On Tue, 7 Dec 2021, 04:26 Rick C, <gnuarm.deletethisbit@gmail.com> wrote:
On Monday, December 6, 2021 at 10:04:47 AM UTC-5, Wayne morellini wrote:
> On Mon, 6 Dec 2021, 09:49 Rick C, <gnuarm.del...@gmail.com> wrote:
> On Sunday, December 5, 2021 at 9:49:42 AM UTC-5, Wayne morellini wrote:
> > On Sunday, December 5, 2021 at 2:41:29 AM UTC+10, gnuarm.del...@gmail.com wrote:
> > > On Saturday, December 4, 2021 at 11:00:13 AM UTC-5, Wayne morellini wrote:
> > > > On Sunday, December 5, 2021 at 1:24:59 AM UTC+10, gnuarm.del...@gmail.com wrote:
> > > > > On Friday, December 3, 2021 at 11:26:41 PM UTC-5, Wayne morellini wrote:
> > > > > > On Saturday, December 4, 2021 at 11:31:55 AM UTC+10, gnuarm.del....@gmail.com wrote:
> > > > > > > On Friday, December 3, 2021 at 6:35:47 PM UTC-5, Wayne morellini wrote:
> > > > > > > > On Saturday, December 4, 2021 at 7:30:47 AM UTC+10, gnuarm.del...@gmail.com wrote:
> > > > > > > > > On Friday, December 3, 2021 at 1:40:16 PM UTC-5, Paul Rubin wrote:
>
> > Rick, you ok?
>
> I'm fine. Are you a native English speaker?

You seem to have goofed up the attribution, but that's ok, I'm going to step out of this conversation. It is clear that either you are not a native English speaker or you literally slept your way through school.

Rick I have seen you carry on this way here over the years. You are not superior to everybody you argue against, and I see below you are keeping it up . This is not acceptable. You have been proven wrong most times in this email too.. Time to give it a break!. It's not I who have problems following things. "It's" was the historical possessive, as explained.

You know that GA, has talked about being approached to develop options to demonstrate (presumably based on the existing chip) but the client obviously went with another option. That is usually protected by NDA. As I rightly said in business, NDAs can cover these things, whatever type of NDA document, incase you were planning on going there.. If you have an axe to grind don't take it out on people. You also notice I, and others here, don't say GA did things contrary to what is known.

> > GA is the only one we are talking about. "It's about where they will be in time, even in 5 years" .

>>"About commercial survival." is also not a sentence.

> Do we get annoyed at street signs, point lists, columns of numbers, forth code. A lot of that are not sentences, and nor do you need a proper sentence structure to convey correct information, as all those examples show. So, it isn't superior to use the best grammer, but it is to convey the correct information, isn't that correct?
>
> Now, I don't know about 'is' actually being used as a verb there, rather than a place (adjective, possession of place, which is obvious), and not traveling to the place, but where the place maybe in 5 years time. But then again "it's about" is that doing something (probably) or possessing something? But, "it's" originally used to be the possessive, and I still use it that way, and it is a form of usage. But the modern usage of "it's" is a corruption of the original word "ist". "Is" is the indicative mood of "Be", so maybe the be in maybe constitutes a verb in the adverb, do it's still makes sense.

Wow! You really need to go back to school to get a proper education. "It's" is not possessive. "Its" is possessive. "Is" is the verb of the sentence. Read the definition of "is" in a dictionary and it will explain how to properly use it. "Is" has a number of meanings, but as used above it is an intransitive verb, present tense third-person singular of be with a meaning as listed here.

https://www.merriam-webster.com/dictionary/be

"It's" is discussed here.

https://www.merriam-webster.com/dictionary/it's

"Its" is discussed here.

https://www.merriam-webster.com/dictionary/its

See my previous true statements about it.

> Anyway, thanks for the heads up. In clarifying that, I used it a few different ways, all of which talked about a future state in 5 years time. But, to do everything mechanistically (and English is not the language to be doing that in, as it has changed so much, and does change, as the living language. It's grammar has become quite a challenge) is to be mechanistic, rather than flexible.

That is just rationalization. Either you had nothing to say or you simply don't know English well enough to say it.

But, one says it with understanding.

> I've seen you debate things here with various people over the years. But, this is not good timing. I have a dieing parent, and I'm not doing the best either. However, I do know of people who get contracted rigid thought when sick. You have indicated you have issues too.

> > You know under NDA, they can report very little, if at all. You know that's how it works.
>
> NDAs do not prevent reporting of activities of a company such as profit or sales reports. They just can't mention the customer. As I said, there is no indication GA has ever bought a second batch of chips. How long has that been, 10 years? More? Sorry, GA has never been a viable, functioning company by any realistic measure. I run a company that sells one product to a company that orders when they get government contracts. So I've had years with zero sales or profit. Based on what we know of GA, my company would be considered a more viable company. At least it has sales and reports them on income tax forms.
> I didn't really indicate that I thought GA was a great success, but they kept going offering funded services.

