Maybe this is not the right place to post about this but I just haven't thought of a better one at the moment so...
I have been interested in this marvelous chip for some years now but never actually had a chance or a possibility to try it out.
Last night, after watching this video from eight years ago I decided it's about time to do so.
However, the problem I'm facing is I can't justify spending that kind of money (plus all applicable taxes and shipping costs) for the eval board while I'm sure it can be made at almost an order of magnitude cheaper price. Well, maybe at least five times cheaper if approached right.
Does anybody else here have any interest in such a board so we could join forces to make that happen ?
Looks like the chips are still available and although not exactly cheap but surely within the reach of a motivated individual (me, for example).
Replacing three expensive and dumb USB>serial chips from FTI with something more modern and universal and far cheaper like the RP2040 for example would surely help reduce the total BOM costs especially than maybe just one RP2040 would suffice.
The Forth2020 group would be the best place to talk about this, after all we had Chuck himself on our October zoom. When I asked him what was the signature app that they use the GA144 for, he had to think about it (bad sign) and said it would make a great Forth PC (albeit with external RAM of course) but other than that it was nada nix nic zip. It was as I suspected because if they had a power app for it then we would hear all about it.
In all this time I could never figure out how to use it, and it seems neither can Chuck. But it's fast although it has to communicate everything through its neighbors, and its really really low power, but effectively useless IMHO. Unless Chuck can demonstrate why we should use it, other than as a Forth PC, then I wouldn't bother with it at all. Mind you, I asked Chuck if he thought it would make a great Forth PC, then why hadn't he done even that in all this time?
Sorry to sound negative, but the GA144 needs to be a GA32 with more memory/core and better I/O connectivity so that it will do more with less. I have other Forth chips such as the PSC1000 ShBoom and Mup21, and RTX2010 etc, and while they are interesting chips, they all need a lot of support chips and memory. Not exactly a single-chip MCU.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Maybe you're right but than I figure if they are still in business as a company then they surely must have some bigger customers than just forth freaks. It's just that they're invisible to us, regular folks. Calling the chip useless might be a bit of a stretch although it might really be useless to many. It is an experimental chip for sure but so was the P1 back then and some other odd ball ones to. If I can get a base board for it done at fairly reasonable price I will go for it anyway. Not every project must succeed and sometimes there are offspring of failed projects that can be of great use for something else.
EDIT:
Regarding the signature app. What's the signature app for the RP2040 or a P2 ? Because I can think of many apps for both but I would be hesitant to call any of these a signature app.
Last edit: Maciek 2021-10-17
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I did ask Chuck a direct question because in all the time the chip has been around there is no user base and no apps, let alone a signature or killer app. I've followed it for many years and the app notes are disappointing.
With the RP2040 and the P2 we can show lots of apps and really useful stuff such as Christof's stuff but I've done a ton of stuff with it too. Also, without any extra memory chips I can have a standalone Forth PC with video, keyboard, SD, etc.
The trouble with the GA144 is that it has 144 little CPUs that have very little memory and depend upon their chain of neighbors to connect them to the outside world. There is nothing easy about configuring the CPUs to work together and do something useful. The P1 and P2 on the other hand are super easy to program and for any cog to access any or all I/O directly.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
If somebody as experienced as yourself, both in hardware and Forth, gives me good reasons against then I would be either stupid or curious beyond sanity limits not to take them seriously. I will not scrap this off my list of interest completely but, say, relocate it further down and do some more research and thinking.
For now, looks like you talked me out of spending at least 200USD plus duties and taxes that could be better spent elsewhere. So at the end it was good to ask this question now and get discouraged early on.
BTW, do you plan to layout the board for that "standalone Forth PC with video, keyboard, SD, etc." any time soon ? The heart of it is already there in a form of the P2D2 (or whatever improved version). I've always been interested in such a device.
Last edit: Maciek 2021-10-17
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I would never want to discourage anyone from "giving it a go" but I share my thoughts plain and simple on the matter. Even if the GA144 was easier to program there is still the problem with external memory. Sure, you can connect it up to all those I/O islands but that it is not efficient nor a good use of resources either. At least the P2 has 512kB of shared RAM and shared I/O in addition to the 496 words each of cog memory. If a "GA32" had that then I'd be interested, even if the F18A memory was still only 128 words.
Yes, I find the F18A very interesting itself but awkward as a GA144 CPU array. We need an MCU.
Yes, on the subject of a Forth PC I have my P2LAB board which forms the basis for that plus it is expandable and embeddable controller or a test instrument. I hope to push out new artwork this month for the P2D2 and the P2LAB etc and then I might jump back onto the Parallax forums 😄
btw, the other kicker is that the GA144 is a 1.8V device and you will need extra buffers even just to light an LED.
Last edit: Peter Jakacki 2021-10-18
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Sounds like a good plan and you are dearly missed there, at the Parallax forums.
