Why Apple will switch to ARM-based Apple A-series-powered Macs

“Somewhere on Apple’s campus, ARM-based Macs are already running OS X,” Matt Richmond blogs.

“In September of 2013, the world’s preeminent independent processor expert compared Apple’s latest iPhone chip with Intel’s fastest tablet chip and concluded that the two perform similarly — even though the Intel chip draws more power, contains four cores versus the A7’s two, and is produced with a more advanced manufacturing technique,” Richmond writes. “If Apple’s chip design team can create a phone processor that performs on par with Intel’s fastest tablet chip, the company’s ‘highest priority,’ then there’s no reason to believe that the same team at Apple can’t design chips powerful enough for any Mac in the company’s lineup.”

“Apple wouldn’t have been able to create Touch ID if the iPhone were powered by an Intel chip instead of an Apple-designed one. There wouldn’t have been a “secure enclave” on the iPhone’s processor to store the fingerprint data, nor would there have been perfect hardware-software integration. Apple was able to implement Touch ID because it designed the A7 chip in tandem with the iPhone 5S’s software and the rest of its hardware,” Richmond writes. “I’d bet that there are features Apple envisions for the Mac that simply can’t be built while Intel designs the chips inside of them.”

Much more in the full article -recommended – here.

Related articles:
Why Apple won’t dump Intel x86 for its own ARM chips in MacBooks and the Mac Pro – August 5, 2014
Intel-powered Macs: The end is nigh – August 4, 2014
Intel’s Broadwell chips further delayed; not shipping for most Macs until early-mid 2015 – July 9, 2014
Apple will inevitably drop Intel for their own A-series processors in the Mac – June 26, 2014
How long before Apple dumps Intel from MacBook Air? – June 26, 2013


    1. That is a big issue. I mean the OS’es can be incompatible. But once you switch processor architectures the VM’s people rely on will run dog slow. Anyone remember having to run the x86 emulators on the first OSX machines with the Power PC CPU? Yuck!

      1. The VM’s people will have to be happy with Windows, Ubuntu or other such systems. Of all the things that could possibly hold it up, maintaining non-OSX compatibility goes to the bottom of the list.

      1. OSX simply cannot run on a RISC processor! Except for OSX 10.5, well ok, BUT it could never run on a RISC processor now because…

        Look, CISC and RISC, see how they’re spelled differently? That means there’s a difference. And difference is a thing that, you know, prevents stuff!


    2. Lots of words: Yes it will happen! Apple isn’t going to stop advancing their A-chips faster than the industry. Apple isn’t going to stop moving up and down the supply chain. They want to control and customize all the critical technology and nothing is more critical than the processor. They already design the best mobile processors in the world. Their vertical control of developer tools would allow developers to migrate to a new instruction set very quickly.

      Apple won’t do it until they can release a great product, but for many people a mid-range/cost MacBook Air A-series would be “a glass of ice water to someone in Hell.”

      The ONLY significant advantage Intel has is its fab leadership. That gives them a price and computing power advantage over ARM, but Apple has a way of solving problems nobody thought they could.

      (Pure speculation, but as Intel continues to scramble for a hold in the mobile game and reverse ongoing processor share losses, Apple could offer them an interesting deal. Apple could give Intel all their A-processor (mobile and desktop) fab business if they gave Apple access to those fabs. Intel has made huge pivots before to stay on top, and a move like this would put them back at the top of the whole processor business in the fastest growing desktop segment, Macs. And it wouldn’t undercut their existing Intel designed processor business as PC makers would not get access to Apple’s processors.)

      1. No, it won’t happen. I expect a long and FAST future for Apple’s A-Series ARM chips. I expect lots more Apple devices based on this RISC technology.

        But to say that Apple would even consider dumping Intel CISC chips to go RISC again at this point in time is pure IGNORANCE! (Yes, I’m talking at you Louis Gassée! For shame!) I rant some more further down the thread…

        1. Your impression of RISC v. CISC seems superficial at best.

          With a decent compiler that will create good code (say the LLVM compiler in Xcode, for example), the difference for the VAST majority of developers is NIL.

        2. Shutup and do your damned homework, lazy child. IASSOTS where tykes like you pretend recoding from CISC to RISC is as easy as recompiling.

