Solved! ARM Apple High-End CPU - Intel replacement

[Richie Rich]

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:

 

[Eug]

Isn't that the 10c Comet Lake part?

Yes, semi-custom for Apple, if it proves to be true.

BTW, the 2018 iPad Pro is as fast in some tests as the 2017 Core i7-7700K I had briefly, before I returned it to get something quieter.

 

[DrMrLordX]

Yes, semi-custom for Apple, if it proves to be true.

Okay, thought so (we've been over that already). So Apple isn't running a 95W in-house CPU in that role. Yet!

 

[blckgrffn]

Nah, now that Jony Ivy is no longer ruling Apple's designs with an iron fist they won't sacrifice battery to get another mm of thinness like they always did. Notice how they went back from that disaster of a keyboard after he left?

Just look at the iPhone 11 line, those have almost the best battery life of any smartphones out there, including ones with far larger batteries. If Apple wanted to be "decent but nowhere near top of the heap" like they were in smartphone battery life comparisons for many years they could have shaved a bit of thickness and weight off the iPhone 11. Ivy would have gone in a more svelte direction if he was still in charge.

I think they will have more battery life, but can't say for sure they won't shrink that battery somewhat as well. After all, there's a point where additional battery life in a laptop is pretty meaningless. If you can get 20 hours for instance, who is going to think it is worthwhile getting a 21st hour? That's sort of like how I view phones too, beyond "all day" battery life it is pointless because humans have to sleep so there is always a time to charge it.

¯\_(ツ)_/¯

What does it take to dominate a laptop chart right now? looks like anything in the 12-14 hour range is really compelling and arguably "all day". 14-16 would be outstanding (but to your point about all day usage, necessary? and how for how many people?). They could likely still massively reduce the battery and hit those targets depending on far they stretch the CPU/GPU TDP. I don't expect them to sacrifice anything for the sake of getting thinner. Anything that could be sacrificed basically already has been.

My Lenovo P52s lasts ~10+ hours and rarely have I found myself wishing for more. How awesome that is compared to the old 2-4 hours...

I found this to reference, not vouching for the site in any way:

Laptops with best battery life in 2021 - longest lasting laptop batteries

The laptops with best battery life will outlast work or school

www.laptopmag.com

 

[lobz]

He's enough of a fan of it to say that it's the single best test of over-all int perf. Yeah, it's not a great proxy for a lot of UI-intensive loads, which will indeed hit L1I harder (especially games), but that's actually the opposite of what beginner99 was saying - beginner99 was arguing that Apple's cores will be great for consumer loads but somehow atrocious for compiling. Which is a statement utterly without evidence.

You are literally putting words in his mouth, and rather disingenuously at that.

 

[Doug S]

You are literally putting words in his mouth, and rather disingenuously at that.

In whose mouth, Linus'? If you follow RWT you would know she's not putting words in his mouth, she's repeating the exact same words he's written on many an occasion.

 

[Eug]

All I remember is that Linus HATED Geekbench 3, but he wasn't too negative about Geekbench 4.

Real World Technologies - Forums - Thread: Geekbench 4

www.realworldtech.com

Anyhow, is this Geekbench 5 benchmark for A14 fake?

Sina Visitor System

www.weibo.com

My guess is it's fake, but I can't read the page contents.

However, if it turns out to be at least close to ballpark speeds, this would be more than sufficient for an entry level Mac laptop.

 

[defferoo]

It should be noted that the MacBook Air with 13" Retina screen, 8 GB RAM, and 256 GB storage is just $999.

The iPad Pro with 12.9" Retina screen, 6 GB RAM, 256 GB storage, and Magic Keyboard is $1448.

IOW, to get a similar setup with the iPad Pro, you have to spend 45% more than what it would cost for the MacBook Air. Even if you remove the Magic Keyboard, it's still priced higher than the MacBook Air, by $100.

Yeah, it’s a complicated comparison because the pricing model is so different between the two product lines. For example, I was referring to the base 11” iPad Pro which is $799 USD versus the $999 MacBook Air (granted, the $999 12.9 Pro might be a better comparison given the screen size). To your point, the iPad Pro is a premium product and the Air is considered Apple’s budget notebook, so they might skimp on processing power in the Air.

That being said, we know the iPad 4+4 SOC works in a fanless environment and the Air already has a fan built in. I’d imagine they could take the iPad Pro SOC and directly implant it in the Air and not have any thermal issues so I’m not sure why they would take a step down and put the equivalent of a phone chip in the Air even if it is faster than the current Air. I feel like they would definitely get some flak for that.

