Explore the new system architectire of Apple Silicon Macs

On Sat, 04 Jul 2020 14:01:28 -0400, Paul wrote:

FFS Lewis, will you stop with that nonsense!

Hi Paul,

Now you know what I deal with every day on the Apple newsgroups...
o Type III apologists like Lewis aren't capable of adult logical thought.

Lewis truly thinks...
Since, with oodles of money, they can make a super computer using ARM silicon...
o Apple can therefore make a best in class PC CPU (calling it "Apple Silicon").

That's the same logic as if a red race car with STP stickers wins a race...
o Then people like Lewis run out to buy STP for their red Honda Civic.

Type III apologists like Lewis are _that_ incredibly immune to logic!
o Type I, like nospam, are essentially informal Apple MARKETING spokesmen
o Type II, like sms, are no malicious, just influenced greatly by MARKETING
o Type III, like Lewis & Alan Baker believe _everything_ MARKETING claims

Note: It's critical to realize that while Type I apologists like nospam
"sound" like Type III apologists like Lewis, the crucial difference is that
nsopam doesn't believe half of what he claims, while Lewis actually and
truly is telling us how his belief system works.

That is, while nospam realizes you can't compare a super computer to a PC
computer for this purpose, Lewis actually _believes_ that, since they "can"
make a super computer using ARM technology, that the Apple PC will, in
essence, be a super computer (of sorts).

Type I apologists don't even believe, themselves, most of their claims.
Type II apologists mostly just don't think all that much about the claims.
Type III apologists, sadly, actually believe everything that they claim.

The fact Type III apologists flock to purely imaginary belief systems...
o Is why (admittedly brilliant) Apple MARKETING holds sway on them.
--
Apologists like Lewis believe anything MARKETING feeds them to believe.

On 2020-07-04 14:01, Paul wrote:
Lewis wrote:
In message Paul
wrote:
Alan Baker wrote:
You get that Apple's homegrown chips are what they've been using in
iPhone since 2012, right?

If ARM was intended (by ARM Holdings) to be a rocket powered
car, it would already be a rocket powered car. I don't think
the ARM staff had any intention of going "head to head with Intel".

The world's faster super computer is an ARM based machine.

It is about 3 times faster than the #2 machine.

https://top500.org/
The new top system, Fugaku, turned in a High Performance Linpack (HPL)
result of 415.5 petaflops, besting the now second-place Summit system by
a factor of 2.8x.Â* Fugaku, is powered by Fujitsu’s 48-core A64FX SoC,
becoming the first number one system on the list to be powered by ARM
processors.


FFS Lewis, will you stop with that nonsense!

TOP500 computers run parallelized "perfect scaling" applications.

No. While the problem may be a rectangle, cube or higher order matrix,
the programming is such that it need not execute in lock step. Thus,
threads are thrown at cores as they become available and the solution
happens when the last thread completes. For most of the program run all
cores are working on the problem. At the end the sums happen (or
whatever other math) and the solution moves forward to the next iteration.

Perfection is the enemy of good enough - ESPECIALLY where super
computers are concerned.

Another example is protein folding. A massive problem well divided into
work units and distributed amongst many computers - different OS' and
CPUs and GPUs. Each does its package and returns its result and starts
on its next package.

Indeed, such held the record for super computing as recently as April.


There are very few things you can do on a TOP500 computer,
that have the necessary scaling properties. Not even Cinebench
would scale on there.

The majority of desktop code that matters, does not parallelize.
It's serialized.

It's multi-threaded. Which is why apps like Handbrake can process so
much, so fast, on multicore machines (to some limit). Which is why it
also scales to GPU's well (to some limit).


A desktop computer needs at least one computer core that runs at "top
clock".

This is why, a desktop processor that Intel made that turbos on
two cores to top clock, was such a win. It provided a core for
interrupt handling and various background activity, plus a core
that could run straight-line code at top speed. But even turboing
on one core is handy. (That's because Intel improved on the time
interval needed to change power states. It no longer takes 100us.)

Is that what ARM Holdings designed the ISA for ?
Can they turbo on one core ? To 5GHz ? And retire four
instructions per clock tick (IPC) ?

Like everything including x86, ARM's _evolve_. They are not even
properly RISC processors anymore - just as x86 is no longer solely CISC.