They are operating the way I do, essentially zero overhead like a spider waiting in his web for a "customer" to land.

> In a publically traded company, you have certain reporting responsibilities. In a privately traded company, you don't. They can keep their books and deals closed under NDA, to regular people. Reporting those things can adversely affect the amount of new work.

No one worries about keeping "books" closed other than the information pertaining to them. A customer has no reason to ask GA to never make any announcements about the activities of the company.

That's definitely wrong. You said you have a one product company for some military application? You wouldn't know from that experience. It's the activities related to the client they keep safe, and their books otherwise, to attract more clients. Occasionally they shared something, and the client would have OK'd that.

So what is your military product you mentioned?

This is not worth arguing about. The bottom line is GA is a failure by nearly any objective measure whether you wish to accept that or not.

You are arguing about it. Failure is a classic subjective term. What's failure to one, is mild success to the other doing it. If I ran something like this, I would be depressed. There is just so much that could be done.

> > > > > So how would money be made from such designs?
> > > > What they have been doing, yes, but it is scary design for people, and the regular arm is more comfortable. In the 1980's, it would have been a great design, even in the 1990's, but it really needed the 18 bit 640kB address space, even back then. At least one processor with access, if not most or all of them. Now. I wonder why for nearly two decades. I remember, you talked about doing software radio with it, but it was just st too our there and restrictive for a modern high datarate format. My recent designs proposals are suitable for that, but this needs to work at lower data rates. Where a custom asic can dominate it. I am concentrating on how changing tac might produce a better marketable product.
> > > In the 1980s feature sizes crossed 1 um. The GA144 would have been a much larger chip (around a square inch)
> Hmm, I should have said "design concept", where the size of the array can vary. Speed doesn't matter (except in relation to th competition) it was about the style of the concept, and the level of technical proficiency required to program it. It reminds more of something coming out of the 1970's. But, the point was, by the end of 1990's, people would have been so used to easier development solutions, that it's day would have passed, and ASIC integration was the future. A single 640KB direct addressable memory would have compensated a bit. But, now, the market is ultra basic, small, low or high energy, or integrated. I would still find use for a piece of wire small 5Ghz single or 8 chip, mind you, just because it can do everything in an application.

I think you don't understand the chip very well. It is not an MCU like any other.

How do you get that from what I. I am writing? I know this is very limited, and what could have helped a bit.

At that point the device has lost all it's advantages and have about the same transistor count per CPU as the 8086.

With memory, that the 8086 didn't have, otherwise it would be close to 1000-4000 transistors per actual processor. I would dump the execution memory for the core hooked to external execute memory. Actually, were we talking about a piece of wire (ultra narrow) chip, 5ghz, maybe 8 core? That was talking about a modern process. So, it should fit.

> But, about designing it with an end product in mind. Processors are designed to be suitable to develop many end products, Probably not too many with only the be product in mind.

You are clearly not familiar with the MCU market. No, an MCU is not designed for a single app, but they are definitely targeted to market segments. That's why there are so many different MCUs out there. Each one has a specific combination of features designed to be optimal for some range of applications. That means the applications have to be known as well as a good estimate of production volumes to be expected. No one says, "Let's build this MCU because someone *might* be interested in this range of features". It doesn't work that way.

That is what I said, different versions for different markets. Clearly, I am familiar with the market, unlike what you said. But I also addressed that the designed for one product issue. If GA had done what was suggested, they could have made a software configurable more general purpose, MCU. You seem to be randomly skipping through things and getting things wrong. I often read through and deliberately avoid that, otherwise I would post illogical garbage instead of following the conversation. Speed reading, good, dense information, writing, is to just miss most of the information. I can see the top of the text below, that auto correcting on Android has just trashed the information in a typo and substituted nonsense. Sorry about that. I've been sick and posting before going to sleep, with all the other issues. and missed that.

One thing I must say. Is the GA, can be thought as a chip where you can stick custom routines in each address space in a fab run, and use the processor memories as further customisations and working memory. Otherwise, it just gets close to being a software filter stage, and not much more fitting in the memories of the current end user chip. My old application, a projection video game, the single core with a main memory, would have been fine, as I was going do vector graphics, and use the array for calculations and limited 3D calculations. Very simple and close to what I needed. But, they didn't give one core that mount of power. I could sit here all day and talk about ways to improve the design, then I would be just giving away most of my secrets here for nothing. The issue was, they didn't do this. Why didn't they sit down and nut out some simple improvements like these?

> Was the x18 such a chip, no. People would have preferred a child thru could use multiple ways. Giving the x18 some main memory and simple programming, is not the solution now either. That time has past. Now, you need something which targets the iot, wireless data, data, and integrated markets. Generating bit coins costs money, having something to do that for a lot less money, is a product. But, bit coin becomes less relevant, or they find other ways to generate them cheaper, then that product becomes less relevant too. But, there is something that has a future market 10 years ago. Server chips had a market too, but companies have wised up to doing them better. So, what's available has shrunk and changed, and what's left are still pretty big markets. But the current chips aren't there. Their advanced 32 bit was/is, and probably a 16 bit array version of that for alongside a single 32 bit or 16 bit (depending on market) with full main memory access. But, it is likely there would need to be specific integrations for different markets, with heavy integration in the integrated market, to over compete with what's out there on a desirable level.