BTW and completely off topic here but I've noticed some new, cheap 8051 chips from SiLabs emerged at Mouser and DigiKey. I'm tempted tho I do not have immediate use for them right now but these are real MCUs with 1.8V to slightly over 5V supply voltage. Nice and convenient.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Thinking about that rather dated eval board they offer, I wouldn't think that they would sell many of them. If I were GreenArrays I would give away a swag of them to all kinds of devs who are interested. Even I would be interested and invest my time in seeing what is possible, but I am not going to invest time and money in taking a risk. The devs would all sign up to a GA forum of some kind where they would share their work and ask questions etc. For a tiny investment on top of their big investment GA would have some real apps and a growing knowledge and expertise base. Who knows, I might even implement my zoom audio array processor in one of these devices.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Surprise surprise, I discovered that GreenArrays have a GA4 with four of the improved F18B CPUs in a tiny 2x2 or 3x3 DFN package. I love the idea of that much more so than the awkward GA144 so I emailed them telling them so and asking how do I get my grubby little hands onto some chips.
I can see these as very fast and low-power I/O "dots" that can be distributed around a board and perhaps have the main MCU boot them up, or vice-versa.
Let's see what happens.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Interesting. I missed that information completely on their site. Need to dig deeper or more thoroughly. Maybe both. Thanks for the tip.
EDIT: I've found it ! But I also recall why I completely neglected that chip in the first place. It's still experimental after all these years and probably will never make it to be put on the shelves. But then, maybe they have lots of them buried deep in their drawers gathering dust.
Last edit: Maciek 2021-10-19
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Greg Bailey shot an email back even though it must have the middle of the night saying thanks and thanks but they only have about 50 each of those chips and the rest they used for test boards. He said many other things but the main point I get is that the GA144 is the chip to use since it has everything and if someone has a specific need then they will tailor a chip for them. They need the $$$$$$ and I ain't got the $$$$$$.
I noticed too on the website that they had prototyped different versions including a GA32 and GA40 and also one that ran I/O at 5V etc. But all this stuff is really old now and I wonder have they survive as a company.
Anyway, I could test with a GA144 but what's the point because the chips are $20 a piece and there is no smaller and cheaper GA32 or GA4 for sale if I try to use these in production. I think if the P2 used an F18B per I/O then they would be a lot more powerful than a smartpin or an RP2040 PIO.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Maybe this is not the right place to post about this but I just haven't thought of a better one at the moment so...
I have been interested in this marvelous chip for some years now but never actually had a chance or a possibility to try it out.
Last night, after watching this video from eight years ago I decided it's about time to do so.
However, the problem I'm facing is I can't justify spending that kind of money (plus all applicable taxes and shipping costs) for the eval board while I'm sure it can be made at almost an order of magnitude cheaper price. Well, maybe at least five times cheaper if approached right.
Does anybody else here have any interest in such a board so we could join forces to make that happen ?
Looks like the chips are still available and although not exactly cheap but surely within the reach of a motivated individual (me, for example).
Replacing three expensive and dumb USB>serial chips from FTI with something more modern and universal and far cheaper like the RP2040 for example would surely help reduce the total BOM costs especially than maybe just one RP2040 would suffice.
Here is the link to the Green Array site.
Last edit: Maciek 2021-10-17
The Forth2020 group would be the best place to talk about this, after all we had Chuck himself on our October zoom. When I asked him what was the signature app that they use the GA144 for, he had to think about it (bad sign) and said it would make a great Forth PC (albeit with external RAM of course) but other than that it was nada nix nic zip. It was as I suspected because if they had a power app for it then we would hear all about it.
In all this time I could never figure out how to use it, and it seems neither can Chuck. But it's fast although it has to communicate everything through its neighbors, and its really really low power, but effectively useless IMHO. Unless Chuck can demonstrate why we should use it, other than as a Forth PC, then I wouldn't bother with it at all. Mind you, I asked Chuck if he thought it would make a great Forth PC, then why hadn't he done even that in all this time?
Sorry to sound negative, but the GA144 needs to be a GA32 with more memory/core and better I/O connectivity so that it will do more with less. I have other Forth chips such as the PSC1000 ShBoom and Mup21, and RTX2010 etc, and while they are interesting chips, they all need a lot of support chips and memory. Not exactly a single-chip MCU.
Maybe you're right but than I figure if they are still in business as a company then they surely must have some bigger customers than just forth freaks. It's just that they're invisible to us, regular folks. Calling the chip useless might be a bit of a stretch although it might really be useless to many. It is an experimental chip for sure but so was the P1 back then and some other odd ball ones to. If I can get a base board for it done at fairly reasonable price I will go for it anyway. Not every project must succeed and sometimes there are offspring of failed projects that can be of great use for something else.
