Intel Could Succeed in the Android Market with HDRC

Posted on 31 March 2012 by



Back at IDF September, Intel and Google finally announced that they’d be working together to get Android up and running on x86 devices. While there were a number of Android-running x86 tablets and a smartphone prototype or two floating around IDF, it wasn’t immediately apparent what the major advantage of Android 0n x86 devices would be for your everyday consumer. In fact, it wasn’t even apparent exactly why any of the existing Android manufactures would  want to create x86 Android devices, given that up until now, pretty much all of their R&D has been focused on ARM devices. However, Intel may actually be perfectly positioned to be able to stimulate the growth of an upcoming segment of Android device — one which truly converges mobile and desktop functionality into one device. Chippy has coined such hybrid functionality: ‘High Dynamic Range Computing’ (HDRC), and the time might just be right for Intel to ignite this segment and find their own place in the Android market.

Before moving on, you might want to visit this link to see Chippy’s look at HDRC from last year.

Any consumer-available Android device that you can get your hands on today uses ARM architecture which is fundamentally incompatible with the x86 architecture that Intel products are based on. Android was originally built to run exclusively on ARM (though being open-source, some community projects were able to do some porting to x86). It wasn’t until several years after Android was on the scene that Intel and Google finally got together to work on full hardware-level Android on x86 support. That work is still ongoing. We’ve had our hands on Android devices running with Intel’s x86 architecture, but it is clear that there is still much optimization to be done. Once everything is complete though, won’t a device running Android on ARM be, for the user, indistinguishable from a device running Android on Intel’s x86?

If ARM has battery life, Intel has power. It’s an interesting dichotomy — we’ve watched as ARM-based devices have continuously scaled up to meet performance demands as the Android device market has grown. Intel has the opposite problem; they’ve got power, but have been constantly trying to scale it down to work with mobile at the tablet/smartphone level. Intel’s Atom series is a notable effort in the last several years to scale things back far enough that users could get reasonable performance and reasonable battery life out of a netbook. Once Intel can achieve the same thing at the smartphone and tablet level (and they’ve been working on this for years), they’ve got the expertise to push the processing end of things far beyond what we currently see from ARM — not to mention that the same x86 architecture that will be found in Intel-based phones and tablets is capable of booting full-fledged desktop operating systems.

If Intel plays their cards right, they could do very well in the Android market by stimulating the HDRC segment. HDRC isn’t really a mainstream thing at this point — most people have their desktop computer and they’ve got a smartphone and maybe a tablet. They view these two devices as fundamentally different. The promise of HDRC is creating a device that scales so well that it can converge these two categories of devices, which are viewed as different, into a single unit. This is a serious challenge because essentially it asks for a single device that is instant-on and has phone-like (all day) battery life, but, when plugged in, can be as powerful as one would expect from a laptop or desktop. Intel has the expertise for the high-end of the HDRC spectrum, we see this daily from the desktop computers that we work on. If they can combine this with phone/tablet-like low-power functionality, they could blow ARM out of the water and define the HDRC space that mobile technology has been steadily moving toward for the last 5 years.

(continue reading on page 2…)

Pages: 1 2

Categorized | Opinion

Tags : , ,

1 Comments For This Post

  1. Realty says:

    Ben, interesting read

    I think you underestimate a few things.

    1). Microsoft really is trying to bring Windows to ARM.

    2). ARM is not sitting still and their newest designs A15? may be as powerful as Intel if not more so.

    3). Ease of use is important but price will be more so. There is more to running a Windows OS than merely using a X86 chip instead of ARM. It’s the other necessary electronics that have made past Windows tablets more expensive than comparable Android tablets. If Android tablets are being sold for $250 and Windows tablets for $700, the Windows tablets will have a tough sell in the consumer market.

    4). Android is heading for the desktop market themselves which means you will not have to dual boot to get a desktop experience. One Android UI will cover both mobile and desktop computing.

    HDRC would be wonderful but you might be able to get it using just ARM processors and maybe at a lower price too.

