Solved! ARM Apple High-End CPU - Intel replacement

Page 13 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Richie Rich

Senior member
Jul 28, 2019
470
229
76
There is a first rumor about Intel replacement in Apple products:
  • ARM based high-end CPU
  • 8 cores, no SMT
  • IPC +30% over Cortex A77
  • desktop performance (Core i7/Ryzen R7) with much lower power consumption
  • introduction with new gen MacBook Air in mid 2020 (considering also MacBook PRO and iMac)
  • massive AI accelerator

Source Coreteks:
 
Reactions: vspalanki
Solution
What an understatement And it looks like it doesn't want to die. Yet.


Yes, A13 is competitive against Intel chips but the emulation tax is about 2x. So given that A13 ~= Intel, for emulated x86 programs you'd get half the speed of an equivalent x86 machine. This is one of the reasons they haven't yet switched.

Another reason is that it would prevent the use of Windows on their machines, something some say is very important.

The level of ignorance in this thread would be shocking if it weren't depressing.
Let's state some basics:

(a) History. Apple has never let backward compatibility limit what they do. They are not Intel, they are not Windows. They don't sell perpetual compatibility as a feature. Christ, the big...

DrMrLordX

Lifer
Apr 27, 2000
21,807
11,161
136
Not sure if it's ever been attempted to take the PCB out and ditch the battery - many modern phones have HDMI/DP alt mode over USB C so that should take care of display/audio, assuming that you can shunt enough power with mains alone.

Unfortunately, if you do that, you do not have a phone anymore. Sounds like you'd have one of those Snapdragon reference platforms that costs over $1k for the board alone.

I have a board with a Cortex-A75 (Snapdragon 845), but yeah I'd like something with more performance such as A76 (and preferably that does not reach crazy temperatures due to no fan).

That's why I got an ROG Phone II. It ships with its own fan (hah) and has a dock accessory that, among other things, also has a fan on it. It also lets you hook up kb+m and a monitor. Not sure if it can bypass battery power, but basically, when it's docked, it's on a charger. Too bad it's so expensive.

Between it being so hard to get modern ARM cores in a package that allows any kind of hacking and it Apple cores being locked up behind iOS, it sure is hard to play with anything ARM.
 
Reactions: soresu

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
Some people work for love of the product, some for the paycheck. If the paycheck is ALL that motivates you, it's probably not a good idea to work at Apple (or any other environment that expects more than that).
Somehow I doubt that the "love of the product" people would take a pay cut lying down or quietly.

It's one thing to take pride in your work, but generally people expect to get paid for that time put in, and equal to the contribution involved - though as a former menial worker I can soundly say that such a thing as contribution/effort is seldom honoured with appropriate pay in that line of work.
 

Nothingness

Platinum Member
Jul 3, 2013
2,757
1,405
136
Of course I worked there for ten years and loved every minute of it...
They treated me extraordinarily well, I had very good colleagues, they let me control my hours as long as I delivered.
That's a very interesting feedback. Can you tell more (here or by PM)? In what division did you work (my ex-colleagues and friends work in the CPU or GPU design divisions)? Was that an engineering role? Was it long ago (company changes and often not for the better alas)?

Some people work for love of the product, some for the paycheck. If the paycheck is ALL that motivates you, it's probably not a good idea to work at Apple (or any other environment that expects more than that).
I think my post made it clear: money is not my motivation, or I would have accepted one of the offers I got. Of course having a correct salary is important but what matters most to me is a challenging work, freedom of speech (including criticizing management/company decisions), flexible work rhythm, great colleagues, freedom to choose what I work on (as long as it is useful to the current project(s) obviously).

When someone tells you that when he sat at his desk in an Apple building, the guy seating next to him was not allowed to even tell what he was working on, that's definitely killer for me.
 

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
When someone tells you that when he sat at his desk in an Apple building, the guy seating next to him was not allowed to even tell what he was working on, that's definitely killer for me.
That certainly would seem counter intuitive, for morale sake if not for efficiency sake - all work and no play make Jack a dull boy.
 

