Recent

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

Intel Corporation Message Board

  • stocktrade14 stocktrade14 Oct 5, 2012 7:55 PM Flag

    Apple's A6 processor is good for Intel

    No doubt Apple's A6 processor is superb. It is reasonable to assume that it is 3X faster than A5 while maintaining the same battery life.

    But this is going to be a problem for rest of the ARM camp. They won't be able to match A6 performance easilty. This is where Intel has opportunity to step in.

    Then how much superb is A6? Let us do the maths.

    A6 shrinks to 32nm from 45nm in A5. That makes A6 1.4X fastener.
    A6 possibly has 1.5X more transistors. Since I don’t know exactly, assume 1.3X. Another 1.3X speed.
    32nm enables A6 to run at 1.6X clock speed. Another 1.6X bump

    Total speed bump =1.4*1.3*1.6 = 2.9

    So Just about 5% improvement from internal design efforts!

    Now 3X speed with same battery life? Does it mean 3X efficiency? No.

    Suppose you have got 2X faster car. It will get you to workplace 2X faster. But assuming same MPG and tank capacity, and you drive the new car for the same purpose, both old and new cars will last same number of days between tank fillings. No doubt, 3X speed will improve user experience.

    Similar move to 32/28nm by rest of the ARM camp will reap them similar benefits. But this is a comfort zone for Intel, not for the ARM camp.

    With recent Razr Vs Razr i comparison, I feel 32nm Intel is outperforming 28nm Snapdragon, though it may still have some shortcomings that will be overcome in few months. Not only the node size, but the maturity and quality of process matters.

    SortNewest  |  Oldest  |  Most Replied Expand all replies
    • Apple licenses the processor designs from ARM and then makes a system-on-chip.

      The processors are not pushing higher processor speeds but going to multiple cores so that the operating system runs faster from threading of the instructions to the multiple processors.

      But see, an ARM tablet is mostly for just connecting to websites. And an application on an ARM tablet is often just a connection to a website.

      But here are my questions. Will the consumer connect to their financial account websites while using their ARM tablet on a WiFi network ? And how does the consumer use their ARM tablet at home ?

    • as long as the stock price dosnt shrink- 4%div with some upside.
      i will be a happy camper.

    • Even with their process lead, the A6 will be difficult to match for Intel, next year even, as long as they stay x86 based. The only way they can match is if they adopt ARM architecture, which I doubt they will do. My best case prediction for them is they come out slightly ahead of the A6 in 2013 with their 22nm OO chip and Apple will retake the lead 6 months later. Slightly ahead will not warrant a switch by either Apple or Samsung to IA. To stay in business, that's the kind of design win Intel needs. Otherwise, they're going to spend a lot of money and get very little market response. It's only now a matter of when they decide to can this whole low power IA effort.

      • 3 Replies to khitchdee
      • "Even with their process lead, the A6 will be difficult to match for Intel, next year even, as long as they stay x86 based. The only way they can match is if they adopt ARM architecture, which I doubt they will do."

        [You must not have seen the Razor comparison showing that Intel already outperforms ARM, even using 32nm. You just keep spouting the same old stupidity...]

        "My best case prediction for them is they come out slightly ahead of the A6 in 2013 with their 22nm OO chip and Apple will retake the lead 6 months later."

        [Yeah the prediction from the guy who demonstrates daily that his knowledge of fabrication will fit on the head of a pin. Intel will easily be out in front with 22nm and hugely out in front with 14nm. It's all about the fabrication. Intel has it. ARM doesn't. Intel's fabrication is ramping up while ARM's coasts to a stop. ARM has nothing after 20nm - if they ever get there...]

        "To stay in business, that's the kind of design win Intel needs. Otherwise, they're going to spend a lot of money and get very little market response. It's only now a matter of when they decide to can this whole low power IA effort."

        [To stay in business. LOL. Intel is going to offer the best performance and the greatest power efficiency. And there is nothing ARM can do about it. After 2013 there will never be another reason to buy an ARM processor. ARM will be the next AMD - relegated to the low-end price value bargain basement section that Intel doesn't want. Intel has already won the low-power war. We're just waiting for it to play out. All about the fabrication, fanboi. All about the fabrication...]

        Sentiment: Strong Buy

      • Read the comparisons between the razor i and razor m on the wedsites Engadget and The Verge. Both show an advantage for Intel in both performance and energy efficiency and thats with a 32mm single core medifield chip.

      • For the umpteenth time............................Intel does NOT need Apple mobile business to succeed. It now has access to the Droid Market and Droids outsell Apple phones !!!!!!!!! Intel just needs to do what it does best and even if all it ever does is do "just as well" as Apple in chips, that will be fine because they'll be inside a zillion US-based Droids !

    • A6 shrinks to 32nm from 45nm in A5. That makes A6 1.4X fastener.
      A6 possibly has 1.5X more transistors. Since I don’t know exactly, assume 1.3X. Another 1.3X speed.
      32nm enables A6 to run at 1.6X clock speed. Another 1.6X bump

      Total speed bump =1.4*1.3*1.6 = 2.9
      ------

      Shrinks by themselves do not increase the speed of a processor. More transistors in the SoC do not increase the speed of the processor unless the transistors have been spent on the CPU. The A6 runs at 1Ghz while the A5 @ runs at 800 mhz.

      The A6 has around double the raw performance of the A5. This increase in performance is due to the new architecture, increased clock speed and a better memory interface.

      As for the rest of the ARM camp, the A15 can deliver this kind of performance (depending who/what you read, it will be faster) but the real question over the A15 is it's size (transistors) and power consumption. There is still a lot of leg room in the existing A9, with improved process tweaks (even at 40nm) meaning higher clocks for the same power envelopes, improvements to memory bandwidth etc

 
INTC
28.64+0.42(+1.49%)Feb 12 4:00 PMEST