You've probably seen a lot recently about the possible introduction of Macs running ARM processors sometime in 2021. Many of us are expecting Apple to move to its own processors for at least part of its Mac lineup, but until recently, details have been light on what that move might look like.
Even now, as a clearer picture of the ARM transition starts to emerge, people may be left wondering what any of this means. What does it mean for Apple to move from Intel to ARM? Crucially, what makes this transition so important to the future of the Mac?
Well, I think you'll find that overall, it has less to do with ARM, and more about Apple taking greater control over its computational destiny, but I'm getting ahead of myself.
So just what is ARM, and what does it have to do with Apple?
ARM Holdings is a British company that designs and creates its own sets of CPUs and other chips. While ARM chips have a long history of powering a variety of devices (such as the Acorn Archimedes), today, the chips designed by the company and most of its licensees are found in things like embedded systems around the world.
But as well as creating its own chips, ARM also licenses out its instruction set architectures (an instruction set architecture is, essentially, what tells a chip how to execute code on a particular processor or type of processor). This means that companies buy a license that allows them to build custom processor cores that implement an ARM instruction set, rather than purchasing or modifying ARM processors themselves.
This is how Apple does what it does with its A-series systems-on-chips, and the distinction here is crucial. Apple designs its own CPUs and CPU cores that implement ARM instruction sets. The company's work is completely custom, rather than a repackage of ARM processors. Theoretically, Apple could license x86, the instruction set architecture used in processors from Intel and AMD, and build custom desktop and laptop chips that way, but the team is by now versed in ARM, and chips made with the ARM instruction set are known for their lower power consumption compared to x86.
This is all a way of saying that "ARM transition," while a convenient shorthand, doesn't fully describe what we expect to happen with upcoming Macs. We expect that, like the A-series chips in iPhones, iPads, and Apple TVs, Apple's Mac processors will be completely custom.
CPU vs. System-on-a-chip
When we talk about Apple's custom-designed silicon, we usually use phrases like "A13 processor" or "A13 CPU," but this terminology is imprecise. Apple's custom chips, while certainly dominated by the central processing unit CPU, are more than just single processors. They're a collection of essential components that are needed to run most computers, including Apple's iPhones and iPads.
An A-series system-on-a-chip consists of a processor, the memory (RAM, not storage), and the graphics processing unit (GPU) all on a single die. Components like storage, battery, radios for Bluetooth and Wi-Fi, and more are located outside the system-on-a-chip.
A system-on-a-chip is great for mobile devices due to its compact nature. It also takes less overall power than more separated systems. On systems like laptops, however, a system-on-a-chip approach isn't the common way of doing things. Instead, processor, memory, and GPU all generally have their own separate places on the logic board.
With the chip design team at Apple having cut their teeth creating small, contained systems like the A-series and S-series, we have to wonder whether this will carry over into the Mac. It seems unlikely. After all, the RAM on laptops and desktops both have access to more available power than the memory on a smartphone, so it's likely that Apple would want to take advantage of that in its Mac chips.
Additionally, a recent report indicates that Apple is currently working with a chip based on the as-yet-unannounced A14 chip, which is expected to come in a new line of iPhones to launch later in 2020. These are supposedly chips with eight high-performance cores and four high-efficiency cores, for a total of 12 cores. By contrast, the Apple A12X and A12Z have the highest core counts yet in an Apple system-on-a-chip at eight, with four high-performance cores and four high-efficiency cores.
Given that the A13 found in the iPhone 11 line features six cores (two high-performance, four high-efficiency), it seems likely that the phrased "based on" is very important in Bloomberg's report. The A14 that we'll see in the fall is likely toned down from a 12-core version said to be in testing for Macs. I would venture a guess that the Mac chip is a more powerful variant of the A14, much like the X-variants found in the iPad Pro line have been (the A12 was a six-core chip, the A12X and A12Z are eight-core). But they're all likely based on the same ARM instruction set.
Conservation of energy
One of the big reasons for Apple to switch to its own ARM-based processors is power consumption. Apple's chips are generally seen as more power-efficient than Intel's, and those results hold up. While Apple claims between 11 and 12 hours of battery life for its most recent MacBook Air and MacBook Pro, real-world usage generally sees considerably less life than that (it's generally between six and eight hours, according to a number of third party outlets).
Meanwhile, on the iPad Pro, I'm generally in the ballpark of Apple's claimed 10-hour battery life in a day of typical use. That can vary if I do something like play a resource-intensive game, but the maximum variance I see is generally about an hour less than expected at most, and that's on heavy-use days. It's more than enough to get me through a full workday.
It's not about power, it's about Power
So why do Apple's chips seem to be more consistent in their power usage than Intel's? Well, it goes to the reason that Apple would make this transition at all: Apple has a better idea of how its own chip is going to perform. When Apple owns the entire stack, both hardware and software, it can optimize everything with extreme specificity.
