The Future of Computing (According to Me)

Started by PabloMack, May 22, 2019, 07:29:52 PM

Previous topic - Next topic

WAS

You seem to have similar interests as me, though it seems you're decades more experienced in the fields. I just research and occasionally dabble with CPUs with emulation and OS's.

Particularly I love your interest with TG as a target application. Definitely interested in hearing about updates here.

I recently have had peak interest in Windows 10 ARM64. I just started figuring out what's wrong with some drivers and trying to diagnose to write/wrap my own drivers. My goal is to get Windows 10 running on my older LG with AM64 Cortex-A53 in preparation for use on Cortex-A76 (my LG G8 ThinQ). I am hoping to get a native installation going, solve driver issues for what is needed... and  than try to install Terragen so I can see all it's errors and what it needs (if it's not compatible. I'm not sure what all it's based on, I know some win32 apps work fine in ARM64).

I want to take TG with me anywhere, and honestly, the Cortex-A76 with it's 8 cores, 8 threads, at 1,800 MHz, 2,420 MHz, 2,840 MHz is pretty adequate for TG and could get decent performance on mobile machines. And with Microsoft targeting ARM64 (and ditching exclusive phones) for new small end devices it's something id' like to see.

I know you're not hot with RISC, and neither am I in the end, I'd love to see more conventional architecture in mobile devices and, even computers so your post here really peaks my interest.

PabloMack

#16
Quote from: WASasquatch on May 31, 2019, 03:44:39 PMYou seem to have similar interests as me,

Thanks for the post. Now I don't feel so lonely. I have written a couple of recent
articles on LinkedIn that you might be interested in. I don't think I need to replicate
that information here.

https://www.linkedin.com/pulse/hypocrisy-hollow-promises-risc-paul-mckneely/
https://www.linkedin.com/pulse/lessons-learned-cpu-design-paul-mckneely/

I was just looking at the ARM64 manual and I'd hate to have to write a code
generator for that monster. I really miss Motorola's manuals because they
were so much better than anyone else's. What I want for the average
programmer is something that they can easily understand. I think it is a
tragedy when CPUs are so complicated that it puts the control of the whole
market into the hands of a few billion-dollar companies.

WAS

Quote from: PabloMack on June 03, 2019, 03:06:34 PM
I think it is a tragedy when CPUs are so complicated that it puts the control of the whole
market into the hands of a few billion-dollar companies.

I agree! That's been a serious issue of mine as well. "Research Teams" and "Labs" for researching CPU architecture and innovation, like it's national security level development. LOl

I remember a guy brought in a CPU he made to school, even was attempting his own CGA graphics. I was just simply amazed that his apparatus was running, and even ran basic ASCII games (fairly well). That was the real start of my interest over just "hey this tech is cool, I want it".

Checking out your papers now.

N-drju

Quote from: PabloMack on June 03, 2019, 03:06:34 PM
I think it is a tragedy when CPUs are so complicated that it puts the control of the whole
market into the hands of a few billion-dollar companies.

On the other hand, this is just business and this is how things work. You can't blame them for winning a market or two.

Do you seriously think these companies have turned "billion-dollar" overnight? Back in the days, they had to make some "first steps" too!
"This year - a factory of semiconductors. Next year - a factory of whole conductors!"

WAS

#19
Quote from: N-drju on June 05, 2019, 07:36:48 AM
Quote from: PabloMack on June 03, 2019, 03:06:34 PM
I think it is a tragedy when CPUs are so complicated that it puts the control of the whole
market into the hands of a few billion-dollar companies.

On the other hand, this is just business and this is how things work. You can't blame them for winning a market or two.

Do you seriously think these companies have turned "billion-dollar" overnight? Back in the days, they had to make some "first steps" too!


Intel's 4004 research was actually quite costly, additionally it came in 4 variants, and each chip at $5 dollars for production cost... That would be like developing a chip with a fab cost of $31 dollars today.  By contrast today Intel is punching out about 10-15$ dies. The rest is all markup based on FAB and consumer market.

Most these companies are actually invested in. They're hardly ever out of pocket.

PabloMack

#20
Quote from: N-drju on June 05, 2019, 07:36:48 AM
On the other hand, this is just business and this is how things work. You can't blame them for winning a market or two.
Do you seriously think these companies have turned "billion-dollar" overnight? Back in the days, they had to make some "first steps" too!


I think you know I never said otherwise so I don't know where it is you think we disagree.

Quote from: WASasquatch on June 05, 2019, 12:01:30 PMIntel's 4004 research was actually quite costly, additionally it came in 4 variants, and each chip at $5 dollars for production cost... That would be like developing a chip with a fab cost of $31 dollars today.  By contrast today Intel is punching out about 10-15$ dies. The rest is all markup based on FAB and consumer market.

This is interesting. The engineering technology that is available today to you and me makes
designing processors far easier than it was back in 1970 time-frame when this took place.
That's pretty exciting, don't you think?