 

[Eug]

That being said, we know the iPad 4+4 SOC works in a fanless environment and the Air already has a fan built in. I’d imagine they could take the iPad Pro SOC and directly implant it in the Air and not have any thermal issues so I’m not sure why they would take a step down and put the equivalent of a phone chip in the Air even if it is faster than the current Air. I feel like they would definitely get some flak for that.

Cuz there are lot of complaints about the fan noise on the current MacBook Air.

I love my fanless 2017 MacBook 12" Core m3-7Y32, but the performance of A14 (non-X) will blow it out of the water. Hell, even A12 (non-X) blows it out of the water.

Currently the top end MacBook Air runs a Core i7-1060NG7, and people complain about its fan noise. Yet this chip is already bested by A13 in synthetic and real world benchmarks, and A14 will do even better.

Basically I see no good reason to put A14X in a MacBook Air (or revived 12" MacBook), aside from marketing, as you suggest. (I wonder how they would spin putting a 5.4" iPhone chip in a 13" computer.)

What I might expect is that they would stick with A14X in most of the MacBook Pros, but would consider designing it for lower thermals. Either that or up the clock rates, etc.

Plus, these chips aren't super small. A12X is is 122 mm2. A12 is only 83 mm2, so much cheaper.

 

[ThatBuzzkiller]

Now that's gonna stir up some heated conversation here with the ARMada 😂

Seriously it gets annoying seeing all the evangelism for ARM based PCs ...

At the end of the day ARM based systems are going to have to pay the overhead of emulation so any gains that they had in efficiency from ISA design are more than going to be negated like we see with the Surface Pro X.

High-end designs for x86 cores are going to be inevitably more efficient for native x86 programs and no amount of hardware engineering will change this even from Apple. ARM CPUs need to maintain a higher clock to have comparable performance against lower clocked x86 CPUs when emulating/running native x86 applications so if it comes down to it both AMD or Intel have the option of down clocking their CPUs to obtain similar or better power efficiency ...

 

[Eug]

Seriously it gets annoying seeing all the evangelism for ARM based PCs ...

At the end of the day ARM based systems are going to have to pay the overhead of emulation so any gains that they had in efficiency from ISA design are more than going to be negated like we see with the Surface Pro X.

High-end designs for x86 cores are going to be inevitably more efficient for native x86 programs and no amount of hardware engineering will change this even from Apple. ARM CPUs need to maintain a higher clock to have comparable performance against lower clocked x86 CPUs when emulating/running native x86 applications so if it comes down to it both AMD or Intel have the option of down clocking their CPUs to obtain similar or better power efficiency ...

Nah. Apple consumer apps are already ARM native. Apple pro apps are also natively running in the lab on ARM but need more work. MS Office and Adobe apps are also already running natively on macOS ARM in the lab. (I suspect Apple paid MS and Adobe $$$$ to get these up and running quick.) iPadOS apps can be ported over now and run natively as well.

On the flip side, the next Mac releases are actually Intel based, to ride out the transition, which I believe will be complete from a software perspective in three years, esp. considering ARM Mac dev kits already in the wild.

This works well. Pro houses with macOS x86 software needs buy Intel in 2020 and 2021, with their usual 3-year service plans, and then go ARM at the next major purchase cycle after the software is ready.

 

[ThatBuzzkiller]

Nah. Apple consumer apps are already ARM native. Apple pro apps are also natively running in the lab on ARM but need more work. MS Office and Adobe apps are also already running natively on macOS ARM in the lab. (I suspect Apple paid MS and Adobe $$$$ to get these up and running quick.) iPadOS apps can be ported over now and run natively as well.

On the flip side, the next Mac releases are actually Intel based, to ride out the transition, which I believe will be complete from a software perspective in three years, esp. considering ARM Mac dev kits already in the wild.

This works well. Pro houses with macOS x86 software needs buy Intel in 2020 and 2021, with their usual 3-year service plans, and then go ARM at the next major purchase cycle after the software is ready.

Apple's in-house "pro apps" are for amateurs and their other development tools are a sick joke as well ...

Visual Studio > Xcode Chromium > WebKit Adobe Premiere Pro > Final Cut Pro X Vulkan > Metal Windows > macOS

I know tons of developers who are going to ditch macOS for Windows because at least Microsoft doesn't promote a toxic development culture of high maintenance and constant refactoring of low-level code. It's extremely naive to believe that professional users would migrate within 3 years when many professional developers behind these complex applications work on much larger time scales of either 5 or 10 years ...