          NO IT’S NOT. Go look up all the hardware APIs that are part of the Intel CISC ships. Then wet your pants in embarrassment. You can’t ‘compile’ that stuff you little twit.

          Oooo, you baby coders are soooo newborn.

          — I actually find it funny, as well as sad, that people are willing to stick their necks out and demand that their ridiculous rubbish is FACT! How entirely human. My cynicism is showing.

          Homework! Get to it!

        3. You’re right. Any function you write in Xcode can be compiled to run on OSX and iOS. I’ve tried it and it works as expected. However, today’s venture will be to determine if there exists any function that will compile to run under OSX that will not run under iOS.

          And IASSOTS seems to only mean something to a guy named Derek Currie according to a Google search 🙂

        4. Any function provided by the OS would be up to Apple to provide in an ARM OS X. And one can assume that would be done before any announcement.

          As for the base languages — C, Objective-C, C++, Objective C++, Swift, etc. — a simple recompile is all that’s necessary to change processors.

          One would also assume that Apple would provide a Fat file format where a developer compiles to BOTH Intel and ARM and puts it in one executable file — just as they did with 68K / PPC and PPC / Intel.

          Both CISC machines and RISC machines are general purpose computers. What can be done on one can be done on the other. Only the OS, device drivers, and very specialized utilities will care what the underlying processor is.

        5. Actually, someone else pointed out that, due to the Mac App Store, they wouldn’t have to do the Fat Binaries thing. You log into the store, you only see the applications compatible with your device, and you download only the code you need.

          There are options available to Apple now that never existed in the past. Their own compiler, their own repository to hold most all of the applications people use, and as of this fall, their own programming language.

          I’d imagine Apple’s just iterating processors internally waiting for a time when they see parity between Swift code compiled on Intel versus Swift code compiled on ARM.

  1. I sure hope this doesn’t happen…. 95% of all engineering software, Europe or North America runs on intel. Because of this, parallels runs as fast as a windows PC.

    Switching to ARM will slow this way down.

    Fingers crossed.

      1. 68K and PPC both offered some clear advantages over Intel at the time Apple decided to use them.

        If you can speculate on one or two significant and meaningful advantages (“progress”) an Apple desktop A-series CPU would have at launch over an Intel one, please suggest it.

        1. That’s the thing, there doesn’t have to be a technical advantage. If it can run Safari JUST as good, run iTunes JUST as good, run the Mac App Store JUST as good and runs most all of the apps JUST as good, then Apple’s device roadmap JUST becomes entirey theirs to define.

          Until that time, there’s JUST no reason to consider a transition.

        2. How would it be “theirs?” Apple doesn’t fabricate chips. Building a plant takes years, even if you can license all the IP for the manufacturing process and hire enough experienced engineers. Even with Apple-grade secrecy, we would know if that was afoot.

          In a shorter timeframe, the only company other than Intel that has the fab capacity for high-density high-yield SOC manufacture is (drumroll) Samsung. Is Mr. Cook going to put the Mac’s future in their hands?

        3. It becomes theirs to define in the same way that it’s theirs to define in the mobile space. The A8 chip isn’t a secret at all, BUT the timing of it’s release is for Apple to decide, not Intel, not ARM. We also know that after A8 will be another processor and after that will be another, each one fabricated and delivered to Apple on their timeline. The problem isn’t fabrication, there’s a ton of companies out there that would fab for them. The problem is delivery.

          For the mobile devices, there’s not a situation where they’ve planned for a year to release new product then Intel comes back and says, “Hey, you know what? We goofed. We won’t be ready but in another 9 months, we’ll have some product for you… maybe”. Not being held to anyone’s schedule but “their own” makes their Mac device roadmap “theirs” to define.

      2. The real impediment is that the largest 3D CAD vendors & similar high end software and then thousands of add-ons to those programs from 3rd party software shops to enhance capabilities for FEA, plastic mold flow, heat and fluid flow are Windows only.

        It will take a lot of years for a major engineering software package and all its add-ons to add the Mac as a platform. That may mean not until past 2020.

        1. But….. Where’s the problem? As long as they make Windows computers, then, you know, buy one.

          The only problem will be for people buying a Mac to run Windows applications. On this point I’m curious. You’re a big business, your industry depends on Windows software, and in order to RUN that WINDOWS software…. You’d buy a Mac? Why?