N-drju

Quote from: PabloMack on June 05, 2019, 05:26:16 PM

I think you know I never said otherwise so I don't know where it is you think we disagree.


The first quote.
"This year - a factory of semiconductors. Next year - a factory of whole conductors!"

PabloMack

#22
Quote from: N-drju on June 06, 2019, 06:50:29 AM
Quote from: PabloMack on June 05, 2019, 05:26:16 PM

I think you know I never said otherwise so I don't know where it is you think we disagree.


The first quote.


I'm still puzzled. Of course what you wrote is true and I haven't written anything to the
contrary.  You do imply that I said that billion dollar company's don't deserve market share
and that they happen overnight but of course I never said any such things. I will accept
that as a misunderstanding on your part. So I guess we don't have an issue seems to me. 
What is it that is bothering you and you are not saying? ;)

N-drju

I just thought that it was hard for you to accept the fact that the aforesaid companies are leading the CPU development and are not eager to share their findings with anyone else. The "tragedy" bit gave me this impression.

CPU architecture is not a trivial thing in general, so, in a way I don't consider it "lamentable" that this equipment is complicated and some companies consider it (rightfuly, I'd say) their "specialty". But then again, by the end of the day, you have millions of users who don't care at all and will just buy their PCs and laptops with whatever CPU on board is offered.

I'm not trying to be a jerk, please understand. I just think that considerations about regular CPUs are not relevant when a CG-dedicated CPU is what you try to investigate. These are two completely different concepts that can co-exist, regardless of what happens to the other concept.
"This year - a factory of semiconductors. Next year - a factory of whole conductors!"

PabloMack

#24
N-drju,

I think we are on the same page as they say. What you say is all true. But I
just noticed your Personal Text and I think it expresses what I have been
trying to say:

"When dictatorship is a fact, revolution becomes a right."

Thank you for saying it for me.  :D

Because you are not a programmer I think you may not fully appreciate that programmers are
under duress because CPUs are made unnecessarily complex. It is not just you who think
they are not easy to deal with. Mega-corporations are known for killing small company's so that
they can take their market share.

One time my sister married a bank president because she wanted a husband that could provide
her with lots of money. This guy had scary ties to the Mafia. She said one time he was drunk
and told a story about his "boss" who hired a hit-man to murder a competitor. The crime was
never solved. Of course it wouldn't bode well for a company like Intel to become publicly
embroiled in a scandal like that. But lot's of large companies still do unethical things that they
are sure they can get away with. It happens.

One tactic that really works is that they can add a lot of unnecessary complexity to technology
they control so that only they can bear the burdens. Of course, it is not always intentional and
it often happens because of poor designs that come out of "design by committee". I saw this
happen on a large scale when ARM came out. A lot of small software tool vendors folded. Of
course, this strategy can backfire when a small company finds a better way to do something.
Many innovations often happen with small companies because they can make decisions much
more quickly and don't carry the baggage that comes with large companies. That's why we
need our small business to remain healthy. But that's also a reason why the big companies
don't want them around. Many big companies have strong lobbies in Congress so that they
can build barriers that keep small businesses out. When a few large corporations control a
market it is a form of dictatorship. As in political dictatorships, the average bum doesn't want
to rock the boat because of the trouble he can see that it could bring to him personally. He's
not concerned about anyone else like you say. But this bum doesn't understand that the other
things that put him in chains (I'm speaking figuratively) are indirectly caused by the very things
that he doesn't see any reason to change.

I have a client who is having difficulty getting his products certified because there are people
on the standards committee that have adopted standards that are aimed at killing their
competition. They expressly will not allow my client to use a technique that is very effective
and safe because the companies they work for don't do it that way. My client became a member
of the standards committee so he can get the rules changed. He talked with the people who
were responsible for the rules and they admitted that they made the rules so they could
eliminate their competitors and not for any logical reasons that relate to safety or anything
else that the rules exist for. Only large companies can afford to send their employees around
the world to participate in this form of dictatorship.

N-drju

#25
Quote from: PabloMack on June 06, 2019, 07:43:46 AM

"When dictatorship is a fact, revolution becomes a right."

Thank you for saying it for me.  :D


Touché.  ;)

I get your point. You are right - I do not know much about the computer industry and its business intricacies but I can believe that companies like Intel would lobby to secure a customer base for themselves. However, note one thing - don't you think it's telling that a billion-dollar company is afraid of small tech companies influence and ideas? This reminds me of the other discussion down here that I had with archonforest and other guys where it was agreed that once people get used to something, they loose their interest trying to seek alternatives...

Just a rhetorical question - would it be possible at all to cooperate with a big company, in order to capitalize on a concept? Even if for a 15% profit share?

On a different note, I would like to remind you of an interesting story that perfectly illustrates how small companies can beat the big ones.

I am sure you know Maxis - the company which made millions from the "Sims" and "Sim City" series.

