iPhone 14 Pro and Redesigned MacBook Air Reportedly Stuck Using Technology Behind A15 Chip

[brofkand]

Actually, let's recall some of the rationalizations we slung to each other when Apple announced going from Intel to Silicon:

• Without having to wait for Intel to deliver new chips, Apple will be able to update Macs much quicker than only annually or so
• Without having to pay the fat Intel premium, new Macs can be more powerful and cost less.
• Without having to deal with Intel incompetencies in upgrading chips, Apple will be able to deliver huge power upgrades year after year.

...and of course...

• Updating Intel Mac software will be just a flip of a switch in the compiler

Now certainly, nobody could anticipate Covid effects and global responses back then, so none of that is really taking a hard poke at Apple (sans maybe bullet 2) but still.

I suspect the bulk of "us" certainly did expect a more robust upgrade schedule and big power jumps generation to generation.

Those around for the last transition knew that bullet 4 was an outright lie some of us were spinning to the rest... else that must be one super-heavy switch to throw, given we're almost 2 years in now and there's still plenty of non-native Mac Apps.

And I never doubted where the "savings" of that Intel premium would go. ???

I own M1 Ultra and glad that I do but it seems we should be on M2- and anticipating M3 soon- by now... and not by painting a 2 in place of a 1 on mostly the same chip technology. It appears M1 basically rules the roost until perhaps 2023 now. Thankfully, Intel and the graphics card makers have basically been stagnant during this whole time.

Anybody that knows the Tim Cook Apple knows that they certainly were never going to pass the chip savings on to the customer.

We still don't know what kind of uplift M2 will have. But even if it's "only" 10%, that's still more than a typical "tock" Intel year gets.

The biggest complaint I have with the M1 variants is they're all just more cores - single-core (and in most cases, perceived computer speed) is equal amongst all the machines. I expected some single-core uplift with M1 Pro for example, but got none. Maybe they are saving it for M2. Or maybe they are at the limit of what they can achieve with this fab and need to wait for 3nm for the more substantial uplift.

 

[fourthtunz]

That's quite slow release cycle.

If they release M2 in late 2023 that would be exactly 3 years since M1 debut.

For comparison Intel updates their line up every year.

AMD every 1-2 years.

Thats true but Apple have released some Kick AS@ computers with the studio line
regardless. I am waiting for the new air but hope it has more ports!

 

[Isengardtom]

It won’t be the same SoC. it’ll be an A15 style upgrade with slightly better performance/efficiency but with 2 more GPU cores.

and I think they‘ll just call it M2

the bigger upgrade will be M3

 

[Psychicbob]

It’s an M1.5, an M1 with go faster stripes and a big spoiler

 

[HobeSoundDarryl]

Anybody that knows the Tim Cook Apple knows that they certainly were never going to pass the chip savings on to the customer.

We still don't know what kind of uplift M2 will have. But even if it's "only" 10%, that's still more than a typical "tock" Intel year gets.

The biggest complaint I have with the M1 variants is they're all just more cores - single-core (and in most cases, perceived computer speed) is equal amongst all the machines. I expected some single-core uplift with M1 Pro for example, but got none. Maybe they are saving it for M2. Or maybe they are at the limit of what they can achieve with this fab and need to wait for 3nm for the more substantial uplift.

Try Ultra… and then come to learn that not all of “double everything” MAX translates to double MAX power. Apparently, the software side needs to catch up with the hardware advancements.

I’m hoping the Mac Pro release comes with macOS refinements to maximize that hardware… because presumably optimizations for the rumored 4 MAX chips in QUAD means that the mere Ultra will enjoy the same software lifts.

One can hope anyway. I’m also hoping for full speed ports on this Ultra.

 

[Mr. Dee]

Looks like Apple is already reaching a plateau with even its chips. Its a scary situation, since we are seeing innovation among all their devices decline. Performance is already negligible between releases and SKUs, features are being announced and taking months to release or are unfinished. Apple needs a reset on how it brings new products and services to market. Maybe having 2 year gaps for both hardware software, along with some out of band updates would be a better strategy.

Our usage has become narrow anyway, you buy a new iPhone or MacBook for the 98% you are likely gonna be using it to do something in a web browser. At my work, staff like to order MacBook Pro’s because they think they are better, yet, they are using the same Citrix based application. Everything down to the printing is virtualized and remote.

The 2% who do anything creative with these devices wouldn’t be enough revenue for Apple to even survive on for a day. Users are obviously keep their devices much longer too. So, I think Apple really needs to rethink the need to bring new revs to market every year.

It was critical during the PowerPC and even Intel days. It would have been awkward to be on Haswell while every Windows OEM was on Broadwell or Skylake. But they are building their own hardware from the Silicon up to the software. They are in full control.

