Google looking for chip partners to enable Android to compete with Apple's A9 chip

1246

Comments

  • Reply 61 of 108
    dasanman69dasanman69 Posts: 13,002member
    mstone wrote: »
    gatorguy wrote: »
    They're also on their way to commodity status over the next few years. 
    Maybe some phones but not iPhones. That is why you see such attention to the case design and materials. People will continue to pay for quality and luxury. Almost all Android phones are already a commodity.

    The Android OEMs is whom was being discussed.
  • Reply 62 of 108
    dasanman69dasanman69 Posts: 13,002member
    herbivore wrote: »
    If Google really wanted to follow the Wintel model, they should have focused entirely on the OS. They should have worked more closely with Samsung as a hardware manufacturer.

    They have emasculated Samsung who was and continues to remain the most competent hardware manufacturer in the business outside of Apple. If Samsung exits the smartphone business, a Android will become a total mess as the remaining companies are simply in a race to the bottom. At least Samsung is actually trying to release a high end and compelling product. But Samsung's priorities and google's priorities really don't mesh.

    Google should have had the foresight to see that they really need to have a strong hardware manufacturer. Else Android phones will be perceived as junk like they currently are in China. So much so that Samsung, which makes high quality hardware is taking a beating, because of that perception while Apple phones are viewed as high quality and aspirational products. Now Google is trying to eliminate all distinction between their Android products and make them all inexpensive and "cheap." How is Samsung or any company who wants to make high end hardware supposed to compete? Google doesn't care as getting their handsets into as many users as possible will bring more eyes and more advertising dollars. Except that people who don't have money can't spend any and therefore poor advertising targets.

    They are now realizing that Android phones are being perceived as junk. The only hardware outside of Apple that I find enticing are Samsung and LG. But Google is attacking the individual brands. They want everything to be known as Google's Android only with the actual manufacturer a non-entity. Why any handset manufacturer would want to be a part of that ecosystem defies logic.

    What Samsung needs to do, is bury their animosity with LG and perhaps bring Sony and HTC to the table. They need to pool their resources and turn either WebOS or Tizen into a viable platform that is able to run Android applications. Or perhaps incorporate an Android runtime environment that can operate Android apps. They can then design their own CPUs and build in high end functionality into their handsets. They can leave Google and perhaps entice the developer community to build apps for their own OS.

    Huawei and Xiami phones are junk. I love the OLED screens of Samsung and LG along with their cameras. However, as long as they run the spyware OS known as Android, they will never be considered for purchase by me. Since the manufacturers are interested in winning over customers that have more in the way if disposable income, it would behoove them to leave Android and develop an alternative platform.

    Google is going to contract. That is an inevitability. Their model is no sustainable. What is happening in China will be duplicated elsewhere. India won't save them. And neither will attempting to get into processor design. Intel x86 CPUs will never be cost competitive. They will be forced to do it on the ARM platform. Maybe Qualcomm or Mediatek would be willing to work with Google. So what happens when Google's new processor gets fabbed on TSMC's manufacturing lines? Does Apple take it's considerable business and move it all back to Samsung? Samsung could manufacture such a CPU, but why would they want to? The low end phones are destroying Samsung's handset business. Perhaps Intel will do it with Google providing considerable contrarevenue of their own. Maybe Global Foundries?

    None of this makes sense. But it does show that Google is getting desperate. And how smart technical minds can be extremely poor businessmen.

    Samsung smartphones didn't get really popular until 2012 which was a few years after Android was released? So why would they have chosen them?
  • Reply 63 of 108
    gatorguygatorguy Posts: 24,095member
    herbivore wrote: »

    What Samsung needs to do, is bury their animosity with LG and perhaps bring Sony and HTC to the table. They need to pool their resources and turn either WebOS or Tizen into a viable platform that is able to run Android applications. Or perhaps incorporate an Android runtime environment that can operate Android apps. They can then design their own CPUs and build in high end functionality into their handsets. They can leave Google and perhaps entice the developer community to build apps for their own OS.
    Personally I don't think it would be worth either the time or investment to do so. For what purpose? You really think they could steal premium share from Apple over the next several years by starting yet another mobile OS, develop some compelling functionality that no one else has, spend the dollars on enticing developers to concentrate on a third OS/app store. . .