Maxis was a monopolist and a pioneer as far as the city-building genre is concerned. Somewhat like Intel now. They scored one success after another since SC 2000. SC 3000 (and its "Unlimited" edition) were cash cows plain and simple. Note that all of this has been intertwined with new editions of Sims and games like "Spore" which were also increasing the firm's popularity and prestige.

And then, Sim City 5 came. This game has been a spectacular flop since day one. This title has been plagued with problems. Overloaded servers (the game was online-exclusive at one point, unlike other titles) and a ludicrously small building area for a city (one square mile? seriously?) were among some of the most serious. Maxis tried hard to salvage the title with new updates and patches but it was a mess beyond belief.

Few months later, a small, virtually unknown company named "Colossal Order" (don't be fooled by the name - it's just about 10 people!!) reluctantly released their own city-building game - "Cities: Skylines." And guess what? Every gamer out there, jumped right on this train as the game proved to be more interesting and technically superior to its counterpart. The CO's release was a complete sucker punch to Maxis and believe me when I say they lost a lot of their customer base back in 2013.

So as you can see - there is hope and a lot of possibilities for small companies. Sure, sometimes it means that you need to wait until they screw something up.  ;) But I generally think that consumers are not as much of a sheeple as businesses would like to think and one should also count on their support. A good example of that appeared a couple of days ago when Apple announced their new touchscreen product and declared that one can optionally buy a $999 stand for it. Contrary to what Apple spin-doctors assumed, the audience were not reaching out for their wallets...
"This year - a factory of semiconductors. Next year - a factory of whole conductors!"

PabloMack

N-drju,

I am with you on what you wrote.  ;)

Oshyan


PabloMack

#28
Quote from: Oshyan on June 08, 2019, 09:40:38 PM
Looks like the EU has a venture to "reinvent" CPU architecture as well:
https://www.techspot.com/news/80425-european-union-cpu-development-branch-delivers-first-designs.html

Any thoughts?

- Oshyan

Thank you for the link. I was unaware of such a thing.

According to the article, it is nothing but a re-branding of ARM and RISC-V.
Apparently, Advanced RISC Machines has been lobbying in government
to garner more (mandatory) support for its products. No surprise there. And
RISC-V is license free. No wonder about that. The article also says that it will
not be used by the consumer markets but only in embedded commercial
applications. That's the part of the article that I like.

To my knowledge, there has been no new development of a 64-bit CISC
architecture since the whole world has jumped onto the RISC band wagon.
RISC proponents would claim that RISC performance is better than CISC
on the basis of comparing 30-year old CISC processors to new RISC ones.
How objective is that? There are many flaws in the arguments by RISC
promoters and one is that RISC processor instruction sets are more
orthogonal that CISC. That is just a lie. Because RISC instructions are
forced to be all the same size, they can't be as orthogonal as a well-
designed CISC processor CAN BE. Nobody has seen what CISC can do
with modern methods and component densities because nobody has
tried in 30 years. The x86 is an old architecture that has had many face-
lifts but current implementations perform admirably when compared
modern RISC machines. Imagine how well a MODERN CISC machine
could perform when implemented using modern methods. Nobody
knows because nobody has tried because of herd mentality.

While almost all processors are targeted at using current software technology
(an understandably short-sighted strategy), ϕEngine is a co-design effort to
correct past mistakes and bring about changes that can't happen unless the
ways we do things are changed simultaneously. The attitude has been "I'm
not going to change unless you do it first." So we fail to move forward because
nobody wants to take the risk (no pun intended).

But change HAS happened and not for the advancement of computer technology
itself. It has happened to accommodate ancient writing techniques. It is called
Unicode. A while back I wrote this article that you might find interesting.

https://www.linkedin.com/pulse/what-most-people-dont-know-ascii-good-bad-unicode-paul-mckneely/

My dream has been to develop a system that implements many advanced features
that have been lacking in our computers since the beginning. This system will be
used mostly by enthusiasts who want a better way of doing things. The machine
will be easy to write code for because it will make sense. It will not depend on
hardware and compilers that can only be developed by multi-billion dollar corporations
for obscure, incomprehensible or undocumented processor architectures.  Apple has
always been a pretty closed system and they soon plan to dump Intel and will be
using processors of their own design. This seems to be a good strategy for
getting control over all of the aspects of its design so that 3rd-party developers can
be left out in the cold. You will no longer be able to get documentation on how
to program the processor because it will all be inside confidential information.
So all you can do is to choose to buy what they offer or go somewhere else.

My vision is to do something like Linux has done in the way of open-source operating
system development. But Linux has gone in so many directions and it carries a
lot of old baggage. It would be too difficult for it to evolve into what I see as the
future computer. The ϕSystem is more revolutionary. It will incorporate many of
the good standards we are already using but it will discard things that should be
relegated to the past. Of course, we will have to use our current systems to develop
the new platform.

Oshyan

Thanks for the detailed reply. I sort of wondered if you might be able to seek funding from them given the stated goals of the EPI itself (separate from the current choice of approach).

- Oshyan