the necessary additional
info could be made available .. like ALL videos, ALL
codes of and for forth
BUT ..
maybe it is time to take a quantum jump .. and move
away from 'classical' flavours of forth (even though
they have a huuuuge following !) .. and to move into
the newer area .. ColorForth
i don't particularly feel the same way about it being
only 'cute' and therefore 'will-survive'
i strongly feel that ColorForth is the culmination of
decades of hardwork, insight, experience, trials, tests,
intelligent work .. and the creation of a suitable chip
to be able to 'do' the conceived executions
it is truly 'revolutionary' ! Putting 40 processors
on one die is even more revolutionary .. and putting
9x40 processors on one board for 'supercomputing' is
definitely amazing !!
the power of inline color usage for coding is neither
appreciated yet .. and (probably) nor understood fully
such a paradigm does not exist .. and so comparisions,
associations, similiarities can't be made/found and
understanding the full impact of this method is just
not happening
i feel that ColorForth-GA144, ColorForth-SAE40,
ColorForth-Haypress2 (360 processors on one board)
combinations are signposts to multiprocessor use
and are the way to go
since this solution is a code+chip thing, it has
far more potential of changing the world of computing
than the only code things like Linux, Android, Simple,
C++, Beos, etc
peekay
---------- Forwarded message ----------
From: Wayne <news_putmynamehere...@optusnet.com.au>
Date: Aug 15, 6:00 pm
Subject: GA4, GA32 and GA144 review .. GA32 talking to ColorForth on a
PC
To: comp.lang.forth
On Wed, 12 Aug 2009 03:54:16 +1000, Jeff Fox
<f...@ultratechnology.com>
wrote:
> On Aug 11, 6:03 am, Wayne <news_putmynamehere...@optusnet.com.au>
> wrote:
>> How many targets will the IDE support?
[code compatibility across Misc/nonmisc platforms]
> I thought that the byte-tokenized standard came and went with Sh-Boom.
> Chuck and I thought it made sense in the 80s to use it as a standard
> on the PC in the eighties for compatibility reasons. (Also because so
> much stuff on a PC is byte centric.) But almost twenty years ago
> things moved to five-bit tokens which do not map as well to Pentium
> hardware. On a PC with lots of memory one could emulate all word
> wide combinations of five-bit opcodes to get a more efficient
> implementation (in terms of speed not size of the implementation
> given that you had had enough cache memory) instead of emulating
> each five-bit opcode. This would also nicely map the way that
> "dup ..." maps to single opcodes in register designs.
Yes, five bit opcodes and self changing code, I forgot that. I tend
to
steer clear of self adjusting code, as it is not a virtual machines
best
friend. I am talking about compile targets, so five bit opcodes are
not
such a problem but self changing code would be a customisation that
would
be coded per target profile. What I was meaning, is that the compiler
would code to standard color forth instructions for each target
selected
(misc or x86 instructions, with the node branching added for misc but
direct formating on the PC) and the standardised inter-core
communications
(rom) and i/o word sets smooth out the other differences.
For my VOS product a code base that works across platforms and on the
PC/MAC/Iphone etc is important to grow market, but I thought a similar
scheme in color-forth might help GA and color-forth in general, to
also
grow a reusable code base for it's customers.
> Best Wishes
Thanks again.
Wayne.
pks kolkata
15 August Saturday morning
Been busy with layout for coming chip submission. GA4 has several nice enhancements and GA144 gets its first try.I'm working on a clock display: hours and minutes. Driven from the 10MHz resonator. To do this I have to address the issue of multiple applications within IDE. I've refactored some code and reorganized so that each application has its own script. That's working. Now to relay the sync message to the analog nodes
peekay
26 August Wednesday morning
<snipped>
Yesterday John and Charlie were here at the office.
Progress on OKAD's Softsim and code for the GA4 and GA144 chips.
We plan a download of colorForth that can support these chips.
Telecon with GreenArrays guys established tenetive pricing for GA4.
Much interest. GreenArrays' web site is up, tho without much content.
pks kolkata
from www.colorforth.com/blog.htm
4 September Friday morning
At the office. Haven't had much to say. Been working hard on Haypress Creek.
Here's a comparison of GA32 and 2 GA4
packages. We're
testing chips and making mock-ups for possible
applications. Hopefully mock-ups will help raise funds for
real boards. Testing, of course, goes on forever.
peekay
www.colorforth.com/blog.htm
27 September 9:00 Sunday
Tomorrow is moving day. GreenArrays is trading a $5K/mth
office for a $1K. Just across the street. We're hanging in
there, but it's easy to get discouraged.
peekay
www.colorforth.com/blog.htm
28 September 1:00 Monday
Moving day. Greg, Glenn, Dean, Charlie and I are packing and toting.
Trying to discard stuff without discarding value.
pk sharma
from : http://colorforth.com/blog.htm
16 October 10:00 Friday
<snipped>
We've a bunch of stuff to sort and dispose of. Including some 30 masks from a production run. Thinking of selling these for $100 each. The picture shows one such souvenir very well packaged.
<snipped>
photo :