    Sammy already tried most of that with Tizen. Remember them offering substantial payouts to developers who would bring apps to their proposed platform and the single Tizen developer conference they had a couple years back? It appears they've changed their mind (Milk Music, their Apple Music/Google Music/Pandora alternative is gone too), The well-heeled Microsoft is not succeeding in attracting huge developer interest on their mobile platform either are they?
  • Reply 64 of 108
    tmaytmay Posts: 6,247member
    Quote:

    Originally Posted by Gatorguy View Post





    They're also on their way to commodity status over the next few year.

    Perhaps it is better to be a valued brand, "commoditized" in a premium tier, not the low or middle?

     

    In the meantime, Apple continues piling up cash while Google and MS really haven't seen much at all in the mobile space.

     

    Google bought term life with Android; Apple looks to be the great investor in the mobile space.

  • Reply 65 of 108
    dasanman69dasanman69 Posts: 13,002member
    tmay wrote: »
    gatorguy wrote: »
    They're also on their way to commodity status over the next few year.
    Perhaps it is better to be a valued brand, "commoditized" in a premium tier, not the low or middle?

    In the meantime, Apple continues piling up cash while Google and MS really haven't seen much at all in the mobile space.

    Google bought term life with Android; Apple looks to be the great investor in the mobile space.

    And how do you suggest someone become a valued brand in the premium tier? None of the existing OEMs can do it. There would have to emerge a new player, and it would take that player at least 10 years before they can even begin to take market share from Apple.
  • Reply 66 of 108
    mstonemstone Posts: 11,510member
    Quote:

    Originally Posted by dasanman69 View Post





    The Android OEMs is whom was being discussed.



    Perhaps but as I read his post, Gatorguy was suggesting that all smartphones would become commodities which is why Apple was branching out into so many other markets. 

  • Reply 67 of 108
    foggyhillfoggyhill Posts: 4,767member
    Quote:
    Originally Posted by RobertC View Post

     

     

    Samsung's Exynos 7420 was the most powerful phone SoC on the market when it launched in early 2015. 

     

    Apple's A9 now holds that title, but I'm sure the cycle will continue.


     

    Hey, bud, it throttles to death in 2 seconds; at Anandtech it doesn't even beat the A8 on many tests.

    Why on earth are you here?

    Quote:

    Originally Posted by tmay View Post

     

    Most apps are single threaded; so most will benefit from a high single core. Not many apps will use all 8 cores, and for a fact, the only 8 core SoC's are big / little implementations, so wasteful of die.

     

    Thanks always for showing the old Tegra Love; but you do know that nVidia is no longer in mobile; likely as the Tegra was a hot running SoC.

     

    Compare that with A 9 which hardly throttles in even heavy gaming; the definition of balanced design.

     

    Tegra sucks for mobile, every generation was power hungry,  so bringing that argument is B.S.

     

    Edit,

     

    So I read your Anandtech link; basically it only demonstrates that Android can take advantage of multiple cores; it doesn't really answer the question of what is actually better in actual use, nor is any of it applicable to iOS. I will retract my "wasteful of die" statement; it would appear that this is a successful strategy for Mediatek which sells to the low and midrange OEM's.


     

    The thing is that "using the core" and actually loading the core without waste, is not the same thing.

     

    Say, you have identical performing cores: one system has 1 core, the other has 4 of these identical cores.

     

    If you use every core at 30% and waste 5% in overhead per core, you'd get the a performance from 4 core that's the SAME as 1 core loaded at 100%. You'd get a bit less latency with 4 cores (because the 100% loaded core needs can't process everything at the same time), but that's it.

     

    The problem is, that Apple's core are not the same.... They're much much faster; that takes care of the latency advantage of multicore and leaves them needing high loads spread accross many high clocked cores just to compete!  The problem is that apple can still add cores, and even even a bit more clock, while on the Android side, things are getting more and more like space heaters.