DrMrLordX

Lifer
Apr 27, 2000
21,807
11,161
136
So Linux on Android on Linux. That's some inception stuff.

It is, but it's clever. Basically they're taking advantage of the fact that, at its core, Android uses a Linux kernel. So they create a separate filesystem as a binary within the Android file management system once the application has been given permission to do so. Then you have to slowly piece together one of the supported distrios (like Ubuntu or Debian or what have you) package by package from appropriate repos. One of the apps that does this is UserLAnd:


Supposedly UserLAnd can actually run without root on the phone via proot, but I remain skeptical as to whether or not that's a workable solution. There's another competing app called Linux Loader that basically does the same thing (and I think it requires rooting the phone). It's a pity nobody has developed something similar for Apple devices since like . . . iPhone 4 or so? iOS isn't Linux at its core, which makes things complicated. There is some FreeBSD under the hood. But it's complicated. Anyway, being able to run FreeBSD on top of/besides iOS would be hella sweet. Too bad it'll never happen.
 

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
iOS isn't Linux at its core, which makes things complicated. There is some FreeBSD under the hood. But it's complicated.
It's based on Darwin which is BSD derivative, and way back in the lineage of MacOS at that - what code it shares with FreeBSD is likely to have been replaced or overhauled at least once in the years since Darwin was branched/forked from BSD - unless of course Apple decided to lessen their burdens and share some common code directly from the stable FreeBSD code base..

BSD is partially interchangeable with Linux from what I have seen due to their common POSIX roots, certainly programs written for Linux can be run on BSD variants with a not so huge amount of effort.

Linux Firefox was available on PC BSD in mostly up to date form when last I tried it before it became True OS (admittedly a looong time ago now).
 

name99

Senior member
Sep 11, 2010
449
333
136
Silly question, wafer space of course - each thing that takes up less area space on the SoC means more dies per wafer, and as high yield as possible too I should imagine.

Please troll somewhere else.
We are not idiots. Obviously more area is not absolutely free. But more area delivers value. I covered that in my post "Are you happy to save 10mm^2 (so 10% of the SoC area) at the cost of losing 30 to 50% of your CPU performance? What's the point?"

If you don't want to engage in the conversations, just snipe silly comments, can you please do that somewhere else?
 

name99

Senior member
Sep 11, 2010
449
333
136
That's a very interesting feedback. Can you tell more (here or by PM)? In what division did you work (my ex-colleagues and friends work in the CPU or GPU design divisions)? Was that an engineering role? Was it long ago (company changes and often not for the better alas)?

I think my post made it clear: money is not my motivation, or I would have accepted one of the offers I got. Of course having a correct salary is important but what matters most to me is a challenging work, freedom of speech (including criticizing management/company decisions), flexible work rhythm, great colleagues, freedom to choose what I work on (as long as it is useful to the current project(s) obviously).

When someone tells you that when he sat at his desk in an Apple building, the guy seating next to him was not allowed to even tell what he was working on, that's definitely killer for me.

I worked in QuickTime from about 1992 to about 2002. I started off doing the MPEG code, moved to the codec guts of the DVD player, and then became the sort of general codec optimization guy working to speed up everything from Sorenson to JPEG2000 to AAC.

I did not experience any sort of paranoia or feel like things I needed to know about were being held from me. But I suspect a lot of this is cultural/personality. Some people feel that they have the right to know everything about everything, from the salary of everyone in the company to the strategy ten years out to M&A plans. Other people are quite willing to accept that some degree of secrecy (in personal life, in politics, and inside a company) is inevitable. I don't think the company makes much difference -- if you demand total transparency you're going to be unhappy anywhere. If you demand the right to blab company secrets to anyone, you're going to be unhappy anywhere. If you feel the you have the right to make political rants about your colleagues or management, don't be surprised when they decide they're not interested in hearing yet another of your rants.

I knew what I needed to know to do my job. I was as surprised as anyone else by new Apple stuff that wasn't relevant to my job --- new models, new OS features, iPods, ...