Buying software support like Apple does isn't sustainable and macOS will crash and burn in the long-term without any organic developer community growth. Apple can try to replace Intel based hardware as much as they want but they'll sooner come to the reality that Intel based software can't be replaced as Microsoft discovered with the Surface Pro X ...

 

[Doug S]

Sina Visitor System

www.weibo.com

[/URL]

My guess is it's fake, but I can't read the page contents.

However, if it turns out to be at least close to ballpark speeds, this would be more than sufficient for an entry level Mac laptop.

There's no way anyone can do anything but guess since it is very easy to fake, but the numbers are in the believable range. There's also the possibility of "its real but that's not the clock rate the production iPhone ends up with".

I believe we may have seen that latter once before a few years ago where numbers came out in April/May timeframe that were higher than the real ones ultimately were, which in hindsight were likely real because the iPhone shipped at a lower clock rate. Since Apple doesn't bin, they have to choose a clock rate that will allow all or nearly all the working SoCs to qualify. They can't choose a higher clock rate that 20% won't qualify at, because they don't have anywhere to put tens of millions of SoCs that "too slow" for the iPhone.

The transition from N7 to N5 is supposed to buy a 15% clock rate increase at the same power, the above increase would be about 19%. That's possible as some rework to critical paths can goose the clock rate a bit as well, but it is pretty optimistic. Right at or just under 3 GHz seems more likely IMHO.

 

[Doug S]

Plus, these chips aren't super small. A12X is is 122 mm2. A12 is only 83 mm2, so much cheaper.

Do you realize how little difference that makes? You're talking about a marginal cost difference of maybe $10 here. I think Apple can afford that if they feel it makes a $1000 product better.

Though it sure sounds like they won't be using the A14 or A14X, but something specifically designed for the Mac. How that differs from the A14X we'll have to see, but I expect it will be a lot closer to that than the A14.

 

[soresu]

Do you realize how little difference that makes? You're talking about a marginal cost difference of maybe $10 here. I think Apple can afford that if they feel it makes a $1000 product better.

Don't underestimate the penny pinching instincts of a CEO.

Apple's entire brand (ARM core aside) has been about selling less for more.

It's part of what makes me so melancholy at Android manufacturers abandoning the added value features like external memory that made Android phones the better buy years ago, I mean apart from a UI that doesn't make me want to simultaneously vomit and commit homicide.

It's also not just about how much each SoC costs, it's also about how many you get per wafer, especially after binning for yield.

 

[defferoo]

Apple's in-house "pro apps" are for amateurs and their other development tools are a sick joke as well ...

Visual Studio > Xcode Chromium > WebKit Adobe Premiere Pro > Final Cut Pro X Vulkan > Metal Windows > macOS

I know tons of developers who are going to ditch macOS for Windows because at least Microsoft doesn't promote a toxic development culture of high maintenance and constant refactoring of low-level code. It's extremely naive to believe that professional users would migrate within 3 years when many professional developers behind these complex applications work on much larger time scales of either 5 or 10 years ...

Buying software support like Apple does isn't sustainable and macOS will crash and burn in the long-term without any organic developer community growth. Apple can try to replace Intel based hardware as much as they want but they'll sooner come to the reality that Intel based software can't be replaced as Microsoft discovered with the Surface Pro X ...

haha what is this post? just a bunch of garbage. if you don’t like Apple software, there are plenty of alternatives, nobody is forcing you to use it. for the vast majority of apps, Rosetta will be good enough during the transition and should give them plenty of time to re-compile for the new architecture.

 

[Eug]

Do you realize how little difference that makes? You're talking about a marginal cost difference of maybe $10 here. I think Apple can afford that if they feel it makes a $1000 product better.

Though it sure sounds like they won't be using the A14 or A14X, but something specifically designed for the Mac. How that differs from the A14X we'll have to see, but I expect it will be a lot closer to that than the A14.

What? $10 is huge. That might even approach 100 million dollars per year, considering they sell almost 20 million Macs a year, with most being laptops, and more lower end than higher laptops being sold.

If it were really as high as $10 per unit, that would be all the more reason they would do it. Even half of that $100 million would be justification enough.

 

[itsmydamnation]

Fine you can continue with your misplaced doubt about Apple's ARM chips until the new Macs are released in six months and you're forced to eat your words.

An Apple chip being good , has nothing to do with a benchmark being crap.

The simple fact that geekbench is so light my renior laptop doesn't boost ( sits at 2.4ghz) yet all the other benchmarks i tired and even my own C and Perl get all 8 cores @ 4.2ghz should tell you something.