  • Reply 68 of 108
    robertcrobertc Posts: 118member
    Quote:
    Originally Posted by tmay View Post

     

    Most apps are single threaded; so most will benefit from a high single core. Not many apps will use all 8 cores, and for a fact, the only 8 core SoC's are big / little implementations, so wasteful of die.

     

    Thanks always for showing the old Tegra Love; but you do know that nVidia is no longer in mobile; likely as the Tegra was a hot running SoC.

     

    Compare that with A 9 which hardly throttles in even heavy gaming; the definition of balanced design.

     

    Tegra sucks for mobile, every generation was power hungry,  so bringing that argument is B.S.

     

    Edit,

     

    So I read your Anandtech link; basically it only demonstrates that Android can take advantage of multiple cores; it doesn't really answer the question of what is actually better in actual use, nor is any of it applicable to iOS. I will retract my "wasteful of die" statement; it would appear that this is a successful strategy for Mediatek which sells to the low and midrange OEM's.


    That's precisely the point. I made it very clear in my original post.

     

    As for nVidia, they haven't made an SoC for smartphones in years, but for tablets, Google's Pixel C will be using Tegra X1. 

     

    Despite Tegra K1 launching months before the iPad Air 2, and being a whole node (28 nm) behind Apple (20 nm), the performance degradation of Tegra devices was not at all how you make it out to be.

    Unlike Apple's SoC's, the Tegra K1 (and newer Tegra X1) have had full API and feature support. This part is what gives it a huge advantage in an area that nVidia is an expert in, gaming.

  • Reply 69 of 108
    robertcrobertc Posts: 118member
    Quote:
    Originally Posted by foggyhill View Post

     

     

    at Anandtech it doesn't even beat the A8 on many tests.

     


     

  • Reply 70 of 108
    dasanman69 wrote: »
    Samsung smartphones didn't get really popular until 2012 which was a few years after Android was released? So why would they have chosen them?

    Samsung had rapidly ascended to the top of the Android pack. Samsung has nearly matched Intel's manufacturing process and makes the leading high end Android handset. It would make the most sense for Google to work with Samsung as a result. Huawei makes an inexpensive handset and doesn't include Google Play any how. Spending the money to develop SOC functionality to sell a low end and cheap device makes little sense. People who purchase cheap devices aren't who the advertisers are going to want to target anyhow. Samsung is clearly targeting a high end demographic. If Google wants to get Android into the hands of high end users, working with Samsung is the logical choice. It's really their only choice.

    If Google doesn't do anything about the situation, Samsung will exit the handset business and it becomes a race to the bottom. The last company who survives will become an Android monopoly anyway and will dictate terms to Google. The very situation they fear about Samsung.
  • Reply 71 of 108
    pistis wrote: »
    You say that there is not sweet spot for Android phones manufacturers and later say "Android can only succeed if it's installed on huge numbers of devices", huh ? Don't you see these two statements are contrary to each other?

    if they can only succeed if its installed on huge numbers of devices then that is by definition it's sweet spot!  but as we know the race to the bottom is not working so logically it cannot succeed. 

    I think you are a very confused person - do you work at Google?

    You totally missed the point of a very well written post. Those two contradictory statements define the problem of the Android handset market. That's why it's going no where.

    I particularly liked how the author, AlanAudio, described Apple's actions as creating "the perfect storm" for the whole Android/Google gang. The funny part of all that is that Apple is not going to let up and will continue to pull away from that pack of clowns.
  • Reply 72 of 108
    tmaytmay Posts: 6,247member
    Quote:
    Originally Posted by RobertC View Post

     

     


    I fear you will be very saddened with actual vs extrapolated performance of the Tegra X1. I at least have full specs available today for the iPad Pro, and with its release in a week, we will start seeing reviews and benchmarks. To date, there is no weight, no battery size, and vague 10 hr video performance metric. I've read reviews on the existing Shield Tablet, and battery life was not good, but at least it ditched the fan from the previous Shileld.

