This is the chip that could make your MacBook even thinner and lighter

Intel’s current chip architecture, branded as Haswell, uses a 22 nanometer process. They are about to roll out a significantly smaller and faster architecture, known as Broadwell, which uses a 14 nanometer process.

Because Intel’s new chip is so small—25% smaller than its predecessor—and requires substantially less power, it may allow for the creation of tablets and other devices without the whirr and thickness of the cooling fans currently used to keep chips from overheating. That’s not a guarantee, but it’s the first time it’s been a serious design possibility.

Running with less energy may seem like a minor thing, but it’s everything for mobile devices—which are basically giant batteries inside pretty cases, with a few other components. It’s also a broadside against a major Intel competitor, ARM Holdings. ARM chips are cheap to make and don’t require much power—enabling thinner, lighter devices with smaller batteries. They’re used in iPhones, iPads, Androids, Windows phones, and some Windows tablets.

This is an important pivot for Intel. Broadwell is their first chipset designed with mobile devices in mind. Unfortunately for Intel, this solution came late enough to the game to give ARM a real leg up on the market. Is this chip small enough and cheap enough to turn Apple’s head? Remains to be seen.

One thing worth a look. If you look at the slide deck at the bottom of the linked article, check out pages 22-24 for some pretty cool pictures of actual Haswell and Broadwell transistors. For an even better look, here’s a link to the original Intel deck.



  • http://sumocat.blogspot.com Sumocat

    Pfft. Intel pushed the Mobile Internet Device form factor in 2007, was involved with Ultra-Mobile PCs prior to that, and I seem to recall Oak Trail was supposed to take on ARM in tablets in 2011. But, yeah, I’m sure Broadwell, “their first chipset designed with mobile devices in mind,” will finally turn things around for them.

  • chjode

    One could argue that Centrino/Carmel back in 2003 was Intel’s first chipset designed for mobile.

  • Joseph Blake

    Will this draw ARM tablets back to x86? Maybe in the Windows world where that’s important, but probably not in iOS. What these WILL do, though, is to stop any potential switch from x86 to ARM that either Apple or Chromebook/Linux or even Wintel builders may be considering.

    • http://geekfun.com/ Erik S.

      Nope.

      First, it seems unlikely that any Wintel builders are considering switching to ARM. It isn’t really their choice, because they’d have to take what Microsoft and Windows application vendors will give them, which, as we’ve seen with Windows RT, isn’t much.

      Second, There are already ARM chromebooks, often from makers that also have x86 Chromebooks. This isn’t going to do anything to stop their use of ARM now or in the future. On its own, the processor instruction set architecture doesn’t matter on chromebooks like it does for Windows machines.. All that matters is the price and performance. ARM tends to offer a better price/performance ratio, but, to this point ARM CPUs haven’t been able to cover a wide enough range of absolute performance levels. Intel may hold the high end for quite a while, but ARM CPUs are covering more and more performance levels and price-points every year. Chrome.

      Finally, if Apple has an active program to transition Macintosh to ARM (and they probably do), the best Broadwell can do is keep them from accelerating their timetable. It isn’t like it is a surprise. Its just another Intel CPU generation, pumped out on a predictable cadence, give or take a few months. That’s the problem, Apple wants Intel to move faster. Instead, Broadwell is actually late.

  • Herding_sheep

    This whole thing with Broadwell and reading through the Intel presentation just makes me think if Apple ever decides to sell an ARM Mac, that this is the reason. PC and more importantly Mac design is dictated by Intels progress. The Intel slides kind of made me laugh, with their reference designs and “journey to fanless” Intel acts like they’re the ones who pioneered the ultra thin ultra mobile PC and Tablets. While its true that the Air wouldn’t have been possible without the special chip Intel designed, it was APPLE’S desire and perspiration that led to the Macbook Air form factor. Equally with tablets, Intel sat on their asses for 10 years during the early 00s, and it was Apple 10 years later that kick-started the entire tablet market today. Without Intel, using their own skills.

    I imagine Apple has bigger ambitions for Macbooks, but is ultimately hampered by whatever Intel decides is their latest priority. And their latest priority is only a priority for them because of how profoundly Apple has disrupted the tech landscape. Not of their own ambition, but rather a response to a market trend that set sail 5 years ago.

  • AppleWatcher

    “turn Apple’s head”..?? For now, Apple is an Intel customer. They will take the most appropriate, newest chips Intel has available for their products, so there’s no head turning needed. No doubt the delays in getting the 14nm process up to volume production for Broadwell will have annoyed Apple, probably enough to have them running Mac OS on ARM chips in their R&D labs although that was probably already happening anyway. They’re doing a lot of chip design with the recent A-series chips and have made investments in chip foundries like TSMC. It’s probably just a matter of time until Apple is controller of its own destiny in processors but until then, they’ll just keep using Intel chips without needing their heads turned.

  • http://geekfun.com/ Erik S.

    The cluelessness is strong in that piece.

    The first ‘graph in the quote is constructed in a way that suggests that the size of the chip part of whats going to enable thinner devices. Thats just silly. On the other hand, it misses that the smaller feature size is a significant contributor to the lower power consumption of the Broadwell.

    It then goes on to write about the possibilities this opens up as if thin tablets without the bulk of fans have never been seen before, when, in fact, they dominate the market for tablets.

    What’s more interesting is that Intel has been trying to crack the mobile market for years (though it started years too late) and it still hasn’t made a meaningful dent in it, despite being willing to practically give chips away to get design wins outside of its traditional strength in PCs running Windows. Moreover, by most accounts, Broadwell is LATE.

    Also, please, don’t call Broadwell, or any other intel CPU a “chipset.” Traditionally, the chipset involved two or more chips, and supported the CPU by connecting it to other key subsystems, like memory, and peripherals. These days, more and more functions are being integrated into the CPU itself, but that doesn’t make it a chipset.