And like the x86, ARM lends itself to multicore architecture as well as
all the tricks of the trade (simultaneous execution while a branch is
being evaluated, etc.)

On my high-core-count machine, most of the time the
excess of cores are idle. The only time I get good usage
of the machine, is doing 7ZIP compression. Some
compression runs last for 24 hours.

I have a variety of apps that will keep 4 cores saturated and their HT
'ganger loaded to 70% or more. Indeed if I run 2 instances of
HandbrakeCLI I get perfect saturation.

At that point the CPU is thermally limited only.

If your machines were up-to-date, you'd be doing those runs in far less
time and use less energy to do so.

If all of the code in the ecosystem could be converted
to perfect-scaling parallelized code, then this exercise
has a hope of succeeding. I've not heard of any
breakthroughs in this regard.

Not needed. At all. Anymore than it is now on machines running over a
dozen cores.


Apple will simply try to bury the issue. And everyone
will leave the scene in a hype-mobile. As it should be.

I've got three Apple computers on the table I'm sitting at.
I've lived it. I know what it's like to run Unreal at 20FPS
and pretend I'm enjoying myself. Apple fixed that when
they moved to Intel. I would have no right to complain
about the platform today. But now what ? Well, I
guess we'll find out some day.

Very soon. And like when people were terrified of the PowerPC
abandonment for "crappy intel", it will all be forgotten a year later
when Apple/ARM shows its stuff on Macs.

Apple are not going backwards. But you're staying lost 2 decades worth.

Alan Baker wrote:


It has now been shown over and over that Apple designs its own chips.

What prevents people from running off and designing processors ?

It's called a patent portfolio.

There are like, a thousand patents you have to dodge, while
designing such hardware.

If anything is perceived as a threat to Intel, you
can be damn sure they'll decap that chip and have
it scanned for patent infringement. There is a
company that specializes in this topic, of scanning
for patent material. Intel doesn't even have to warm
up their own electron microscope.

Oh, my, is that a ring oscillator I see in there ?
Is that a charge pump on some dynamic circuitry ?
Who has the patent on that ? And so on.

When Microsoft said "we're going to make an x86 software interpreter
for our new ARM box". Intel threatened them. Microsoft backed off.

Which company has the richest patent portfolio ?

Will they go to patent war with one another ?

Will Apple boxes be stopped at the border, on their way from China ?

Paul

On Sat, 04 Jul 2020 15:19:44 -0400, Paul wrote:

What prevents people from running off and designing processors ?
It's called a patent portfolio.

Hi Paul,

It's similar to the hurdle that Apple/Intel found out when they tried (and
failed, so far anyway) to build something as simple as a "modem".

... *Even after spending _billions_ of dollars on the modem project!* ...

HINT: As you noted, it's all about the patent portfolio.
o Now compare a "modem" to a PC CPU in terms of potential patents

(Although, unlike Lewis, I realize they're fundamentally different so I'm
only asking for a comparison on size & complexity, not on digital
performance versus analog/rf performance, which is mostly the real
technology comparison between the CPU & modem comparison).

(Also note that licensing the ARM technology obviates at least the ARM
patents, which, of course, is how and why ARM makes money off of Apple.)

Apple has never created a best-in-class modem; Apple has never created a
best-in-class GPU; Apple has never created a best-in-class smartphone
CPU... and yet, these apologists "believe" Apple will create a
best-in-class ARM CPU for their beloved Mac.

Maybe Apple will pull it off...
o But more than likely, Apple MARKETING will tout it more than it's real.

Remember, of _all_ of high tech, NOBODY can find a _single_ high-tech
company that has a _lower_ R&D spend % than Apple. Nobody.

They _hate_ that Apple is all MARKETING and almost no actual R&D...
o They _believe_ only in the MARKETING (which, admittedly, is brilliant).

And yet, the MARKETING works on them basically because they're ignorant.
o I assess, personally, that Alan Baker's IQ is roughly around 40 or 50.

Just as I did for Snit.

In two decades on Usenet, I've only bothered to killfile two people:
o Snit & Alan Baker
Where that alone should tell you what I think of anything Alan Baker says.
--
2 kinds of people post to Usenet: Those who add value; and those who can't.