What you are describing above is the idea of designing a chip for an application, a target market. That's what didn't happen with the GA144.

That is basically what I said. It was on what they should have done instead of what they did do.

If the software side of things were adequately developed, it might well be practical to use a GA type chip for high performance processing such as crypto or stock trading. But it would need to be tailored for that. I suspect the current devices for those markets are already well optimized and it would be tough to insert a produce remotely like a GA144.

That's what I said here.

> About the 1 micron in 1980's chips you mentioned, didn't they get 0.8 and 0.6 devices on the market by 1990?

So? You want to split hairs and work with the latest technology from 1989? I seem to recall GA not using the latest technology because of the costs. It was less pronounced then, but there was still a cost hit for using the latest and greatest processing technology.

Rick, you are splitting hairs bringing in old 1 micron, and the speed and size compared to the current as if that was the only thing available. While, the move to 0.8 or 0.6, makes a huge difference to your arguments, not splitting hairs.

It would be reasonable to say the architecture fits at whatever speed and size and number of processors it could do back then. The only logical thing, is how it would compare to other chips back then. Didn't the mup get 80-100mips in 1990, so we could reasonably expect lower energy for a unit of speed. You would also know, that the reason 180nm was chosen, was the energy performance of the process, not just that it was an old process.

> What is transmitted means nothing. For a direct conversion the carrier frequency is probably limited to 100 MHz or less. That would require a significant input signal with good SNR.

> Yes it does. Modern high end communications standards deal with a lot of bandwidth. Even old wireless HD (was it, I get the names confused. It was the one from the developers of HDMI) was 27Gb/s, or maybe more. If we are talking about some sort of digital ham survivalist wireless network, then sure, the data rate you could get out of it maybe good. But, in many cases, an integrated mass produced chip is going be ok.

Bluetooth is digital and has a data rate similar to serial ports, a few hundred kHz, maybe a couple of MHz. Don't connect things that aren't connected.

Why are you ignoring new interfaces, why are you trying to justify on old standards only, and ignoring what is being talked about. If it was me, I would have something better tomorrow. I advise you to look up the latest Bluetooth standards, as with many things here. The GA solution just can not be as good as a good ASIC version can be, for an new Bluetooth solution, or for many solutions. It is also obvious that what is being transmitted matters on the x18, as "modern high end communications" is beyond it, and Bluetooth is old. I'm talking about sports car tyre, and you throw in a wagon wheel in comparison.

If you say Bluetooth is suitable, then do top end Bluetooth and prove it. You have the wealth and health and been talking about this for many years.

> The idea of a F18A like device with an 18 bit address space, fully populated by memory is a contradiction. The speed of the F18A is due to every part of the device using optimal structures, often very different from what is used in most designs. Give it a 256 kW memory and it would run much slower.. Even with a 4 kW memory it would be slower.

> It was originally stated that the 18 bit bus matched an 18 bit static ram cache chip, which presumably can attach to just one processor in the array, making nearly 0 difference in the energy consumption and performance if not attached, and great differences in performance of certain product applications, of used.

I have no idea what you are trying to say with that.
You should.

A static RAM chip would use more power than the GA144. The entire memory interface was a failure. They never connected it to DRAM which would have been a useful interface and SRAM was not useful because of the high power. This is another poorly thought out feature with no actual application in mind.

You are not following again. That's the talk back then, right or wrong back then. It uses virtually 0 power if not connected, so you shouldn't be trying to change what I said.

> Again, that was back then, and now those chips are probably going be expensive. However, memories that use die capacitance are a viable alternative to static memory, integrated in the same die with the processor. There is probably a market for a 32 bit advanced version with 4-16GB of memory able to be powered off in banks that aren't used, somewhere out there.

Really? That's your claim that there must be *some*

Stop! You know I said "probably" and you wrongly said I said 'must'. This is not on, and it's not the only time things have been changed to make a falsified argument.

application, *somewhere*??? DRAM is not at all optimal for the same process as used for MCUs. You have to sacrifice to combine them. Do they even make 16 GB DRAM? My computer came with 16 GB and I'm pretty sure that's multiple chips. I think the chips today are 4 GB and below. Here's a 4 GB device for over $100 each. It also uses up to 404 mA of current. But this is actually a twin die package so the dies are 16Gb or 2 GB. Even then the part is bigger than a GA144. So you might be able to incorporate 256 MB or 512 MB on a common die, but at what cost?

You are not being true again. Re-read things, it's talking about a type of memory capable of being put on a ASIC process. Please don't come and try to win and muddy everything, again. Whole irrelevant posts do not help, it's time to stop.

Yeah, you lost me in all that. It's getting too close to a manifesto.

Then it probably pays to be silent, if you can't talk to others properly.