    James Reply:

    While ARM’s role in the market is increasing there are still issues that have to be overcome…

    1) MS won’t be providing legacy support on ARM and are insisting that there would be no way to unlock secure boot on ARM. So while they are bringing Windows to ARM, there will also be some limitations compared to the x86 implementation.

    2) Neither is Intel standing still, while the new ARM Cortex A15 are the next gen and will dominate 2012 for the low end market that Intel ATOM’s normally fill but in early 2013 Intel is going 22nm and doing the first major architectural update to the ATOM since it was first introduced.

    While right now ARM is having problems with going 28nm and those manufacturing issues could effect them till next year.

    3) The Intel ATOM market has worked with lower profit margin than pretty much the entire rest of the market for years, ARM included, and that offsets the normal cost advantage ARM has… Add Intel going 22nm when ARM will still be 32nm and 28nm could means Intel could match or at least come very close to similar pricing. While still potentially offering even better performance and a wider range of capabilities.

    Even in power efficiency, Medfield already shows Intel has made significant progress at getting nearer to ARM level power efficiency and that’s not even using their next gen technology yet.

    Mind also that Windows 8 will require the top of the line ARM offerings to run well. So pricing is likely not going to be that cheap.

    4) Android may be heading to the desktop market but it’s still a mobile OS and that means limited by design. More separates mobile OS from a desktop OS than just how they look and how the UI works.

    Also ARM is still a 32bit processor that’s only started using 64bit memory management and won’t be able to go fully 64bit for at least a couple more years, at least outside of the server market and alternate solutions like what Nvidia is working on.

    Though the growing number of hybrid design systems could mean a eventually merger between tablets and what has been the netbook market up till now.

    There have already been attempts to make systems that would switch from a ARM Tablet to a laptop when docked. However, MS seems to be developing a patent for a docking system that will allow a device to switch to a more powerful processor when docked. While Intel has access to technology to quickly switch between two OS’es.

    So these may either use a combination of ARM and x86 but seamless switching will likely be easier for a low end x86 docking to a higher end x86 for easier scaling and thus better fit the HDRC goal.

    Realty Reply:

    James,

    You make some interesting arguments. Here are my thoughts point by point.

    Point # 1). We have been trying to discuss the difference between Windows 8 on ARM and Windows on Intel in the forums. The consensus seems to be that the differences are not positive but will only be transitory. If not, then why would anyone make a tablet with the permanently crippled Windows on Arm OS?

    Point # 2). You may be right about the chip manufacturing future. Only time will.

    Point # 3). I think you are dreaming. Dell and HP are not in the business to make computers or tablets, they are in the business to make money. I expect no charity from them to keep Windows tablets or UMPCs price competitive. I think you would have been closer to the truth if you had said, “People will gladly pay a premium for a Windows tablet or UMPC.” I question whether smaller Intel chips will mean cheaper but again time will tell.

    Point # 4). You make good points however I disagree concerning the OS UI. Ben was discussing HDRC which means usage on all form factors. The Android and Metro UIs scale up nicely from 3 inch screens to 24 inch screens. Windows Legacy UI is worthless below a 10 inch screen. (Sorry fellow UMPC lovers but I’m getting old and Legacy Excel on a 5 inch screen does not work for me anymore.) Android has 400,000 apps that can be leveraged up onto a desktop in short order. I question how long it will take the Windows Legacy Apps to be rewritten to the Metro UI? ( I think Microsoft does not see it happening in mass until 2015.) Therefore I will stick by my statement that if you want to see an HDRC device, you might see it first on ARM using Android. This is not a put down on Intel. They are a great company however they have to work with the cards they have been dealt. They may have the chips but if the OS and supporting Metro desktop software is not ready yet, there is not much they can do except try to get in on the Android craze for now. Interesting times.

    animatio Reply:

    short replay on point4)
    not everybody does agree with james point of view. it is not the gui that’s important, its the application and thus kernel support for the control of their main components (window sizing, menus etc). in this regard win 8 is a lousy trickery by microsoft, trying to fore their whole market to change paradigms of windows from desktop to tablet. from precision to fingering, so to say by squezing the desktop into a app window (actually a sort of a virtual machine asset (same trickery they tried to do with XP apps in win7 in a clumsy virtual machine btw, by hiding the truth that there are far better ways to run almost everything – down to the old dos btw – in win7). we shall see if professinal markets will accept this attitude or not. the win XP example tells us another lecture. in fact the paradigm of protection of investment is a straight contradiction to this.