In article , Paul
wrote:

If anything is perceived as a threat to Intel, you
can be damn sure they'll decap that chip and have
it scanned for patent infringement. There is a
company that specializes in this topic, of scanning
for patent material. Intel doesn't even have to warm
up their own electron microscope.

apple isn't making anything that infringes intel ip, so no.

apple is making arm chips of their own design, which they've been doing
for a decade or so.

Arlen Holder wrote:
On Sat, 04 Jul 2020 15:19:44 -0400, Paul wrote:

What prevents people from running off and designing processors ?
It's called a patent portfolio.

Hi Paul,

It's similar to the hurdle that Apple/Intel found out when they tried (and
failed, so far anyway) to build something as simple as a "modem".

... *Even after spending _billions_ of dollars on the modem project!* ...

HINT: As you noted, it's all about the patent portfolio.
o Now compare a "modem" to a PC CPU in terms of potential patents

(Although, unlike Lewis, I realize they're fundamentally different so I'm
only asking for a comparison on size & complexity, not on digital
performance versus analog/rf performance, which is mostly the real
technology comparison between the CPU & modem comparison).

(Also note that licensing the ARM technology obviates at least the ARM
patents, which, of course, is how and why ARM makes money off of Apple.)

Apple has never created a best-in-class modem; Apple has never created a
best-in-class GPU; Apple has never created a best-in-class smartphone
CPU... and yet, these apologists "believe" Apple will create a
best-in-class ARM CPU for their beloved Mac.

Maybe Apple will pull it off...
o But more than likely, Apple MARKETING will tout it more than it's real.

Remember, of _all_ of high tech, NOBODY can find a _single_ high-tech
company that has a _lower_ R&D spend % than Apple. Nobody.

They _hate_ that Apple is all MARKETING and almost no actual R&D...
o They _believe_ only in the MARKETING (which, admittedly, is brilliant).

And yet, the MARKETING works on them basically because they're ignorant.
o I assess, personally, that Alan Baker's IQ is roughly around 40 or 50.

Just as I did for Snit.

In two decades on Usenet, I've only bothered to killfile two people:
o Snit & Alan Baker
Where that alone should tell you what I think of anything Alan Baker says.

Ah, yes. I showed you things I could do very easily in iOS and my Mac that
you could not replicate. Freaked you out.

Love it!

--
Personal attacks from those who troll show their own insecurity. They
cannot use reason to show the message to be wrong so they try to feel
somehow superior by attacking the messenger. They cling to their attacks
and ignore the message time and time again.

On 2020-07-04 12:19 p.m., Paul wrote:
Alan Baker wrote:


It has now been shown over and over that Apple designs its own chips.

What prevents people from running off and designing processors ?

It's called a patent portfolio.


And?

There are like, a thousand patents you have to dodge, while
designing such hardware.

If anything is perceived as a threat to Intel, you
can be damn sure they'll decap that chip and have
it scanned for patent infringement. There is a
company that specializes in this topic, of scanning
for patent material. Intel doesn't even have to warm
up their own electron microscope.

And Apple is designing and building chips based on ARM's ISA...


Oh, my, is that a ring oscillator I see in there ?
Is that a charge pump on some dynamic circuitry ?
Who has the patent on that ? And so on.

You think that Apple doesn't have people whose job it is to make sure
their designs don't infringe on patents they don't own?


When Microsoft said "we're going to make an x86 software interpreter
for our new ARM box". Intel threatened them. Microsoft backed off.

Which company has the richest patent portfolio ?

Will they go to patent war with one another ?

Will Apple boxes be stopped at the border, on their way from China ?

Would you care to make a small wager?

Which company has more MONEY: Intel or Apple?

Get this through your tiny, narrow, silly little mind:

APPLE IS ALREADY DESIGNING AND USING ITS OWN CPUs AND HAS BEEN SINCE 2012.

On 2020-07-04 12:36 p.m., Arlen Holder wrote:
On Sat, 04 Jul 2020 15:19:44 -0400, Paul wrote:

What prevents people from running off and designing processors ?
It's called a patent portfolio.

Hi Paul,

It's similar to the hurdle that Apple/Intel found out when they tried (and
failed, so far anyway) to build something as simple as a "modem".