 

[RasCas99]

Reading this thread for a long time , still no converters to either side , some folks here even going out of their way to invent crazy propositions that fit their agenda without any basis to stand on , when you read your own post and all it has is "but what if , and imagine that, consider that maybe if " instead of relying on facts that are in front of you , its time to admit your bias or even worse bringing agendas and personal dislike to a company into a technical discussion (happens a LOT around here for Intel/AMD).

It will be great to see the reactions when we have a real system in our hands , my bet here is that ppl will STILL be entrenched in the same opinions they have now.
If reviews/extensive benchmarks say Apple are behind , lets see the Armada say they were wrong !
If the reviews/extensive benchmarks say Apple hit it out of the park , lets see the Nay crowd say they were wrong !

No need to feel bad about being wrong , it happens when you have to pick sides with limited information , but once everything is out in the wild there is no excuse for not changing your stance and brace the reality whatever it might be.

"set reminder for the end of the year".

 

[Doug S]

Reading this thread for a long time , still no converters to either side , some folks here even going out of their way to invent crazy propositions that fit their agenda without any basis to stand on , when you read your own post and all it has is "but what if , and imagine that, consider that maybe if " instead of relying on facts that are in front of you , its time to admit your bias or even worse bringing agendas and personal dislike to a company into a technical discussion (happens a LOT around here for Intel/AMD).

It will be great to see the reactions when we have a real system in our hands , my bet here is that ppl will STILL be entrenched in the same opinions they have now.
If reviews/extensive benchmarks say Apple are behind , lets see the Armada say they were wrong !
If the reviews/extensive benchmarks say Apple hit it out of the park , lets see the Nay crowd say they were wrong !

No need to feel bad about being wrong , it happens when you have to pick sides with limited information , but once everything is out in the wild there is no excuse for not changing your stance and brace the reality whatever it might be.

"set reminder for the end of the year".

You're assuming the outcome will be black and white. You still see plenty of AMD vs Intel argument, because while AMD's Zen is better than Intel on some fronts, they remain behind on others. There's no chance Apple's ARM Macs will beat x86 in every metric, nor any chance they will be behind in every metric, so those with an axe to grind will move the goalposts and claim that those outlier metrics are the ones that matter.

 

[soresu]

Rosetta will be good enough during the transition and should give them plenty of time to re-compile for the new architecture.

Dear lawd, please stop with this perpetual re compile nonsense, C code can be re compiled to ARM with no great effort assuming that extra API's are not necessary to address, but hand written, optimised assembly is a different kettle of fish.

Big, old code bases like Maya, 3DS Max, Avid etc will have a ton of hand tuned x86 SIMD assembly code to replace, and probably a lot more now than there was at the time of the PPC -> x86 transition due to more features added in the last 15 years.

Those ancient behemoths bloat in size over time for a reason, and it's not just lazy coders.

The apps that have been pursuing mobile applications for awhile like Photoshop will likely transition much faster - or in the case of the likes of Affinity Photo already have both x86 and ARM code bases to draw from.

Hopefully in time new ML based compilers will be able to take C/Rust etc and churn out SIMD assembly that can match or exceed human written code - I have seen academic papers that address this subject already, so maybe not too far down the road if we are lucky, then perhaps automatic gfx API conversion can follow in time to migrate old gfx code to modern API's like Vulkan.

 

[ThatBuzzkiller]

haha what is this post? just a bunch of garbage. if you don’t like Apple software, there are plenty of alternatives, nobody is forcing you to use it. for the vast majority of apps, Rosetta will be good enough during the transition and should give them plenty of time to re-compile for the new architecture.

Sadly for you that's not the case in the real world ...

How am I going to offer my apps in the App Store if I don't use a trash IDE like Xcode ? What alternative is there to macOS since Catalina was a huge disaster ? What are going to happen to those alternatives if they aren't in the App Store ?

"Plenty of alternatives" is a bull crap statement when Apple has proven plenty of times that they'll take them away in order to pursue their anti-developer walled garden ...

Rosetta being "good enough" for the transition ? I think it's irrelevant whether they offer a compatibility layer or not. It made no difference when Apple was transitioning from PowerPC to Intel because it was largely down to the industry converging to Intel that macs still managed to survive. Are you naive enough to believe that the same scenario will unfold for ARM architecture despite a number of big corporations attempting this ?