Because that's what it's really about. It doesn't matter that both the 2018 and 2020 iPad Pro match the 2020 MacBook Air in single-core performance benchmarks (and absolutely thrashes it in multi-core), or that it closes in on the Intel Core i9 in the 16-inch MacBook Pro in single-core, too. I mean, those things do matter, but they're secondary to the larger reason that Apple would make this move from Intel to its own processors.
It's about control.
Apple is a company that craves control. Not like it used to be, and it's certainly more flexible in terms of how customers can use its devices than its been in many years. But when it comes to actually making those devices, Apple wants to own as much of the production stack as possible, from hardware to software. Because processors are an essential component to all of the hardware that Apple makes, the idea that Apple should make the chips that power the Mac, rather than continuing with Intel or switching to AMD, makes a lot more sense.
If you view this through the lens of Apple wanting control, then the picture really starts to come into view. Because Intel has been letting Apple (and its customers) down for years now. It's consistently been late with new processes, die shrinkages (which means it's been late with better power efficiency, too), and it's lagged behind AMD for years. It's more mobile-friendly processors are continually matched or beaten by Apple's, and it's increasingly clear that Apple's Intel partnership, which was a great boon to both companies when it began 15 (um, wow!) years ago, has almost run its course.
And while AMD has been delivering very impressive results in its desktop and, much more recently, laptop chips, it's unlikely that Apple would have switched to them instead of going its own way. Why would Apple enter into an agreement with another third party chipmaker who, I would bet from Apple's perspective, will inevitably let them down? The good times can't last forever, as the Intel partnership has proven, and AMD is unlikely to be immune from the issues that have plagued Intel.
What using its own processors in the Mac gives Apple is the same thing it has over the iPhone, iPad, and Apple Watch: full control over hardware and software. Apple's chip design team has put out hit after hit for years now, but even they will slow down eventually, releasing incremental improvement after incremental improvements (they do that already — hello A12Z). But even when they do, Apple's going to know exactly when new chips are coming, and plan product releases around them. It'll also be able to release updates to Macs more consistently.
This way, everything that Apple's learned about making chips, everything it's learned about performance, power management, creating new processes, and everything else, will finally benefit the Mac line. And hopefully, Apple will take anything it learns from making Mac chips, and bring that back down to its mobile devices, too.
So for Apple, it's not really about ARM, per se. That might be the technology that the chip team excels in, but it's not the main factor in the move. It's about the fact that these would be Apple-made processors, through-and-through. Apple could finally be on its own schedule for all of its products.
Bringing it all together
Apple's transition to custom processors for Mac is all about control. It's just, hopefully, the kind of control that will offer a benefit to customers, as well as Apple. I, for one, am excited to see where the company may take its ARM-based Macs. Once we have the kind of power management Apple's mobile processors are known for in the Mac, could we see Apple finally make a cellular-equipped MacBook?
I'm also curious to see these processors run on desktops, like the iMac or Mac mini? Apple's chips are already powerful on the mobile side of things, particularly the AX-series systems-on-chips of the iPad Pro lineup. What can they do when you can count on your computer to always draw power from an outlet?
As intriguing as all of this speculation is, though, Apple hasn't announced anything yet. We still probably have a bit of time left to wait before we get official confirmation of any of this. Personally, I think if it's coming next year, Apple could let developers know during the all-virtual WWDC 2020. And even after it's announced, don't be surprised when we have to a while between the announcement and actual product launch.
But I'm excited to see what Apple has in store next for the next era of the Mac. I think we're in for a very interesting time indeed.
Joseph Keller is the former Editor in Chief of iMore. An Apple user for almost 20 years, he spends his time learning the ins and outs of iOS and macOS, always finding ways of getting the most out of his iPhone, iPad, Apple Watch, and Mac.
Since Apple has never stated its intentions in this regard, don't
you think it is pundits who want ARM on Mac? What do you do for a high data rate connection? How do you daisy chain devices? How do you handle floating point accuracy? How do you support a multi-platform storage architecture? Multiple users? Caching?
This is not a “transition”. Intel Macs are not going to go away. Ever. ARM will be for something like the MacBook Air. Light, no fans, great battery life. Of course, it may not go anywhere. Windows runs on ARM too, and Windows ARM laptops have not exactly flown off the shelves. Unlike Microsoft, Apple already has great mobile products. So Mac on ARM is not a desperation move, trying to get SOMETHING going on the mobile space.
I think although it might be possible on the MacBook Air, it will probably take a while longer too reach the MacBook Pro and the desktops. It will probably be even longer to take on the Mac Pro because that modularity would get even more complex that it already is with ARM architecture. But theoretically it world be great for a Apple when they can know exactly when the next chip sets are coming and basically be able to build everything around that AX/Z SoC.
Thank you for signing up to iMore. You will receive a verification email shortly.
There was a problem. Please refresh the page and try again.