Recent

% | $
Quotes you view appear here for quick access.

Intel Corporation Message Board

  • wallisweaver wallisweaver May 3, 2013 8:29 PM Flag

    Intel: Silvermont - The ARM Killer

    Silvermont, which is the new, 22nm replacement for the five year old Atom design will be introduced on Monday, May 6, 2013 .

    Since the Silvermont is the new Atom architecture, it will be the foundation for a family of SoCs (System on a Chip) with different names. So Monday, we will actually hear about a SoC called Bay Trail, the heart of which will be the Silvermont CPU.

    The BayTrail is a quad Silvermont core version SoC, which will be aimed directly at replacing the ARM (ARMH) based SoCs used in tablet computers. The manufacturers in the cross hairs are Samsung (SSNLF.PK) (including the Apple (AAPL) fab versions of ARM derived parts), Qualcomm (QCOM), Nvidia (NVDA) and other sellers of ARM based chips for use in tablet computers.

    At 22nm and chronologically the newest design available, Bay Trail will outperform any ARM based chip design while delivering longer battery life.

    There is some subtle genius in Intel making tablets the first serious target in mobile devices. First, only an x86 device can run full Windows 8, we're not talking about Window RT, which is an abbreviated, and largely failed, version of Window 8. Bay Trail will run full up Win8.

    Above excerpt from article by Russ Fischer

    SortNewest  |  Oldest  |  Most Replied Expand all replies
    • Intel will wage its battle for tablets on two fronts, 1). Overall performance and battery life, and 2). Software flexibility giving the consumer access to real Windows and its millions of apps, plus Android and its apps.
      In all probability between Haswell and Bay Trail Intel will enjoy performance superiority across the entire range of tablet price points, but even if Intel's performance is only roughly equal to ARM tablets Intel still has the huge advantage of being able to give consumers access to all OSes and their apps. The latter puts the purchase decision on a qualitative "It runs Windows or not" question. This is something the consumer easily understands and the debate about benchmarks don't even matter.

      Now if Intel wins the performance benchmarks too, it won't really be much of a choice.

    • The reason why this is not true is not anything to do with the Silvermont core, but with the SoCs that have been built around this core. They lack integrated 4G/LTE and hence can't compete with Qualcomm's Snapdragon which does. For a smartphone or tablet vendor shipping to the US market 4G/LTE is a critical feature. They only way Intel can provide the functionality is through separate standalone modems which are power hogs. This shows you an example of why ARM's IP licensing model makes better business sense.

      • 1 Reply to khitchdee
      • 'For a smartphone or tablet vendor shipping to the US market 4G/LTE is a critical feature. They only way Intel can provide the functionality is through separate standalone modems which are power hogs.'

        Rubbish. It would be the same functionality just on another chip and there is nothing inherently power wasteful about that no matter how many times you spout this fallacy. Intel (Infineon) modems are known for their low power usage anyway. Modems burn less power than the application processor as well so it is not an important point in the first place.

        'This shows you an example of why ARM's IP licensing model makes better business sense.'

        What an idiotic statement. Whether a modem is on the same chip as the application processor has absolutely nothing to do with its architecture as Intel modems are still ARM based for historical reasons i.e. that is what Infineon used. If you did not spout nonsense you would not spout any sense at all !

    • The next Intel move in mobile will be with the Merrifield SOC. The Merrifield is expected to be available by the end of the year. The Merrifield is dual Silvermont core SoC that is optimized for smartphone use.

      The existing 32nm Clovertrail SoC, already beats ARM based phones on almost every benchmark. The 22nm Merrifield chip will add to performance and battery life.

      The smartphone segment presents problems not present in the tablet segment. In smartphones, nearly two-thirds of the chip business is locked up with Samsung and Apple. The SoCs with these two companies are captive in that the Apple device is a custom chip manufactured by the arch Apple enemy, Samsung. And of course, Samsung SoCs are manufactured by, surprise, Samsung.

      Penetrating the smartphone business will be a little trickier, but even Apple and Samsung will have to adopt a superior technology if they are to remain competitive in the smartphone market.

      Intel will need a lead customer to get the conversion process started. That lead customer could well be Google/Motorola with their X-phone, which is to be introduced later this year. Intel and Google (GOOG) have worked together for nearly two years to optimize Android to work with x86 processors. The higher performance and longer battery life of Merrifield in an Xphone will cost Samsung and Apple smartphone sales.

      • 1 Reply to wallisweaver
      • The existing 32nm Clovertrail SoC, already beats ARM based phones on almost every benchmark. The 22nm Merrifield chip will add to performance and battery life.
        ----

        No. There is only one 32nm Clovertrail SoC in the market place today. Once Clovertrail+ arrives (in phones) we'll have a better test environment to compare and it will be clear that ARM SoC's will beat in it in Javascript, Java, Geekbench etc.

 
INTC
33.45-0.10(-0.30%)May 22 3:59 PMEDT