Google tried to make the industry converge to ARM but they failed since AMD/Intel are still alive and their Stadia cloud gaming platform uses x86 CPUs too for added irony. Microsoft has been working with Windows on ARM for many years but there are still no signs of increased OEM uptake. Seeing as how Apple is now a hardware company rather than a software company since they can't maintain a platform market share of above 10% if their life depended upon it, they'll fail proportionally 10x harder than Google or Microsoft did ... (we'll soon see iOS at sub-10% soon enough since Apple seems to like buggy software so much)

 

[defferoo]

Dear lawd, please stop with this perpetual re compile nonsense, C code can be re compiled to ARM with no great effort assuming that extra API's are not necessary to address, but hand written, optimised assembly is a different kettle of fish.

my bad, re-write and re-compile. happy now? either way, most apps don’t need a ton of rewriting. they realistically have 2-3 years, the only hard deadline they have is when Rosetta 2 is no longer supported. They’ll have an incentive to make their apps run well on new Macs because they have customers who want to continue using them.

 

[defferoo]

Sadly for you that's not the case in the real world ...

How am I going to offer my apps in the App Store if I don't use a trash IDE like Xcode ? What alternative is there to macOS since Catalina was a huge disaster ? What are going to happen to those alternatives if they aren't in the App Store ?

"Plenty of alternatives" is a bull crap statement when Apple has proven plenty of times that they'll take them away in order to pursue their anti-developer walled garden ...

Rosetta being "good enough" for the transition ? I think it's irrelevant whether they offer a compatibility layer or not. It made no difference when Apple was transitioning from PowerPC to Intel because it was largely down to the industry converging to Intel that macs still managed to survive. Are you naive enough to believe that the same scenario will unfold for ARM architecture despite a number of big corporations attempting this ?

Google tried to make the industry converge to ARM but they failed since AMD/Intel are still alive and their Stadia cloud gaming platform uses x86 CPUs too for added irony. Microsoft has been working with Windows on ARM for many years but there are still no signs of increased OEM uptake. Seeing as how Apple is now a hardware company rather than a software company since they can't maintain a platform market share of above 10% if their life depended upon it, they'll fail proportionally 10x harder than Google or Microsoft did ... (we'll soon see iOS at sub-10% soon enough since Apple seems to like buggy software so much)

what a load. you don’t have to release your app on the iOS app store, you choose to because the market dictates that it’s good business sense. I don’t like to do tedious things, but if I get paid then it doesn’t suck as much. Let’s be honest here, if you dislike Xcode, you’ve probably never liked it, so how exactly does this transition affect you? You can continue using an Intel Mac to build your app as long as it can run the latest macOS which apparently is so bad that you could not possibly use it, yet millions of users continue to use it to run their businesses and as their personal machine.

The difference here is that all the other attempts at using ARM CPUs in a computer we’re terrible and half-hearted. Nobody besides Apple has the capability or the will to make a decent ARM chip for a laptop or desktop. You call Windows on ARM an attempt? They basically said here’s a Windows machine, but you can’t run anything you‘re used to running, also it’s kind of slow, but battery life! Why would anybody buy that? If you can’t tell the difference between that and what Apple is doing, then this conversation is over.

 

[DrMrLordX]

The simple fact that geekbench is so light my renior laptop doesn't boost ( sits at 2.4ghz) yet all the other benchmarks i tired and even my own C and Perl get all 8 cores @ 4.2ghz should tell you something.

I know, right? It just doesn't make my chip work very hard. Definitely not in MT.

 

[soresu]

my bad, re-write and re-compile. happy now? either way, most apps don’t need a ton of rewriting. they realistically have 2-3 years, the only hard deadline they have is when Rosetta 2 is no longer supported. They’ll have an incentive to make their apps run well on new Macs because they have customers who want to continue using them.

That 'incentive' part is a sticky proposition at best - far more stick than carrot, if any carrot exists there at all for the application developers.

A real incentive would be Apple actually paying developers handsomely to do their dirty work of supporting their new platform, rather than a threat of possible loss of users/customers because Apple felt like ditching x86.

If you follow that logic to its conclusion, those same customers will need to buy new ARM Macs if x86 support will be entirely ended in the future.

While some would likely follow the Apple party line, there is no small chance that some would stop and take stock in such a situation, especially if it involves a company wide change of dozens to hundreds of computers - from there some might consider moving to a completely different OS platform entirely, so long as the same software exists there.

Which it does for many of the big application suites used on Mac's like Autodesk, Adobe CC and Avid.

Apple like to have their own way, but they are playing with fire here, and they may well get burned if they do not take the trouble (and the cash) to sweeten the deal alot more than just loaning out some last gen based SoC devkits during the Rosetta changeover period.