... *Even after spending _billions_ of dollars on the modem project!* ...

HINT: As you noted, it's all about the patent portfolio.
o Now compare a "modem" to a PC CPU in terms of potential patents

(Although, unlike Lewis, I realize they're fundamentally different so I'm
only asking for a comparison on size & complexity, not on digital
performance versus analog/rf performance, which is mostly the real
technology comparison between the CPU & modem comparison).

(Also note that licensing the ARM technology obviates at least the ARM
patents, which, of course, is how and why ARM makes money off of Apple.)

Apple has never created a best-in-class modem; Apple has never created a
best-in-class GPU; Apple has never created a best-in-class smartphone
CPU... and yet, these apologists "believe" Apple will create a
best-in-class ARM CPU for their beloved Mac.

Apple is demonstrably producing the best-in-class CPUs for smartphones.

'As of March 2020, the best mobile processor is Apple A13 Bionic which
powers the iPhone 11 Lineup.'

https://www.techcenturion.com/smartphone-processors-ranking

https://www.myfixguide.com/best-smartphone-processors-ranking/

https://www.phonecurry.com/benchmarks

In article , Alan Baker
wrote:

On 2020-07-04 12:19 p.m., Paul wrote:
When Microsoft said "we're going to make an x86 software interpreter
for our new ARM box". Intel threatened them. Microsoft backed off.

Which company has the richest patent portfolio ?

Will they go to patent war with one another ?

Will Apple boxes be stopped at the border, on their way from China ?

Would you care to make a small wager?

Which company has more MONEY: Intel or Apple?

Get this through your tiny, narrow, silly little mind:

APPLE IS ALREADY DESIGNING AND USING ITS OWN CPUs AND HAS BEEN SINCE 2012.

actually, 2010 with the a4, and prior to that, they designed a wide
variety of custom chips.

On 2020-07-04 2:12 p.m., nospam wrote:
In article , Alan Baker
wrote:

On 2020-07-04 12:19 p.m., Paul wrote:
When Microsoft said "we're going to make an x86 software interpreter
for our new ARM box". Intel threatened them. Microsoft backed off.

Which company has the richest patent portfolio ?

Will they go to patent war with one another ?

Will Apple boxes be stopped at the border, on their way from China ?

Would you care to make a small wager?

Which company has more MONEY: Intel or Apple?

Get this through your tiny, narrow, silly little mind:

APPLE IS ALREADY DESIGNING AND USING ITS OWN CPUs AND HAS BEEN SINCE 2012.

actually, 2010 with the a4, and prior to that, they designed a wide
variety of custom chips.


To be fair, the A4 was not an all-Apple design.

I've been looking into this and the first all-Apple chip appears to have
been the A6.

In article , Alan Baker
wrote:


APPLE IS ALREADY DESIGNING AND USING ITS OWN CPUs AND HAS BEEN SINCE 2012.

actually, 2010 with the a4, and prior to that, they designed a wide
variety of custom chips.


To be fair, the A4 was not an all-Apple design.

I've been looking into this and the first all-Apple chip appears to have
been the A6.

maybe not entirely but it wasn't just an off the shelf part either.

and then there is the mystery of the a1, a2, and a3, which presumably
were internal prototypes, with the a4 being the first public release.

On 2020-07-04 4:22 p.m., nospam wrote:
In article , Alan Baker
wrote:


APPLE IS ALREADY DESIGNING AND USING ITS OWN CPUs AND HAS BEEN SINCE 2012.

actually, 2010 with the a4, and prior to that, they designed a wide
variety of custom chips.


To be fair, the A4 was not an all-Apple design.

I've been looking into this and the first all-Apple chip appears to have
been the A6.

maybe not entirely but it wasn't just an off the shelf part either.

and then there is the mystery of the a1, a2, and a3, which presumably
were internal prototypes, with the a4 being the first public release.


Don't forget the A7f

On Sat, 4 Jul 2020 16:56:35 -0400, JF Mezei wrote:

Rosetta 2 will cnvert an x86 OS-X binary to an
ARM binary before the applicatiion is launched. So the application
launches as a native ARM app and runs natively.

Hi JF Mezei,

You're an adult, so the conversation is _different_ than with nospam.