Most of the things you are talking about are issues about which reasonable people come down in different places for different topics and different circumstances. If you are willing to be a reasonable person, and to accept that other people --- perfectly DECENT REASONABLE people -- have different opinions, then act like an adult, find a job you love, and accept that it's a job, it's not a marriage, where you expect every single detail to be shared with you, and agreement on every single aspect of how life is to be lived.


I would say one final thing. The happiest people are people who are honest with themselves. They know what they are good and bad at. They know what they want from life and from a romantic partner. They know what they want from a work environment.
Then they choose and optimize based on that.

That all may seem obvious. But in fact 95% of the people I know don't honestly know themselves. They started projecting a persona sometime around 8 years old, and have projected it so long they can no longer tell that's it's not their true self. This is especially obvious when it comes to socially contentious matters, where they have spent so much time stating more and more extreme political claims that, in fact, they kinda just slid into. But it seems like now they're stuck in this loud persona that just isn't them. And of course it's even worse if you have 20yrs of public social media statements along these lines...

So that's my point. What I hear (maybe I'm wrong, but this is what I hear in your comment) is someone who has been swept up in various aspects of techno-hysteria over the past few years. For some it became privacy hysteria. For some it became walled-garden hysteria. For some it becomes work environment hysteria. But in all cases close to zero knowledge coupled with massive echo-chamber feedback led people to lay down lines that (honestly) they just don't care about. But now it feels like they've told everyone that Apple (or Google or Samsung or Microsoft or Intel) is the devil because [...] and they can't walk that back to go work there.

So that's the issue:
- Do you know your real self as opposed to the persona you've constructed?
- Are you brave enough to destroy that persona and create one that better matches who you now are?
- Are you smart enough to ignore the next wave of hysteria (and the next, and the one after that) to admit "I don't need to have an opinion on that"?
- Are you willing to accept that not everything has to be viewed as a political battle to the death, to accept that some people, people smarter than you, even people more ethical than you, may have looked at the same facts you have, and concluded something different? And still be willing to treat those people with respect?
- Is your response to any disagreement going to be "can we figure out a solution that works for both of us"? Or is it going to be "someone is TRAMPLING ON MY RIGHTS. I DEMAND SATISFACTION!"

Almost none of this has to do with Apple. But this is what life in large organizations is like. If that is not the sort of person you are, stick to small organizations.
 
Last edited:

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
Please troll somewhere else.
We are not idiots. Obviously more area is not absolutely free. But more area delivers value. I covered that in my post "Are you happy to save 10mm^2 (so 10% of the SoC area) at the cost of losing 30 to 50% of your CPU performance? What's the point?"

If you don't want to engage in the conversations, just snipe silly comments, can you please do that somewhere else?
It was not intended as a trolling exercise, but rather a basic maxim of die area economics, no more no less.

I did not call anyone an idiot, you asked a question to something I believed to have an obvious answer (hence the "silly question" remark) - I am astounded at how often I have experienced reasonably intelligent people miss obvious things right in front of them, I'm sure there's probably a psychology paper or 10 on the subject.

Value is a subjective term relative to the market space the product is aimed at, and what the product itself is.

There are plenty of SoC's that do not remotely aim for maximum performance - hence why so many SoC's have only A53 or A55 cores (why it's so hard to get recent big core SBC's), in those circumstances area would be the main consideration as the SoC itself is the product to be sold.

For Apple the iDevice is the product, with the SoC being only part of it, and the bragging rights of highest performance (pointless but I digress) being their goal - so in those circumstances I can imagine they would sacrifice a sight more than 10mm2 to maintain said bragging rights.
 

name99

Senior member
Sep 11, 2010
449
333
136
It was not intended as a trolling exercise, but rather a basic maxim of die area economics, no more no less.

I did not call anyone an idiot, you asked a question to something I believed to have an obvious answer (hence the "silly question" remark) - I am astounded at how often I have experienced reasonably intelligent people miss obvious things right in front of them, I'm sure there's probably a psychology paper or 10 on the subject.

Value is a subjective term relative to the market space the product is aimed at, and what the product itself is.