        2. What’s the purpose of being able to legally run OSX if there’s no industry software that runs on it? Especially if the computer will always be running Windows (if it’s not running Windows you’re not “doing work”/”making money” with it).

        3. The reason is because we understand the reliability of the Mac platform. We are running (7) mac pros, a dozen regular mac’s (mini or iMac or iPads etc). Mac servers….no antivirus software, no files lost running apple servers, etc. we understand all of this…price to quality, etc ….unfortunately, the expensive BIM software that run, that we make money using…$12,000 a license…only runs on windoze…. You forcing me to go back if they do this? WTF?

      3. Apple will release ARM Macs when they are ready, but they will keep selling Intel Macs for Pros for long after that. Mid-range ARM Macs would be great for most people, but Apple has shown they are still there for the pros.

        So you can rest easy!

      4. Can you think of how far advanced we’d be without the scourge that is Windoze???

        OMFG yes! I continually point out to people that we’re going to continue to be in The Stone Age of Computing until every copy of Windows has been ERASED at the STAKE! 💥😆💥😎💥😋

        I’ll be dancing on its ashes.

        But also imagine if IBM had kept up their end of the PowerPC RISC CPU deal with Apple (which they did NOT!). We’d still be intel-Free! But we also couldn’t virtualize Windows or run the crapware in Boot Camp. Like it or not, there are still occasions when Windoze is a requirement in these Stone Age times. Woe is us. (;_;)

        1. Not as much as you may think. One of the reasons Apple is making such progress with its A chips is because it could not find what it wanted, and thus determined it was better to create its own chip. The flip side of that is Apple’s iOS is the dominant OS for developers, especially ones who like to make money and not malware, so they choose to develop for iOS and the A chips.

          Even if Windows had not taken such a hold as it did, Intel chips were still the dominant chip architecture long before Windows came out. Perhaps another company could have challenged Intel more, but some certainly tried and either failed or had moderate success (AMD).

        2. Total agreement about the profound wisdom of going RISC again on NON-Mac software dependent hardware. Apple invested in ARM way-back-when and has consistently pushed for improved RISC technology. The ONLY reason they ate the Intel CISC slop was because the progress of the PowerPC effort literally died thanks to IBM. Apple was lied to for a full three years that there was a mobile G5 chip coming. It NEVER happened, to this day! So we’re no Intel CPUs now thanks to IBM.

          There’s nothing to love about CISC from my POV. But it was required to get the Mac this far. There was, and continues to be, no RISC path for Mac at this time. These rumors to the contrary are filler fodder for the cattle who don’t know any better.

    1. So basically what you’re saying is Intel controls the entire PC industry and we will have to follow Intel forever because we’re scared of change. I don’t think it works that way at all. Intel will be in charge until something better comes along. I totally respect Intel’s processors but I think something else can come along that’s just as good or better. I’m not looking for a change and will continue to happily use Macs with Intel processors unless Apple can make some huge leap forward in processing power and I don’t expect that to happen. I’m completely satisfied with Intel processors and small incremental increases in processing power. I can’t even imagine what Apple could offer using ARM chips for a desktop computer that would make me switch. Not even dropping upgrade support would cause me to switch until I was good and ready to buy a newer computer.

  2. At least… not happening until the A-series chips are so fast that they can run Intel applications (Windows, but more importantly, Mac legacy code) under emulation without a significant performance hit.

    1. Apple won’t stop selling Intel Macs for pros when they introduce ARM Macs for the “rest of us”. The point will be to better serve the non-pro market (which is huge).

      Windows compatibility or high-end Intel processor power are not holding up this transition, or in danger of being dropped any time soon.

  3. Absolutely stupid reasoning. The flaw in the logic is laid out plainly in his own words: that Intel leads the desktop/laptop CPU performance space, but can’t hack it in mobile space.

    So what the hell makes him think Apple, a leader in the specialized mobile CPU space, will come anywhere close to matching Intel desktop/laptop CPUs right out of the gate?

    1. “what the hell makes him think Apple will come anywhere close to matching Intel CPUs right out of the gate?”

