2015: 64-bit ARM chips in iPhone 5S serve up taste of Intel-free future for Apple

2015: 64-bit ARM chips in iPhone 5S serve up taste of Intel-free future for Apple

Summary: The most important thing that was heard at the iPhone 5S launch event was talk of a "Desktop-Class" architecture using 64-bit ARM chips of Apple's own design.

TOPICS: Smartphones, Tablets, PCs

Like many others on the East Coast, I spent my lunch hour going through the various computer industry liveblogs of Apple's iPhone 5C and 5S launch event.

While many of the things we heard about both devices were covered extensively in various leaks and analysis in weeks prior, by far the most the most important new thing that was announced at the event was that the iPhone 5S uses a 64-bit ARM chip of Apple's own design, the A7.


Back in November of last year I postulated about the future roadmap for the semiconductors used in Apple's mobile and PC products. While it is surprising for many to see a 64-bit System on a Chip (SoC) on a smartphone so soon, I can certainly understand why Apple wanted to get one out into the wild.

The phrases "Forward-Thinking" and "Desktop Class" were thrown around during the launch event when discussing the iPhone 5S and the A7 processor. That stuff isn't just marketing hype in place here, it's a glimpse at the thought processes that are going on within Apple as it pertains to their long-term device and operating system strategy.

While a 64-bit SoC sounds like overkill for a smartphone, there are features within the iPhone 5S that will immediately benefit from or would not easily be possible without it.

Biometric authentication requires a significant amount of CPU horsepower to pull off without being sluggish, as would strong end-to-end VPN encryption, both of which are likely necessary for the iPhone to continue to attract corporate attention.

Image-processing features that are included in iOS 7 on the iPhone 5S also benefit highly from a 64-bit OS, although they can also be achieved on older models that support iOS 7, albeit run much slower.

And of course who can forget the games that are now possible with these chips as well, which up until now have been confined to advanced console systems like the XBOX 360 and the PlayStation 3, as well as gaming PCs.

These are the kind of things that were only possible on desktop computers less than 10 years ago. Yet now, with the advent of 64-bit SoCs, we're going to see them on mobile devices.

But this is just the tip of the iceberg. The reason why Apple wanted to get a 64-bit ARM chip into the wild is largely an issue of platform convergence.

I have discussed the concept of platform convergence a number of times in the column. I most recently talked about it in the context of the Ubuntu Edge, the Linux-based smartphone concept envisioned by Canonical that did not achieve its crowdsourced funding target. 

To summarize, I expect that our smartphones will become the center of our computing experience and even extend themselves to tablets and the desktop, through a unified operating system that runs identically on all three form factors and is supplanted by back-end Cloud services which will do the heavy lifting for our line-of-business applications and data.

To take the concept even further, I expect that smartphones will become the "brain" of modular tablet and notebook PC designs, much like was envisioned for the Ubuntu Edge.

For this transformation to occur, the operating system must be converged first. I believe that within just a few years, all of the major players will have converged systems to offer.

Microsoft has already ported Windows to ARM and is unifying its developer target by bringing Windows Phone and Windows 8.x closer to API parity.

By virtue of its platform having the lion's share of line of business applications on the desktop, and with thousands of ISVs with over 20 years of code to migrate among them (not to mention purely internal Windows LOB apps at major corporations) the full transition to the Modern-style UI as it exists in Windows 8.X and Windows Phone will take years.

Google is clearly working on marrying Chrome OS with Android, and Apple has now taken the first step in its own convergence process by releasing its 1st-generation 64-bit SoC with the iPhone 5S.

Once the 5S -- and presumably also the iPad 5 which will sport a similar chip -- is in the hands of developers, the 64-bit app transition for Apple will be underway.

So the question is, what happens to Macs?

iOS and Mac OS X, much like Windows Phone 8 and Windows 8.x which share a core OS and kernel as well as APIs, also share a common architecture. They both use the same kernel and a shared developer environment.

At some point, iOS will end up on some kind of laptop or tablet convertible device using a 64-bit ARM chip of Apple's design.

I think that will happen sometime around 2015, when the future 64-bit SoCs are powerful enough to actually assume the type of creative content desktop workloads Macs are actually used for today, and private and public clouds have achieved a level of maturity to provide "Extreme SaaS" functionality that cannot be easily achieved on a low-power endpoint device with locally-installed apps.

Because both Mac OS X and iOS are now 64-bit, it will be easier to port the more demanding (but fewer in number) Mac apps to a single, converged future-state Apple device OS.

Whether that OS is called Mac OS or iOS or something completely new is irrelevant considering how much DNA is already shared between the two.

Is the 64-bit A7 in the iPhone 5S a sign of a platform converged future at Apple? Talk Back and Let Me Know.


Topics: Smartphones, Tablets, PCs


Jason Perlow, Sr. Technology Editor at ZDNet, is a technologist with over two decades of experience integrating large heterogeneous multi-vendor computing environments in Fortune 500 companies. Jason is currently a Partner Technology Strategist with Microsoft Corp. His expressed views do not necessarily represent those of his employer.

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
              • Correct. Commoditization eventually takes over in every industry.