There are plenty of SoC's that do not remotely aim for maximum performance - hence why so many SoC's have only A53 or A55 cores (why it's so hard to get recent big core SBC's), in those circumstances area would be the main consideration as the SoC itself is the product to be sold.

For Apple the iDevice is the product, with the SoC being only part of it, and the bragging rights of highest performance (pointless but I digress) being their goal - so in those circumstances I can imagine they would sacrifice a sight more than 10mm2 to maintain said bragging rights.

The A76 is not one of those devices ("that do not remotely aim for maximum performance")...

I'm saying something very simple:

- that ARM made a tactical mistake in prioritizing the size of the A76 to the extent they did. They created a core that is, yes, tiny, but is also less performant than many of their customers would like.
(Of course that doesn't mean ARM would have got the same level of performance as Apple just by using the same area --- but they would certainly get better than what they have today.)

- that the extra area cost of Apple's cores is not THAT large compared to A76, when you consider the entire context of the SoC

Now neither you nor I know enough about ARM internals (financials, what customers told them 5 years ago when design started as opposed to now, what TSMC+Samsung expected of their processes) to be able to say that ARM was WRONG in their decision;
all we can say is that it turned out the way it did, with ARM still FAR behind all these years after the A6 and A7. And that there's something to be said for having ambitious goals rather than narrowly optimizing for immediate financial return.

If you have a complaint with the above analysis, make that complaint. But don't extract random quotations out of context ("What are you saving the area FOR?") and then imagine that you are contributing value to the conversation by telling us all that, oh wow, area costs money. THAT is what I object to -- being treated like an idiot.
Sure, sure, I've never actually designed a CPU from scratch. But I did have some input (on behalf of Apple) into the team that designed Altivec (and was willing to devote 25% of the G4 die to it) so I do have some vague awareness of the tradeoffs between area and consequent performance.
 
Reactions: Richie Rich

name99

Senior member
Sep 11, 2010
449
333
136
It's based on Darwin which is BSD derivative, and way back in the lineage of MacOS at that - what code it shares with FreeBSD is likely to have been replaced or overhauled at least once in the years since Darwin was branched/forked from BSD - unless of course Apple decided to lessen their burdens and share some common code directly from the stable FreeBSD code base..

BSD is partially interchangeable with Linux from what I have seen due to their common POSIX roots, certainly programs written for Linux can be run on BSD variants with a not so huge amount of effort.

Linux Firefox was available on PC BSD in mostly up to date form when last I tried it before it became True OS (admittedly a looong time ago now).

If you're interested in how AppleOS works (REALLY works) as an Operating System, the info about this is available:

The easiest book to read (best place to start, though somewhat out of date) is Amit Singh's book

If you want free (somewhat newer, more difficult to read) there is Jonathan Levin's first attempt:

If you want up to date (as of ~mid-2019) but somewhat expensive. (Three volumes, $75 each. Absolutely worth it, if you actually care about Apple technology and want to be fully informed) there is Jonathan Levin's second attempt:
 

beginner99

Diamond Member
Jun 2, 2009
5,223
1,598
136
"Are you happy to save 10mm^2 (so 10% of the SoC area) at the cost of losing 30 to 50% of your CPU performance? What's the point?"

Citation needed that you loose 30-50% performance.

With a simple die calculator and assuming bigger SOC (Apple) is 10x10mm and the smaller one 10% smaller meaning 10x9mm on a 300mm wafer with defect rate of 0.1/cm2 you get 523 good big dies and 597 good small dies. That's 14% more dies per wafer. And this only increases if yield gets worse. So "small die" is also a risk lowering strategy.

If you go big die, you need to sell you chip for 14% more to make it worth it. Seems like a small number but in times were double digit annual growth is a huge success, it is a lot.

You also assume people buy based on performance. They don't. Especially not when it comes to iOS vs Android. People bought iPhones even when they were comparable to Android in terms of performance. Actually iPhone sales have been stagnating since Apple has this large advanatge while Android sales kept growing. Of course this says nothing about margins and maybe stagnation at high margins is apples goal (AFAIk they want rich, 1st world buyers because these people actually buy stuff from Apple store, the real "margin cow".). So the chip design itself my be a reflection of the companies goal (Apple=high margin "luxury" devices) while ARM designs for the masses (cheaper, less risk) and not about performance.
 