Tim, if you are reading this, delay all new software and hardware especially the iPhone 14 until October 31st 2024. Work on refinement and bug fixes providing long term support for whats already out there on the market.

Also, transform Apple Park into affordable apartments with free parking.

 

[Unregistered 4U]

This is, as far as I know, in recent years they use “marketing” to brand a 1 year old processor and fooling consumers that is a new chip by upgrading the name. Last time they did this was with the Apple Watch S7 which is a S6 processor?

These tricks are and should not be Apple. This is sad.

Apple’s always been clear on their pages which processor is in which device. The only folks ‘fooled” are those paying zero attention. Everyone else is just buying the newest of whatever and wouldn’t care what chip it had as long as it’s an always on screen

 

[Unregistered 4U]

But it’s been 18 months since MBA M1 and there’s still no upgrade.

That’s about right. As far as importance to Apple, the Mac is towards the bottom of the pile. And, since people keep devices for longer, there’s really no pressure to release new processors any faster than they already are.

 

[anthogag]

Even then, they care about the camera and battery life.

The Pros care about everything…the chip, battery life, camera, colour, weight, screen, sound,…take your pick😎

 

[Realityck]

Looks like Apple is already reaching a plateau with even its chips. Its a scary situation, since we are seeing innovation among all their devices decline. Performance is already negligible between releases and SKUs, features are being announced and taking months to release or are unfinished. Apple needs a reset on how it brings new products and services to market. Maybe having 2 year gaps for both hardware software, along with some out of band updates would be a better strategy.

Funny you are attempting to critique AS technology at a plateau when we haven't even seen Mac Pro yet. Neither have the native applications that take full advantage of what these M1 SoC can do is not represented yet. Look at the so late introduction of a Universal version of Plex, M1 was announced Nov 10 2020. There is so much further you can go with the OS and native applications, never mind using the latest Studio Ultra as one hardware technology marker.

 

[appleguy123]

Since when has process node been the defining factor in iPhone performance?

Since Apple (and really all current manufacturers) got so good at making chips that the limiting factors literally became the laws of physics and how small something can be made and still be atomically precise.

After 3nm I’d start to expect every chip update to feel like this for awhile.

 

[jdb8167]

That’s about right. As far as importance to Apple, the Mac is towards the bottom of the pile. And, since people keep devices for longer, there’s really no pressure to release new processors any faster than they already are.

Even in the heyday of Moore’s law fabrication node shrinks only came every 18 months. Now it takes longer. So it won’t be surprising that new Mac SoCs will take a couple of years. Apple might want to do what Intel used to do with a 12-18 month tick-tock cadence. That might be what the M2 is.

 

[Unregistered 4U]

I suspect the bulk of "us" certainly did expect a more robust upgrade schedule and big power jumps generation to generation.

I think anyone that expected a more robust upgrade schedule just didn’t understand where the Mac fits into the big picture. The watches, iPad and iPhone are all more important to Apple than the Mac. In a good year, they may sell 25 million Macs in the entire year where they’ll sell 25 million iPads in a quarter and 25 million iPhones in a month.

Thankfully, Intel and the graphics card makers have basically been stagnant during this whole time.

Even if they haven’t been stagnant, it’s not like any of them can run macOS. Anyone that cares about the highest performance in cross-platform metrics left Apple a long time ago. The majority of folks left just want something that can do what their current model does, just do it a little faster. There are certainly a few left that will always want to be in a waste water ejection contest with their Intel/AMD owning friends, and they’ll always be disappointed. Because, someone will ALWAYS be able to build and overclock one of those other systems such that they’ll show percentage improvements in benchmarks over whatever Apple releases.

 

[jdb8167]

Funny you are attempting to critique AS technology at a plateau when we haven't even seen Mac Pro yet. Neither have the native applications that take full advantage of what these M1 SoC can do is not represented yet. Look at the so late introduction of a Universal version of Plex, M1 was announced Nov 10 2020. There is so much further you can go with the OS and native applications, never mind using the latest Studio Ultra as one hardware technology marker.

I’m guessing that plateau started on March 8th with the introduction of the M1 Ultra. I mean it’s been almost 3 months. What has Apple done recently? /s

 

[NeoSe7en]

How are they “stuck” exactly? Haven’t reports suggested this has been the plan for some time now?

 

[nimitz9]

It appears Apple isn't getting as much of a chip performance this year as they wanted. If they knew about this far enough in advance they could have focused on improving other features as an upgrade model selling point, under glass camera, face id and finger print reader etc. But all of those(or any?) and every change to the iPhone has it's own roadmap/time-line to be released. Apple is very calculating on product marketing.

 

[ilikewhey]

whenever apple realize they are short of making the next iphone appealing, they tack on another feature to make it more worthwhile, therefore...usb-c on iphone 14 confirmed?