                This is why innovation is important to future growth. The PC market is saturated and right now hardware leads software by a pretty wide margin.

                For instance, you can run Windows 7/8 on six-year-old hardware so fewer people are replacing their hardware in order to get the latest OS.

                There is still lots of pent-up demand for tablets - especially for those who do not already own a PC. A tablet can do 90% of what we used to need a PC to do and it is more portable and simpler to use. (and it costs consumers about the same amount of money to start ($200 to $500).
                M Wagner
  • The party started in 2010 with the iPad. That is where the money is today.

    Development of the iPad and development of the Surface both started in 2009. Apple took the "stand alone" approach and Microsoft took the integration approach. A year late to market - yes. A decade late to market - I think not.
    M Wagner
  • Windows (and Linux) will continue to provide the "cloud services" that ...

    ... all those smartphones and tablets depend upon so no worries for Windows in the enterprise - either way. Consumer interest in Windows is a serious challenge for Microsoft though.

    People hold up the $900 million "write-off" on the Surface RT as indication of failure but Microsoft made $853 million on the Surface RT before the write-off.

    Only in Redmond can $853 million be called "a failure". The write-off is nothing more than money which will NOT BE MADE because of the lower price for the Surface RT.

    Post-PC really only means that consumers (who long-ago abandoned desktop PCs for notebook/netbook PCs) are now choosing tablets. Microsoft needs a piece of the consumer tablet/smartphone market to stay in the consumer marketplace.

    Microsoft once led with "Windows XP Tablet Edition" but the hardware wasn't there yet - and OEMs resisted selling them at consumer price-points.

    Today, Microsoft is behind because it took them so long to embrace hand-held devices.

    Still, Microsoft recently slipped into #3 behind Samsung (Android) and Apple (iOS) with Windows Phone 8 (with BlackBerry coming in fourth) in the SmartPhone business.

    It is not clear where they place (as a vendor) with respect to sales of the Surface RT but traditional Windows OEMs are having some success with Windows 8 tablets and Microsoft is selling the Surface Pro with some success so it is way too early to count Microsoft out of this race.
    M Wagner
    • Risky behavior

      Ms didn't "make" $853 million on RT. "make" assumes it is profit. It wasn't profit. That was the sales total, which was less than the write off.

      Microsoft being third means very little, as their position is way behind. In reality, their mobile position is in deep trouble. They needed to buy Nokia's handset division because Nokia was about to declare bankruptcy over debts due in early 2014. If that happened, the 84% of Win Phone sales would have been in danger of slipping away. Even so, they aren't out of the deep waters. If Nokia fans refuse to buy Microsoft Win Phone instead, they could see sales shrink.

      What Microsoft has done is very risky.
  • Computing cycles are computing cycles. They have to come from somewhere.

    No one has the corner on low power chip designs and when you start scaling up performance requirements, you scale up power requirements right along with it.

    Smartphones and tablets get by with less power consumption today because an ever greater number of operations are being moved to the cloud, where servers are doing the work once done on the smartphone (or the tablet, or the PC).

    iOS and Android are single-user operating systems. Windows RT is multi-user but they all lack preemptive multitasking so none of these operating systems will scale-up to server applications.

    The future of Windows (and UNIX/Linux for that matter) for very large scale file servers is NOT in doubt. Only Mac OS X is equally scalable but Apple is not interested in the enterprise.

    Whether Apple or Microsoft (or Ubuntu or Red Hat) ever port their respective OSes to ARM remains to be seen but, over the long haul, don't expect ARM designs to be dramatically more efficient (less power-hungry) than their Intel or AMD counterparts for the same performance.
    M Wagner
    • Wrong about Android

      "...iOS and Android are single-user operating systems. Windows RT is multi-user but they all lack preemptive multitasking so none of these operating systems will scale-up to server applications..."

      Android always had pre-emptive multitasking and official multi-user support came with 4.2
  • You said ...

    ... "The smartphone platforms will upscale to take on the role that the PC does today."

    This certainly isn't true because iOS and Android and Windows RT lack preemptive multitasking - which is absolutely critical to scaling up to multi-user environments.

    ... and you said ...

    "Apple gets it. ARM processors will one day power everything from phones to huge industrial server farms."

    Maybe ARM processor designs may permit that kind of scalability, and maybe not. But, in any event, neither iOS nor Android will scale up today because they are not built upon preemptive multitasking.

    Apple and Microsoft certainly would like to be able to port Windows and Mac OS X to ARM in order to get Intel to lower it's prices but it is not clear that the ARM instruction set can do this as efficiently as Intel can with x86/x64.

    Even Ubuntu was talking abut a hybrid Smartphone platform - running Android on ARM in "handheld mode" and Ubuntu on x86/x64 in "PC mode".

    Certainly, Intel needs stiffer competition than AMD can provide so it would be good for Mac OS X and Windows (and Linux) to be ported to ARM but legacy-software developers would have to come along and port their own apps to ARM in order to complete the transition of which you speak.
    M Wagner
    • preemptive multitasking.