Nothingness

Platinum Member
Jul 3, 2013
2,757
1,405
136
I worked in QuickTime from about 1992 to about 2002. I started off doing the MPEG code, moved to the codec guts of the DVD player, and then became the sort of general codec optimization guy working to speed up everything from Sorenson to JPEG2000 to AAC.

I did not experience any sort of paranoia or feel like things I needed to know about were being held from me. But I suspect a lot of this is cultural/personality. Some people feel that they have the right to know everything about everything, from the salary of everyone in the company to the strategy ten years out to M&A plans. Other people are quite willing to accept that some degree of secrecy (in personal life, in politics, and inside a company) is inevitable. I don't think the company makes much difference -- if you demand total transparency you're going to be unhappy anywhere. If you demand the right to blab company secrets to anyone, you're going to be unhappy anywhere. If you feel the you have the right to make political rants about your colleagues or management, don't be surprised when they decide they're not interested in hearing yet another of your rants.

I knew what I needed to know to do my job. I was as surprised as anyone else by new Apple stuff that wasn't relevant to my job --- new models, new OS features, iPods, ...

Most of the things you are talking about are issues about which reasonable people come down in different places for different topics and different circumstances. If you are willing to be a reasonable person, and to accept that other people --- perfectly DECENT REASONABLE people -- have different opinions, then act like an adult, find a job you love, and accept that it's a job, it's not a marriage, where you expect every single detail to be shared with you, and agreement on every single aspect of how life is to be lived.


I would say one final thing. The happiest people are people who are honest with themselves. They know what they are good and bad at. They know what they want from life and from a romantic partner. They know what they want from a work environment.
Then they choose and optimize based on that.

That all may seem obvious. But in fact 95% of the people I know don't honestly know themselves. They started projecting a persona sometime around 8 years old, and have projected it so long they can no longer tell that's it's not their true self. This is especially obvious when it comes to socially contentious matters, where they have spent so much time stating more and more extreme political claims that, in fact, they kinda just slid into. But it seems like now they're stuck in this loud persona that just isn't them. And of course it's even worse if you have 20yrs of public social media statements along these lines...

So that's my point. What I hear (maybe I'm wrong, but this is what I hear in your comment) is someone who has been swept up in various aspects of techno-hysteria over the past few years. For some it became privacy hysteria. For some it became walled-garden hysteria. For some it becomes work environment hysteria. But in all cases close to zero knowledge coupled with massive echo-chamber feedback led people to lay down lines that (honestly) they just don't care about. But now it feels like they've told everyone that Apple (or Google or Samsung or Microsoft or Intel) is the devil because [...] and they can't walk that back to go work there.

So that's the issue:
- Do you know your real self as opposed to the persona you've constructed?
- Are you brave enough to destroy that persona and create one that better matches who you now are?
- Are you smart enough to ignore the next wave of hysteria (and the next, and the one after that) to admit "I don't need to have an opinion on that"?
- Are you willing to accept that not everything has to be viewed as a political battle to the death, to accept that some people, people smarter than you, even people more ethical than you, may have looked at the same facts you have, and concluded something different? And still be willing to treat those people with respect?
- Is your response to any disagreement going to be "can we figure out a solution that works for both of us"? Or is it going to be "someone is TRAMPLING ON MY RIGHTS. I DEMAND SATISFACTION!"

Almost none of this has to do with Apple. But this is what life in large organizations is like. If that is not the sort of person you are, stick to small organizations.
Huh what a silly and stupid rant. Where did I talk about political views or personal life? If I'm sitting next to an engineer I'm likely working on the same subject as the guy, so I expect to be able to discuss with him about technical subjects related to our daily job.

You left Apple almost 20 years ago, are you the kind of guy who thinks that things don't change? In particular in a company that went to where Apple is now compared to where they were at the start of the century? You always go ballistic when someone says something bad about Apple. You're the one who has built a persona trapped in Jobs Reality Distortion Field. Are you brave enough to leave that behind and accept that people can like your beloved company but can also say bad things about it?