    James Reply:

    Ok…

    1) Part of the problem is we don’t have a clear indication yet of how the final version of Windows 8 will work. So opinions are a little befuddled by what it appears right now they are doing but it’s only a preview and a lot of changes are left for them to make before final release, even if we’re already close to the semi-final beta releases in another month or two… Time will tell though…

    3) No dream, it’s a point of fact that the netbook market is the only one with significantly low profit margins. Many companies in China were making them for pennies worth of profit, which is why a lot of them went out of business.

    Also, it’s one of the reasons why Intel and MS put in so many limitations to keep the netbook market separate from the rest of the laptop market.

    Really, ARM has had a cost and power efficiency advantage for a long time but it’s only when they are finally getting close to providing similar range of performance as Intel ATOM’s that Intel is finally getting serious about the ATOM development.

    Sure, HP, Dell, etc. are going to charge what they charge but for the same reason why ARM tablets are dropping in price so will x86 based mobile devices be prices as low as they can go.

    Originally the ATOM was set on a slow 5 year product cycle that helped ensure all parts would be well vetted and thus as cheap as possible to make. Starting with the 22nm Silvermont though they are officially switching to the same 2 year cycle they put their main chip lines through, putting in new advances practically as soon as they have them available.

    However, Intel knows they need more than just a manufacturing lead to compete in the mobile market and thus the hardware at least will be as low cost as they can manage.

    They’re already doing it with Ultrabooks, basically paying companies to make them and help absorb costs of setting up productions, etc.

    So it has nothing to do with altruism, just simple competition…

    4) Android has legacy as well, older apps not optimized for tablets, etc are being left behind. The difference is just people tend to hold onto legacy apps more for Windows than they would a mobile app that can easily be replaced with a better up to date version.

    With the merger of the Android Kernel with Linux 3.3 it could mean a sharp break from the old apps with a complete re-working of the OS to something that’s more a hybrid mobile and desktop OS, much like how it took a long time to adapt when Android first introduced Honeycomb.

    Features like webtop/web-desktop customization are increasingly being developed and that’s likely how they’ll manage to keep Android relevant as people start using more and more desktop features and capabilities with their mobile devices.

    The only thing is it’ll take nearly as long for Android to make the adaptation as it will take MS.

    However, what I believe you’re missing is that while MS may be limited on ARM for a number more years, there will be much less issue on x86 systems. While as I pointed out before it’s easier for a desktop OS to run both desktop and mobile apps than it is to adapt a mobile OS to start running desktop apps.

    Though it remains to be seen whether Windows 8 will succeed or stumble and thus give Google the time it needs to make Android a viable alternative.

    Thing to keep in mind is Windows doesn’t just have legacy apps going for it but also popular and more capable desktop apps like MS Office, Adobe Photoshop, etc. Along with supporting higher range performance, like gaming laptops, etc. that ARM is still many years from getting into.

    @animatio – Of course the kernel support is important but that’s one of the points of a desktop OS over a mobile one. Even with similar base a mobile OS is designed to for efficiency even at the cost of capabilities and features. While a desktop OS is optimized to offer as many capabilities and features without as much concern for efficiency.

    While either way developers will have to learn both paradigms for developing apps but it’s arguably easier to make more capable apps and features more efficient than it is to take less capable apps and features and expand them as that usually requires a complete redesign.

    Mind up till now Android had been working with a stripped down version of a no longer supported version of the Linux Kernel. It’s only with Linux 3.3 that they re-merged the Android Kernel with Linux and that likely means they’ll be trying for their own version of a hybrid OS like MS is doing with Windows 8 and like MS they’re running their own risks as to whether it’ll be good enough or not.

    While the basic trick for either of them is to get developers to easily be able to develop scalable apps and not just alternative apps.

    Given the Android market fragmentation, it may not be as easy for them to adapt but time will tell.