     

    I'm getting that past Tegra feeling; lots of hype, little in the way of performance metrics.I'm calling this early; I don't think that the Pixel C will demonstrate performance benchmarks in any respect better than the iPad Pro, including its much touted graphics performance for gaming.

  • Reply 73 of 108
    mstone wrote: »
    dasanman69 wrote: »
    The Android OEMs is whom was being discussed.


    Perhaps but as I read his post, Gatorguy was suggesting that all smartphones would become commodities which is why Apple was branching out into so many other markets. 

    In a hundred years...maybe a bit less.... all smartphones may become commodities. However I think Apple, like many successful companies, wants to have multiple streams of income. Apple is too dependent on iPhone profits, and one or two other equally successful products would prevent them from being hurt financially if somehow that product hit a speed bump.
  • Reply 74 of 108
    gatorguy wrote: »
    Personally I don't think it would be worth either the time or investment to do so. For what purpose? You really think they could steal premium share from Apple over the next several years by starting yet another mobile OS, develop some compelling functionality that no one else has, spend the dollars on enticing developers to concentrate on a third OS/app store. . .

    Sammy already tried most of that with Tizen. Remember them offering substantial payouts to developers who would bring apps to their proposed platform and the single Tizen developer conference they had a couple years back? It appears they've changed their mind (Milk Music, their Apple Music/Google Music/Pandora alternative is gone too), The well-heeled Microsoft is not succeeding in attracting huge developer interest on their mobile platform either are they?

    Microsoft's plans are to include an Android runtime environment in Windows. Such an inclusion would increase the desire for Windows as a platform that can run legacy programs along with the multitude of Android applications. A sufficiently powerful mobile x86 processor running a Windows phone that can run the majority of Android applications would be quite compelling. It could really make things difficult for Google and their hardware manufacturers. Especially on commodity hardware.

    It becomes clear why Google would want to develop a CPU using the ARM design with unique functionality. It would mitigate what MSFT is attempting to do. But let's face it, such a chip would have to be manufactured by TSMC or Samsung. It's very doubtful that either of them would want to risk losing Apple's business. And Intel is loathe to manufacture anything non x86 on their fabs. Guess that leaves Global Foundries.

    Android doesn't confer any particular advantage to Google and could still be easily usurped. Unlike iOS where the hardware allows for unique application functionality. Microsoft cannot usurp iOS apps.

    And I'm not at all convinced that Samsung, cannot establish another platform. Especially in conjunction other high end manufacturers such as LG and Sony. Yes, they tried and failed once. They really need to try again. Else they will be rendered irrelevant like the rest of the Android ecosystem by MSFT of all companies.
  • Reply 75 of 108
    tmay wrote: »

    I am really disappointed that MS has essentially failed with ARM, as it is to me a competitive advantage in mobile versus X86, but that ship has sailed.

    MS with Win10 wants to stuff a desktop OS into a phone-sized package... and just because it can be done, doesn't mean it should. As time wears on and portability (weight and size) and battery life become more important, the ARM chip and a matching OS will carry the day...and the Intel/MS strategy will hit limits way before ARM/iOS will.

    The former is just bull-headed thinking, while Apple's way is imaginative thinking.
  • Reply 76 of 108
    tmaytmay Posts: 6,247member
    Quote:

    Originally Posted by herbivore View Post





    Microsoft's plans are to include an Android runtime environment in Windows. Such an inclusion would increase the desire for Windows as a platform that can run legacy programs along with the multitude of Android applications. A sufficiently powerful mobile x86 processor running a Windows phone that can run the majority of Android applications would be quite compelling. It could really make things difficult for Google and their hardware manufacturers. Especially on commodity hardware.



    It becomes clear why Google would want to develop a CPU using the ARM design with unique functionality. It would mitigate what MSFT is attempting to do. But let's face it, such a chip would have to be manufactured by TSMC or Samsung. It's very doubtful that either of them would want to risk losing Apple's business. And Intel is loathe to manufacture anything non x86 on their fabs. Guess that leaves Global Foundries.