I'm not subject to any kind of hysteria, in particular not your Apple hysteria. I also have been working in a large company (8k people and still counting) for 15 years and I'm perfectly happy, and my management also is quite happy with my job.

No really your psychology skills need some more work

EDIT: BTW I'm surprised you didn't call me "boomer" as you insulted someone on RWT.
 
Last edited:

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
all we can say is that it turned out the way it did, with ARM still FAR behind all these years after the A6 and A7
Far behind on raw performance yes, but their performance per watt is comparable despite that, as both I and Nothingness have been over before.

Perf/watt/area clearly seems to be (or at least was) ARM's main focus in core design at the point of A76's inception - and given Apple's manifest woes in battery life that doesn't seem like any kind of tactical mistake to me.

Given the significant changes in A77 we can already see ARM's strategy changing, so we won't have to wait too long to see a core comparable in width from them if Andrei's info about Matterhorn turns out correct, though the lack of a v9 announcement this year makes me skeptical about Matterhorn being v9 rather than v8.6 (which has the MatMul instructions intended for Matterhorn as per the TechCon keynote).
 
Reactions: Nothingness

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
So the chip design itself my be a reflection of the companies goal (Apple=high margin "luxury" devices) while ARM designs for the masses (cheaper, less risk) and not about performance.
That is relative too, Android phones used to be the cheaper option, but now you have to turn to Chinese manufacturers like Huawei to get a sane price on a high end Android phone.

In my country they can get up to £800 with the standard quantity of flash memory, with yet higher prices as you add more, no doubt why they periodically seem to lose microSD ports.
 

Nothingness

Platinum Member
Jul 3, 2013
2,757
1,405
136
The A76 is not one of those devices ("that do not remotely aim for maximum performance")...

I'm saying something very simple:

- that ARM made a tactical mistake in prioritizing the size of the A76 to the extent they did.
That tactical mistake has allowed A76 to be used in mid-range chips where the volume is much higher.

They created a core that is, yes, tiny, but is also less performant than many of their customers would like.
Agreed, but that would not have sold as well.

(Of course that doesn't mean ARM would have got the same level of performance as Apple just by using the same area --- but they would certainly get better than what they have today.)
I also doubt they could have reached Apple level.

- that the extra area cost of Apple's cores is not THAT large compared to A76, when you consider the entire context of the SoC
In the high-end that's right, but for middle-end and below a large CPU would have a disproportionate impact.

Now neither you nor I know enough about ARM internals (financials, what customers told them 5 years ago when design started as opposed to now, what TSMC+Samsung expected of their processes) to be able to say that ARM was WRONG in their decision;
all we can say is that it turned out the way it did, with ARM still FAR behind all these years after the A6 and A7. And that there's something to be said for having ambitious goals rather than narrowly optimizing for immediate financial return.
You indeed definitely have no clue what most ARM customers want and how ARM works. Your Apple way of thinking prevents you from understanding that smaller companies have different constraints and needs.
 

Richie Rich

Senior member
Jul 28, 2019
470
229
76
Nothingness, soresu, beginner99:

Name99 is right. If A76 is so winning design concept (saving area and cost), so why would ARM developed much wider A77 and even wider A78 and Matterhorn? It's simple, ARM desperately tries to close the gap to Apple. And because ARM is approximately 5 years behind Apple in terms of CPU development, ARM needs to develop advanced micro-architecture first and only then spend it in transistors. In other words A76 is small because ARM wasn't able to develop something larger and powerful at that time.
 

Andrei.

Senior member
Jan 26, 2015
316
386
136
Nothingness, soresu, beginner99:

Name99 is right. If A76 is so winning design concept (saving area and cost), so why would ARM developed much wider A77 and even wider A78 and Matterhorn? It's simple, ARM desperately tries to close the gap to Apple. And because ARM is approximately 5 years behind Apple in terms of CPU development, ARM needs to develop advanced micro-architecture first and only then spend it in transistors. In other words A76 is small because ARM wasn't able to develop something larger and powerful at that time.
Your argument is flawed.