EDIT:
Regarding the signature app. What's the signature app for the RP2040 or a P2 ? Because I can think of many apps for both but I would be hesitant to call any of these a signature app.
Last edit: Maciek 2021-10-17
I did ask Chuck a direct question because in all the time the chip has been around there is no user base and no apps, let alone a signature or killer app. I've followed it for many years and the app notes are disappointing.
With the RP2040 and the P2 we can show lots of apps and really useful stuff such as Christof's stuff but I've done a ton of stuff with it too. Also, without any extra memory chips I can have a standalone Forth PC with video, keyboard, SD, etc.
The trouble with the GA144 is that it has 144 little CPUs that have very little memory and depend upon their chain of neighbors to connect them to the outside world. There is nothing easy about configuring the CPUs to work together and do something useful. The P1 and P2 on the other hand are super easy to program and for any cog to access any or all I/O directly.
If somebody as experienced as yourself, both in hardware and Forth, gives me good reasons against then I would be either stupid or curious beyond sanity limits not to take them seriously. I will not scrap this off my list of interest completely but, say, relocate it further down and do some more research and thinking.
For now, looks like you talked me out of spending at least 200USD plus duties and taxes that could be better spent elsewhere. So at the end it was good to ask this question now and get discouraged early on.
BTW, do you plan to layout the board for that "standalone Forth PC with video, keyboard, SD, etc." any time soon ? The heart of it is already there in a form of the P2D2 (or whatever improved version). I've always been interested in such a device.
Last edit: Maciek 2021-10-17
I would never want to discourage anyone from "giving it a go" but I share my thoughts plain and simple on the matter. Even if the GA144 was easier to program there is still the problem with external memory. Sure, you can connect it up to all those I/O islands but that it is not efficient nor a good use of resources either. At least the P2 has 512kB of shared RAM and shared I/O in addition to the 496 words each of cog memory. If a "GA32" had that then I'd be interested, even if the F18A memory was still only 128 words.
Yes, I find the F18A very interesting itself but awkward as a GA144 CPU array. We need an MCU.
Yes, on the subject of a Forth PC I have my P2LAB board which forms the basis for that plus it is expandable and embeddable controller or a test instrument. I hope to push out new artwork this month for the P2D2 and the P2LAB etc and then I might jump back onto the Parallax forums 😄
btw, the other kicker is that the GA144 is a 1.8V device and you will need extra buffers even just to light an LED.
Last edit: Peter Jakacki 2021-10-18
Sounds like a good plan and you are dearly missed there, at the Parallax forums.
BTW and completely off topic here but I've noticed some new, cheap 8051 chips from SiLabs emerged at Mouser and DigiKey. I'm tempted tho I do not have immediate use for them right now but these are real MCUs with 1.8V to slightly over 5V supply voltage. Nice and convenient.
Thinking about that rather dated eval board they offer, I wouldn't think that they would sell many of them. If I were GreenArrays I would give away a swag of them to all kinds of devs who are interested. Even I would be interested and invest my time in seeing what is possible, but I am not going to invest time and money in taking a risk. The devs would all sign up to a GA forum of some kind where they would share their work and ask questions etc. For a tiny investment on top of their big investment GA would have some real apps and a growing knowledge and expertise base. Who knows, I might even implement my zoom audio array processor in one of these devices.
Surprise surprise, I discovered that GreenArrays have a GA4 with four of the improved F18B CPUs in a tiny 2x2 or 3x3 DFN package. I love the idea of that much more so than the awkward GA144 so I emailed them telling them so and asking how do I get my grubby little hands onto some chips.
I can see these as very fast and low-power I/O "dots" that can be distributed around a board and perhaps have the main MCU boot them up, or vice-versa.
Let's see what happens.
Interesting. I missed that information completely on their site. Need to dig deeper or more thoroughly. Maybe both. Thanks for the tip.
EDIT: I've found it ! But I also recall why I completely neglected that chip in the first place. It's still experimental after all these years and probably will never make it to be put on the shelves. But then, maybe they have lots of them buried deep in their drawers gathering dust.
Last edit: Maciek 2021-10-19
Greg Bailey shot an email back even though it must have the middle of the night saying thanks and thanks but they only have about 50 each of those chips and the rest they used for test boards. He said many other things but the main point I get is that the GA144 is the chip to use since it has everything and if someone has a specific need then they will tailor a chip for them. They need the $$$$$$ and I ain't got the $$$$$$.
I noticed too on the website that they had prototyped different versions including a GA32 and GA40 and also one that ran I/O at 5V etc. But all this stuff is really old now and I wonder have they survive as a company.
Anyway, I could test with a GA144 but what's the point because the chips are $20 a piece and there is no smaller and cheaper GA32 or GA4 for sale if I try to use these in production. I think if the P2 used an F18B per I/O then they would be a lot more powerful than a smartpin or an RP2040 PIO.