      "neither iOS nor Android will scale up today because they are not built upon preemptive multitasking. "

      Not sure, what do you mean by "preemptive multitasking" but both iOS / Android are based on UNIX like operating systems FreeBSD / Linux respectively. So if do you mean technical ability to run concurrent processes they are definately can do it. Automatical suspending of non iteractive (background) processes / tasks is a design concept (improvoment) that was implemented in both systems to improve the batterly life.
      Andrey Lartsev
  • No it doesn't need to - but your suggestion that Windows is ...

    ... any less efficient than Mac OS X or UNIX/Linux misses the point.

    Whether or not the ARM instruction set is rich enough to deliver high-performance preemptive multitasking on an appropriate platform is an entirely different question. (The answer to which I do not know.)

    On a handheld device preemptive multitasking is OVERKILL, plain and simple, so why put a powerful OS like Windows or Mac OS X (of Linux) on an handheld device?

    BTW, that is what Windows RT and Windows Phone is all about. No preemptive multitasking means no legacy apps. It's essentially the same OS model as iOS and Android.
    M Wagner
  • I question the analogy...

    but not the point. I do sometimes use my toaster (oven) to cook meals. It's faster to heat up, draws less energy and for smaller items does just as well as an over and faster. Plus, in the summer, it produces much less waste heat that the central air has pump out. So it has it's uses, just like a mobile device. By the way, I don't know if you meant an actual toaster, but I can't think of anything that I would cook in a toaster that I would in an oven.

    Having said that, content creation on a CPU designed for a mobile device? Not anytime soon, unless there are *major* advances in that processor technology coming. I transcode and author DVD's and Blu-Rays on my desktop. I really doubt an ARM CPU is going to match a Intel or even an AMD chip for that by 2015. And it will be a while longer before they can do it in a tolerable time frame. Even if Intel seems intent on focusing on the mobile market with their higher end mainstream processors.

    These processors are good for consumption, which is a large portion of the market. They might be acceptable for editing images, audio and even video at a lower level (read: Low resolution). Again, good for much of the populace and some tasks for the remainder. But that's about it and will be for a while. So the market still has potential for growth. But like you, I don't see these processors superseding traditional desktop CPUs for a while.
    • DVD's and Blu-Rays on your desktop?

      What's a DVD? Blu-Ray? Desktop? What planet are YOU living on where people are still buying DVD's and using desktops?? ;)
      Freddy Sanford
      • What planet are you are on where they don't? - Really

        You actually believe that asinine comment you made? I'm using a desktop, I know 1500 employees at my company that use a desktop every single day. I have OS X, Win7 and several variations of Linux - ALL DESKTOPs sheesh. I love these goofballs that think just because their life is so simple they can handle everything on the cloud through Googles crappy versions of thing that - "hey if I can do it then everyone should be able too". I can tell you right now that Google Docs covers about 20% of what I do with office document software. Until I can do everything I do and my IT customers do with Google Docs then they are always behind IMO.
  • Won't ever be equal

    Give me an example where virtually unlimited power, practically unlimited cooling, and magnitudes more space is ever bested by a smaller, power/cooling restricted sibling... The cheese will always be moved ahead of the mouse.

    Side note: Developers won't write 64-bit apps until there is a large enough market. I'm not claiming I know what number that is, but we are at least a year off given contracts and the rate at which the 5s will disperse. Then again, the masses of Android devices are low-end. Doubt we will see those hitting low-end devices any time soon.
    • you don't need to

      you don't need to "code at 64 bit"....thanks to the improved XCODE the compiler do it for you . you write your code as usual....you compile it for 32 and/ or 64 bit (yes...you just need to press a button) .you didn't heard during the krynote that the programmer of infinity blade 3 needed just 2 hours to compile for 64 bit and review the final result?
      • Are You Saying . . .

        . . . that the XCode compiler "just works"
        • scalability

          Meanwhile in Android, we recompile the OS and not the apps to take advantage of new hardware. Modular, scalable, modern computing. Re-compiling is so 90's.
          That's the definition of JUST WORKS.
          • Huh, isn't that the same thing as re-compiling

            your code in Xcode for 64-bit? new libraries for new hardware recompile? So you write your output for a GalaxyS3 and you say all you have to recompile for it to work on the Note 8, screen? yeah that "Just works"... I think you just confirmed what everyone above just said, albeit I have no idea what Android has to do with it...
          • no, it's not the same

            IOS requires re-compiling the apps to take advantage of hardware features whereas Android apps which complies with the Dalvik VM will simply take advantage of the hardware with upgraded OS.
            That's why there's now fragmented iOS apps for iPhone, iPad, Retina, and now 64bit.
      • words

        the compiler do it for (does)
        you didn't heard during the krynote hear, what is a krynote (keynote?)
        I hadn't heard that the compilers today can compile to different sets easily, I have not programmed much in 45 years. That is what I get for being out of touch. Next thing you will tell me is that one compiler can compile to different languages (C, BASIC, FORTRAN, Forth, VB, VC (and variants), et al) for what ever situation you need.
        • Hey! if it's on Android and Google made

          it then I'm sure it compiles everything and it "just works" even crap...