      That sounds a lot like this quote from Palm’s CEO, Ed Colligan, back in 2006:

      “We’ve learned and struggled for a few years here figuring out how to make a decent phone,” he said. “PC guys are not going to just figure this out. They’re not going to just walk in.”

      1. You’re comparing a device to a *component* of a device. Apple specializes in integrating components into killer devices, and doesn’t compete based on components (specs) themselves, remember? At least that’s the running claim here at MDN.

        To justify an architecture change an A* chip would have to offer some clear advantages over Intel CPUs. I asked another desktop A* CPU advocate to speculate what those might be, and got no answer.

    2. Really? Apple can’t do X?

      In any case you have already answered why and how Apple can succeed with desktop chips: They didn’t have to best Intel processors to take over the mobile market. They just had to beat it in price/power value for the mobile segment.

      They don’t have to compete with the top of Intel’s line to create fabulous mid-range desktop processors with better price/performance in that range. And that is the market where most people are. Very few people really need top Intel speeds.

  4. Never say never, but this isn’t happening any time soon. There’s still a huge gap in processing power between these mobile and desktop chips, and killing Windows backwards compatibility is not something Apple will do without a very compelling reason.

    1. Apple didn’t best Intel to create the best mobile processors.

      They don’t need to best Intel’s top-of-the-line chips to create the best mid-range desktop processors for the majority of people on the planet.

      Apple is creative like that. They make the best solution for a particular segment, and don’t care a bit whether they are the best in theory. Processor power does not necessarily make a product great or not.

      1. Yeah but that suggests fragmenting into three platforms: OS X Intel, iOS ARM, and a new OS X-like platform running on ARM processors. That sounds problematic for developers and very confusing for end users

  5. Folks, this isn’t going to happen. Even ignoring the whole VM market, you would have to buy EVERY SINGLE app you own all over again because they would have to be recompiled for ARM. That’s going to cost a pretty penny. While it’s true that many of the standard apps (Word, Excel, Google Docs, etc.) have moved to the cloud, there are still a TON of native apps which have to be repurchased. This is just too much of a pain…

    1. I made the switch from 68k to PPC, and from PPC to Intel. Apple was not (overly) concerned then about the cost of upgrading legacy software. Apple focuses on future products/sales. If it happens, it will be when Apple decides future benefits outweigh current costs. (That said, I don’t think it’s happening any time soon.)

      1. I’m not happy to say that I went from 68k to PPC to Wintel. But even MacOS 7 was one of those periods where Apple moved forward without a care about people migrating to the new OS – they *knew* we would do it, and it was fairly seamless. Who still complains about the lack of a floppy? But they did it without even providing other options. This, too, will be done when they know we need to move that way. And I also agree that the power isn’t there yet, but the benefits are easy to see – Apple’s control/design, integration with an already large app and developer base. I suspect that Swift was a deliberate step towards smoothing that recompilation bump. I bet *Swift* compiled programs run mightily on both OSes!

    2. Aah someone with some common sense. The only place this might happen in the foreseeable future would be a news range of crossover devices, the larger iPad coming into desktop territory for example not for the desktop or even mainstream laptop products any time soon.

    3. That’s funny… I recently had to repurchase VMware Fusion because it wasn’t compatible with the latest operating system, never mind processor architecture which hasn’t changed at all in my iMac.

      I don’t understand why they would have to be recompiled? Didn’t Rosetta handle the code translation during the move from PPC to Intel? That was actually an extremely smooth process. Don’t you think Apple would do the same if they switched again?

      Furthermore, why would they completely switch? Why not two distinct Mac lines. one for consumer that ran ARM and one for prosumer that ran Intel.

      And there’s no reason you’d need to rebuy EVERYTHING. The developer could simply update the app to run on both architectures.

      1. Rosetta worked because the Intel CPUs were more powerful than the PPC processors they replaced.

        Out of the gate, an A-series desktop CPU is unlikely to be faster than the previous-generation desktop Intel CPU, so “Rosetta 2” users get a double-whammy of emulating a faster CPU on a slower one.


    Again Again Again kids:




    Intel = CISC
    ARM = RISC

    Going from RISC to CISC is relatively easy!
    Going from CISC to RISC is a massive PITA. At this point in time it would be like Apple blowing off one of its hands with an M80. NOT gonna happen. I don’t care if you’re Louis Gassée. NOT gonna happen.