 

[reallynotnick]

Upgrade it to LPDDR5, the GPU cores from 8 to 10 and then move to the A15 architecture along with a redesign and the MBA would be a solid upgrade even if it's not earth shattering.

 

[Jamers99]

Well if the iPhone 14 doesn't have USB-C then it is a hard no for me. I will wait to upgrade our (4) iPhone 12's until the 15 is out. My iPad Pro from nearly 2 years ago has USB-C. Not sure why Apple is dragging their feet.

 

[Zwhaler]

Do people really think Apple’s plans have changed since March when the Senior VP of Hardware Engineering said that the Ultra was the last M1? There are not going to be any new M1s. I have no trouble believing that the M2 is a minor update on the M1 but there won’t be any new M1s. I’m also very skeptical that Apple would introduce a new MacBook Air more than 18 months after the M1 MBA with the same SoC. Logic says that having an M2 with the A15 or A16 CPU and GPU cores is going to be requirement by Apple’s marketing.

The opposite could also be true. From a business perspective, consumers care first and foremost that something is "new". How many of them deeply investigate the details like SoC architecture? Apple can get away with putting largely unchanged internals into an updated form factor, and they will still make near record sales. Many customers will be coming from much older machines; it's not like very many buyers update their laptops every year. Factor in the cost savings and Apple likely won't miss the customers who decide to wait another year.

 

[GooseInTheCaboose]

Anyone know how this new Air will compare to the entry M1Pro in the 14” MBP?

 

[commander.data]

Only every year since A4.

Shrinking transistors each year means Apple can put more in each year. That's how those impressive graphs about 30% or 50% boost come about. Apple can't magically increase performance without adding transistors. When TSMC slows, Apple Silicon slows.

The Apple A9 and A10 both used TSMC's 16nm process and the A10 delivered noticeable process gains through architectural improvements and more efficient die layout. Quickly and regularly adopting new process nodes has made it easier for Apple to increase performance each SoC generation, but it's definitely possible to increase performance on the same process. The best example is probably nVidia's Maxwell 2 generation GPUs which nearly doubled the performance of the previous Kepler generation GPUs without significantly increasing power consumption while using the same TSMC 28 nm process through a comprehensive architectural redesign. No doubt implementing major architectural changes on the same process requires long and careful planning so if TSMC's N3 and N4P processes not being ready for 2022 was only known in the past year there would not be enough time to backport what was intended to be a N3 or N4P A16 to N4 or N5P so the only choice would be to rebrand a refreshed A15 as the A16.

Personally, with new iPhones needing to be released every year in the September/October timeframe, I've been amazed Apple's SoC team has been able meet that no-fail deadline with new SoCs with generational improvements for as long as they have. Hopefully these rumours of problems are wrong and they'll do so again this year.

Do people really think Apple’s plans have changed since March when the Senior VP of Hardware Engineering said that the Ultra was the last M1? There are not going to be any new M1s. I have no trouble believing that the M2 is a minor update on the M1 but there won’t be any new M1s. I’m also very skeptical that Apple would introduce a new MacBook Air more than 18 months after the M1 MBA with the same SoC. Logic says that having an M2 with the A15 or A16 CPU and GPU cores is going to be requirement by Apple’s marketing.

Marketing can always work around previous statements if they wanted to. For example, if they wanted to save the "M2" term for a true architectural successor to the M1 line, they could just add a suffix to brand new refreshed/minor-enhanced M1 chips like M1E for enhanced. It could even be a whole line if required like M1E Pro and M1E Max. Calling it something like the M1E line would still be consistent with their statement that the M1 line is done.

 

[jdb8167]

It appears Apple isn't getting as much of a chip performance this year as they wanted.

Says who? I doubt any of the current TSMC offerings are a surprise to Apple.

 

[Zwhaler]

Anyone know how this new Air will compare to the entry M1Pro in the 14” MBP?

It has to be announced first. M1 Pro will almost certainly be more powerful than an M2, particularly on the GPU side. We don't know if we're getting a full M2 yet. Apple will compare them spec-wise when they come out either way.

 

[Danfango]

This doesn't surprise me. They are reaching the point of viability limit on shrinking features while producing a reasonable production yield. At some point we will not be able to make faster or more efficient CPUs with this approach. Then there will be a massive resurgence in computer science theory because the only way to make things faster will be to change how we approach problems computationally.

I am waiting for this day because one of the ways ahead is to burn all the layers and layers and layers of crap software abstraction that our society is built on. There will be a whole new computing paradigm which isn't basically NeXTstep bunged on top of a research project crossbred with BSD UNIX.

Computers are going to get very boring over the next few years. Also a good thing.