First of all, the A77 is only marginally larger than the A76. PPA on the A77 is better than on the A76.

Secondly, what makes you think that Matterhorn will be the single "big" CPU offering alongside their small cores instead of it simply being the new "giant" CPU offering alongside big and little microarchitectures?

Arm not doing a giant core until now was tied to economics and their customer needs instead of their (un)willingess to actually create one.
 

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
It's simple, ARM desperately tries to close the gap to Apple.
Define desperation?

ARM is not impeded by Apple's progress because Apple ceased to license Cortex cores from A6 onwards, they are not directly competing against Apple at all when you stop and think about it from a purely business perspective.

Only some of ARM's licensees are directly competing against Apple - ie Samsung and Huawei, but even then not on the same software platforms, so again it's not an apples to apples competition going on (pun intended).

In my opinion, ARM is going wider because of the demands of x86 binary translation for WARM have increased the need for wider cores beyond what most Android or Chrome platforms and apps demand - while until now there has not been a drastic need for wider cores on Android platforms since they could run them at a speed that all but eliminates UI jank, and very few Android applications strain the compute capabilties of current big cores (to my continuing displeasure in the slow paced development of ARM based console emulator backends/dynarecs).

I also believe the change of hands to SoftBank may have both changed directions and increased R&D capital for ARM, which after all does not have a very great amount of capital to work with compared to Apple when it comes to development - their license royalties are a constant income, but not nearly as lucrative as the insane markup on Apple products.

The SoftBank angle makes sense considering that Matterhorn is probably being announced just shy of 5 years after ARM was acquired by them, ample time to design a new uArch....

The AAA game market for ARM is mostly on the Switch, with very few translating to either iOS or Android.

The Switch itself runs on the lowly A57, now 4 generations behind A76 and 5 behind A77, so all that width and IPC in Axx is currently wasted on mostly trivial apps until they throw something demanding like path tracing or video encoding at it, which would probably come with MacOS - I say mostly, the one truly impressive real world use of them I have seen was dav1d decode speed vs Snapdragon, though I have yet to see figures for SD 855 to compare against.

As for A78, we have basically no info on it so far - so it is pointless to say it is getting wider than A77 until next May (late May seems to be the annual launch since A73 I think) - the only thing that seems a given is the promised 15% YoY improvement outlined in 2018.

All in I'd say both Apple and Google missed a trick when nVidia was pursuing ARM game ports for the SHIELD, it was certainly as good a driver as any for generational improvements to mobile performance.
 

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
First of all, the A77 is only marginally larger than the A76. PPA on the A77 is better than on the A76.
Ah interesting, I had gained an impression to the opposite effect of that - methinks ARM could do with being a bit more open about these things, they have an annoying tendency to compare figures from much older cores with new core PR, it leaves me more than a little confused about what they are saying most of the time.
 

soresu

Platinum Member
Dec 19, 2014
2,968
2,192
136
Secondly, what makes you think that Matterhorn will be the single "big" CPU offering alongside their small cores instead of it simply being the new "giant" CPU offering alongside big and little microarchitectures?

Arm not doing a giant core until now was tied to economics and their customer needs instead of their (un)willingess to actually create one.
This gels with their earlier A35 and A32 efforts during the A53 era, though they did not go very far with that.

As you say economics of a small company like ARM is an issue when considering design choices, and SoftBank's capital will give them options to offer a wider selection of performance across the spectrum of the market space.
 

Andrei.

Senior member
Jan 26, 2015
316
386
136
This gels with their earlier A35 and A32 efforts during the A53 era, though they did not go very far with that.

As you say economics of a small company like ARM is an issue when considering design choices, and SoftBank's capital will give them options to offer a wider selection of performance across the spectrum of the market space.
I have no doubt SoftBank is one of the reasons for Matterhorn. SoftBank invested an idiotic amount of money to buy Arm and their end goal is to actually have the company go public again in a few years and recoup their investment. The main growth opportunity here is the infrastructure and HPC market and that's where a giant core like Matterhorn comes into play.
 
Reactions: soresu
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |