Apple looking at Intel MID platform for the iPhone

At Computex, Intel demonstrated its Mobile Internet Device concept and its “Menlow” MID platform. Menlow will include a 45nm “Silverthorne” x86 processor, a “Poulsbo” chipset, and WiMax support. The company had some MID prototypes on display, which were running Windows Vista, displaying web pages, and playing videos.

A report by DigiTimes now suggests Apple may be considering Menlow’s successor for the iPhone. The site quotes “OEM sources” as saying Apple is mulling the “Moorestown” MID platform, which will replace Menlow in 2009, for next-generation iPhones. Current iPhones use ARM processors, according to ARM CEO Warren East. Channel sources tell DigiTimes Apple would “reposition the MID market place” if it chose Moorestown for a future iPhone.

Comments closed
    • UberGerbil
    • 12 years ago

    It would be a significant achievement for Intel to get the x86 architecture operating in that niche. Having an x86 chip with reasonable perf/watt for handsets opens up a huge swath of other embedded applications

    • Gandhi
    • 12 years ago

    Given the general secretive nature at Apple, I seriously doubt Apple is going and chatting it up with OEM makers about its future plans. OEMs assemble what Apple tells them to assemble.

    Apple may very well be looking at the Intel Menlow platform (as they should be), but I doubt any Apple engineer has talked about this to any one in China. 2009 is a long ways away – a lot can happen between now and then. Doubt these OEMs are speaking from first-hand knowledge and are just guessing on Apple’s plans.

      • ssidbroadcast
      • 12 years ago

      I like the Samsung-made ARM’s and NANDs in the current iPhone–if it ain’t broke(or unlocked,) don’t fix it. That said, I can see why Apple would consider this: WiMax. If WiMax does as good as analysts are hoping, then it might be a sensible move. Doesn’t Intel own the patent on WiMax? If so, a WiMax based iPhone would basically /[

        • UberGerbil
        • 12 years ago

        1. There isn’t a single “the” WiMax patent
        2. WiMax is 802.16 — it’s a standard, like the other IEEE 802.x technologies, and is not controlled by Intel
        3. Intel isn’t the sole source of WiMax chipsets: Motorola and TI have a solution, and there are a bunch of other, smaller companies — any of whom would jump at the chance to get in a future iPhone.

        Intel has a lot of influence, and sees an advantage to being the first out of the gate pushing WiMax hard, but if Apple cares about WiMax they don’t need to switch to x86, and they don’t need to get the tech from Intel.

        The reason Apple would want to move off ARM is to avoid having to maintain two separate targets for their codebase. That’s a significant reduction in effort (eg, any fancy SSE work they might do to accelerate, say, video for Macs is useless for the iPhone and must be recoded).

          • ssidbroadcast
          • 12 years ago

          Didn’t know WiMax was a IEEE standard. Thanks for the clarification.

          l[

            • UberGerbil
            • 12 years ago

            Yeah, ARM are very efficient. x86 gets its performance by doing a lot of tricks (cracking the instructions into micro-ops, reordering them, dispatching them in parallel, predicting branches, detecting memory access patterns for prefectching, etc etc) all of which use power. The x86 chips also have huge caches. Your typical ARM chip has a few K of cache, runs at a few hundred MHz, and has only a few tens of thousands of transistors. Consequently they only need a couple of W to run. However, in demanding applications like the iPhone they typically run one or more coprocessors to get adequate performance in things like video. ARM chips generally aren’t fabbed at industry-leading process nodes, so Intel does have that advantage.

            Intel claims their Silverthorne chip will require just 10% of the power that the Pentium M did. Even if they pull that off (which would be a huge achievement), they’re almost certainly needing more juice than ARM, but it may be close enough that the other factors make it worthwhile for Apple. Or maybe not: it’s all just vapor ATM.

            • Flying Fox
            • 12 years ago

            I think it is just Apple doing its homework and exploring all options. Intel’s now-accelerated strategy in bringing out new micro-architectures is going to make them improve quickly if they can maintain the pace. Coupled with better processes who knows if they may beat ARM in a few years.

            Of course, ARM can kick it up a notch and still maintain the lead. Competition is good.

          • d2brothe
          • 12 years ago

          A good point about the portability, but I don’t know how much of an improvement it would be. I don’t know anything about this “silverthorne” processor, it says its x86, but that doesn’t mean it supports SSE.

            • UberGerbil
            • 12 years ago

            SSE was just an example (though it’s almost certainly included in Silverthorne). There’s just way more pieces to plug into the toolchain on the x86 side.

          • dragmor
          • 12 years ago

          Except that all of the IEEE 802.x standards use patents owned by the CSIRO which did not has not signed a Letter of Assurance for the IEEE standard past 802.a, but the IEEE went ahead and used them anyway.

          The CSIRO recently won an injunction case vs Buffalo and are now in a lawsuit against Dell, Intel, Microsoft, HP, and Netgear. Funny what happens when the government says its research arm has to make a profit.

    • Damage
    • 12 years ago

    iPhone + good hardware + x86 compatible = sweet

Pin It on Pinterest

Share This