    And no I won’t bother with replies. This rumor is just a waste of time, except of course to debunk it. So there. 😛

    1. CISC and RISC are both types of general purpose computers. Anything that can be done on one can be done on the other.

      Today the vast majority of even the OS is written in a high level language. That places the work of moving from CISC to RISC on the compiler for that language. Much like the LLVM compiler in Xcode can compile for either Intel or ARM CPUs.

      Apple has been telling us since the introduction of the iPhone that iOS and OS X share a common base of code. That is possible because of the work done on the compiler.

      You really are out of your depth, here.

        1. I don’t care if I win. Why don’t you care if you know what you’re talking about? Why do you care to engage anyone in conversations about subjects you don’t understand? I’m not asking for a reply. I’m asking for you to, as I pointed out originally, do your homework.

      1. Stop Kirk!! He said he wouldn’t reply and you’re forcing him to go back on his word. 🙂

        I’ve actually spent a good amount of time looking for anyone that also believes that OSX, which started out as a RISC OS, couldn’t be a RISC OS again. The strongest wording I’ve found only says that the current processor isn’t powerful enough, BUT if it ever becomes powerful enough, it becomes an option. As everyone knows Apple is working on making the A series chips more powerful, it really becomes more WHEN than IF.

  7. Why do people keep under-estimating Apple?

    * Apple can continue selling Intel/Windows compatible Macs (to the small but influential percentage of market needs this) along side mid-range and mass market ARM Macs (for the vast majority of consumers).

    * Apple does not need to best Intel’s speeds or fabs to create a great product for the vast majority of people that don’t need the best Intel speeds.

    * Apple has to grow. They are huge, so to do that they need to be bold. Apple only chooses to grow in a few ways, supply chain cannibalism being one. Making their own desktop chips provides more opportunity for cost savings, differentiation, design freedom, than about any other vertical move.

    * Apple will have no trouble creating a competitive ARM desktop chip. They are the best at the mobile segment. They can be the best at the mainstream laptop and desktop market, without any need to be better than Intel at the highest end.

    * Apple has obviously been laying the ground work for this. They have blatantly called ARM desktop class, even though it is obviously “only” low-end desktop class so far. Their development tool stack is ready to quickly migrate developers to an chips they chose to use.

    * Steve would do it. While he was very pragmatic he wanted to control all the details. However Apple has change, that has not changed yet.

    1. The argument that Apple needs (or would even tolerate) a slow cheap computer in its lineup is just an old story recycled, like the cheap iPhone or the “Apple is doomed without a netbook” meme.

    2. One other point–Apple doesn’t make any of its own chips. The OSX processors are fabricated by Intel and the iOS processors are designed by Apple but fabricated by Samsung. Building a state of the art chip plant is spectacularly expensive, even by Apple’s standards, and not many companies have that expertise.

      They aren’t about to move to A-series for Macs unless they have a secure source for wicked-fast chips.

  8. I don’t see Apple changing Macs to A chips anytime soon. The problem lies more with the necessary computing power for its iMac and Mac Pro models. Apple is now getting major CAD software for OS X that was not available for years, and then it’s going to throw in a processor switch for them? Sorry, but that’s a good way to get some of these firms to drop the Mac again.

    In addition, moving to A chips for Macs isn’t simple. Not only does OS X have to be rewritten/compiled for the new instructions sets, all applications have to be rewritten as well. Sure, Apple could have a Rosetta-like layer to ease the transition, but in the end software has to be rewritten.

    And for what? What major advances would A chips bring to Macs that aren’t already present in Intel chips or which are road mapped to be added? Plus, people don’t upgrade Macs every 1-2 years, it’s more like 4-5, so a transition could take quite a long time.

    No, Apple is better off concentrating its resources on keeping the A chips as fast, iOS-specific mobile chips and let Intel be the OS X chips.

    1. Why would adding mid-range ARM Macs to the line up cause any problems for Intel iMac/Pro models?

      The market for ARM Macs would be a huge new market for Apple, not a replacement for higher end Intel Macs.

      1. Because the software would have to run on both platforms. That is just as workable as supporting both Windows 8.1 and Windows RT on the same computer. If Apple ever switches, it will be across the entire Mac line like the previous times.