You're also not an apologist, so, again, the conversation with you is
different than with nospam, just like the conversation with a five year old
(like Alan Baker appears to be) is, of necessity, quite different than a
conversation with an actual adult.

My point about nospam is that he always parrots Apple MARKETING.

Interestingly, nospam is actually smart enough to know that (roughly) half
of what he claims, he, himself, knows is utter bull**** (sort of like how
most MARKETING is - where a tad is based on truth, but the rest is
innuendo).

That's why there are three types of apologists, with respect to beliefs:
o *Type I*, e.g., nospam, claim it whether or not they actually believe it
(e.g., nospam might claim it's "Apple Silicon" while knowing it's not)
o *Type II*, e.g., sms, simply filter out facts, where they're just lazy
(e.g., Alan Browne might claim it's "Apple Silicon" without even thinking)
o *Type III*, e.g., Alan Baker or Lewis actually completely believe the BS
(e.g., whatever Alan Baker claims, rest assured, he actually believes!)

Obviously the scary ones are nospam, if you believe him...
o But he's not scary once you realize you can predict his every word.

Fundamentally, like a defense lawyer, he'll excuse whatever Apple does
o Mostly by blaming Google or Microsoft or even Amazon - for Apple's flaws
o But often by simple doublespeak or brazen denial of even well-known facts
o Often he resorts to claiming imaginary functionality that doesn't exist
etc.

If you disagree, I'll simply provide you with proof via cited threads.
o *Why do apologists like Alan Baker & nospam desperately try to shift*
*the _blame_ of Apple bugs to Google & Microsoft?*
https://groups.google.com/forum/#!topic/misc.phone.mobile.iphone/LOQx1Ok-79c

The apologists have only 7 responses to fact, none of them adult.
o *What are the common well-verified psychological traits of the*
*Apple Apologists on this newsgroup?*
https://groups.google.com/forum/#!topic/misc.phone.mobile.iphone/18ARDsEOPzM

My point on nospam's use of "Apple Silicon" is that, as a type I apologist,
he _knows_ full well it's not even remotely close to Apple Silicon; and he
knows exactly why Apple is frantic for us to NOT call it what it is, which
is ARM Silicon (e.g., product differentiation); however, deep down, nospam
will never admit it, but he _knows_ it's not even close to Apple Silicon.

The ones to really worry about are the Type III apologists like Alan Baker
and Lewis, who actually _believe_ the MARKETING bull****.

It petrifies me, for example, that they're given the same vote as we are.
--
People _that_ fundamentally prone to believing MARKETING should't exist.

On 2020-07-04 2:22 p.m., nospam wrote:
In article , Alan Baker
wrote:


APPLE IS ALREADY DESIGNING AND USING ITS OWN CPUs AND HAS BEEN SINCE 2012.

actually, 2010 with the a4, and prior to that, they designed a wide
variety of custom chips.


To be fair, the A4 was not an all-Apple design.

I've been looking into this and the first all-Apple chip appears to have
been the A6.

maybe not entirely but it wasn't just an off the shelf part either.

Which doesn't make it an Apple designed chip.


and then there is the mystery of the a1, a2, and a3, which presumably
were internal prototypes, with the a4 being the first public release.

One would assume, yes.

But given that the A4 was an Apple chip with ARM-designed cores, I think
it's only logical to assume that they either used ARM designed
components... ...or were failures to go all-Apple first time out.

On Sat, 04 Jul 2020 17:22:04 -0400, nospam wrote:

To be fair, the A4 was not an all-Apple design.

I've been looking into this and the first all-Apple chip appears to have
been the A6.

maybe not entirely but it wasn't just an off the shelf part either.

Doublespeak.

What's interesting is Alan Baker apparently claims the ARM Mac is "Apple
Silicon", even as he must then ignore that the fundamental patented
technology comes from ARM (hence, it's really ARM Silicon but Apple is
frantic we don't realize that fact - which is why they're desperate for us
to not call it the "ARM Mac" or the "Mac ARM" or "ARM Silicon", etc.)...

And yet...
o He and nospam agree that the A4, isn't.

Doublespeak.
--
The apologists live in an imaginary world of all-MARKETING doublespeak.