        The transitions to PowerPC and Intel worked because the new processors were so fast that they could emulate the existing midrange Macs. The A-series is still light years away from that.

        iOS and OSX share origins, but they do different things. The A7 just has to support one foreground program and a few background processes that are optimized to steal as few processor cycles as possible. My wife currently has 15 programs running on her iMac with over 200 open windows.

        If you tried that on an A7 or anything likely to evolve from it within a few years , the performance would be unacceptable even with native code, much less for Intel stuff run in emulation.

        1. You are absolutely right that emulating Intel on ARM would not be acceptable. Which is why when Apple introduces ARM chips they won’t cancel all their Intel Macs.

          Most people will be happy with the ARM once its ready, but Apple isn’t going to try and out compete Intel on the high end from the get go. They don’t need to, the mass market doesn’t need that kind of speed.

          It will be smooth transition for everyone, except for Intel and Microsoft’s profit margins which will crater as the mid-market gets scooped up by Apple.

    2. “but in the end software has to be rewritten”

      If Apple does their job well, you could just open your old project in the latest Xcode, recompile to the new processor, then you’re ready for testing. It could end up that very little code has to be rewritten, especially if performance is good.

  9. To start, I want to say I’ve never been a big fan of this idea. But that doesn’t mean that such a move is impossible. And most of the objections (including all I can thing of) can be overcome.

    First, the vast majority of software is completely processor agnostic. I can recompile software for ARM, Intel, PPC, and a dozen other processors with very little work using LLVM. Some software, such as Photoshop, has in the past been optimized for specific processors. A small amount of critical code has to be customized for the processor for best performance while the bulk of the code remains common across platforms/processors.

    However, today such software optimizes not for the CPU but for the GPU. Enter Apple’s recently announced Metal — a low level layer between the programmer and the GPU. Making optimization to Metal rather than a specific GPU very attractive.

    Second, device drivers and the underlying OS is where most CPU specific modifications must be made. We’ve been told since the first iPhone that iOS and OS X share a common code base. That tells us that much of the work to get OS X working on ARM has been done. This part of the transition is totally on Apple. If I’m making an app for the Mac App Store, once Apple does it’s work, a simple recompile will make it ARM compatible.

    Third, virtual machines. The transition from 68k (CISC) to PPC (RISC) worked because of the great performance advantages of PPC. By contrast, the first Intel Macs were slower than the PPC machines they were replacing. It worked because of JIT (Just in Time) translating in Rosetta, which allowed PPC code to work at near full speed on Intel.

    Those were software solutions. But we’re not talking about Apple using off the shelf parts that it must force into compatibility. For the first time, we’re talking about Apple designing its own hardware. Imagine a desktop chip from Apple with 4 or 6 ARM cores AND 1 or 2 Intel i64 compatible cores. During the Intel transition, Intel code (including the OS) would run on Intel and PPC code ran under Rosetta. A transition with a hybrid chip, ARM code would run on the ARM cores, Intel code would run on the Intel core(s). Only highly parallel Intel software would suffer in such a situation.

    Again, I don’t see it happening soon. But the closer I look at it, I see the foundations being laid. And the more I look into it, the more is seem not only possible, but probable — eventually (i.e. when Apple is good and ready).

    1. This. Intel has been using RISC cores with a CISC decoder for x86 chips since the Pentium 4. Which is the same thing Apple could do with an ARM chip if they wanted to. If it makes sense spec wise and money wise, they will do it, otherwise they won’t. The “problem” everyone is arguing about was solved almost twenty years ago.

    1. Remember, Apple is using Intel’s integrated graphics because they HAVE to. Intel made a change awhile back that means you can’t use their processor without their graphics. So, got a system with limited space and want to have decent graphics power from AMD or Nvidia? Since you’re forced to have Intel’s graphic chip, you either
      A) try to shoehorn in two graphics chips into a tight space.
      B) just use Intel’s graphics chp.
      There is no option for
      C) use Intel’s CPU and a good performance low tier graphics chip from another company.

      I’m sure that Apple (and other vendors as well) wish on the low end they had a choice other than Intel integrated.

      In addition, this also means that for power machines, if you just want one Powerful chip… You still have to leave space for Intel’s integrated graphics…

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.