    Android doesn't confer any particular advantage to Google and could still be easily usurped. Unlike iOS where the hardware allows for unique application functionality. Microsoft cannot usurp iOS apps.



    And I'm not at all convinced that Samsung, cannot establish another platform. Especially in conjunction other high end manufacturers such as LG and Sony. Yes, they tried and failed once. They really need to try again. Else they will be rendered irrelevant like the rest of the Android ecosystem by MSFT of all companies.

    A simple solution;

     

    Google should pay unprofitable and marginally profitable OEM's  not to make phone at all; a few billion dollars gets spread far and wide in the Android world, and Google could then repurpose Android OS as proprietary with a just few OEM's, including Samsung, HTC, Moto and Sony.

     

    I kid...

  • Reply 77 of 108
    Quote:
    Originally Posted by tmay View Post

     

    I fear you will be very saddened with actual vs extrapolated performance of the Tegra X1. I at least have full specs available today for the iPad Pro, and with its release in a week, we will start seeing reviews and benchmarks. To date, there is no weight, no battery size, and vague 10 hr video performance metric. I've read reviews on the existing Shield Tablet, and battery life was not good, but at least it ditched the fan from the previous Shileld.

     

    I'm getting that past Tegra feeling; lots of hype, little in the way of performance metrics.I'm calling this early; I don't think that the Pixel C will demonstrate performance benchmarks in any respect better than the iPad Pro, including its much touted graphics performance for gaming.


    Don't mistake my intention, I have no personal investment in Tegra succeeding or not. We've already seen Tergra X1 in the Shield TV and the performance was quite good.

     

    The larger iPad Air uses an A9X on a 14/16 nm process with other high end components.  I highly doubt the Tegra X1, a chip that's almost 6 months old, using older ARMv8 cores and a 20 nm process, would be able to compete. These comparisons are between life cycles. The next generation Tegra processor (based on nVidia's Pascal graphics architecture) should be announced in early January, and on the market a few months after.

     

    Regarding graphics, nVidia has had developers, in the past, port PC games using full OpenGL 4.x to Tegra K1 and now Tegra X1.  Up until the A9 and A9X, support for features in OpenGL 4.x or equivalent, was not possible. So it's still not that easy to actually compare their graphics capabilities, especially when popular mobile graphics benchmarks are based on OpenGL ES 2.0/3.1.

  • Reply 78 of 108
    nhtnht Posts: 4,522member
    Quote:

    Originally Posted by Gatorguy View Post





    We should tag these two posts to look at 10 years out. I hope we're both still around to claim we were right afterall!



    Christ, like you two didn't say the same damn thing in 2009?

  • Reply 79 of 108
    tmaytmay Posts: 6,247member
    Quote:
    Originally Posted by RobertC View Post

     

    Don't mistake my intention, I have no personal investment in Tegra succeeding or not. We've already seen Tergra X1 in the Shield TV and the performance was quite good.

     

    The larger iPad Air uses an A9X on a 14/16 nm process with other high end components.  I highly doubt the Tegra X1, a chip that's almost 6 months old, using older ARMv8 cores and a 20 nm process, would be able to compete. These comparisons are between life cycles. The next generation Tegra processor (based on nVidia's Pascal graphics architecture) should be announced in early January, and on the market a few months after.

     

    Regarding graphics, nVidia has had developers, in the past, port PC games using full OpenGL 4.x to Tegra K1 and now Tegra X1.  Up until the A9 and A9X, support for features in OpenGL 4.x or equivalent, was not possible. So it's still not that easy to actually compare their graphics capabilities, especially when popular mobile graphics benchmarks are based on OpenGL ES 2.0/3.1.


    Evidently, gaming is important to you, so battery life isn't a big issue. Buy the Pixel C or the Shield Tablet and enjoy. My point is that nVidia has stated that they are out of the mobile market, the Tegra is now targeted for vehicles, and indications are that nVidia processors aren't competitive for tablets. As all of the specs haven't been announced to date, it appears that the Pixel C is still being developed. With the rumors of the next A series processors being built System on Package (SoP), it is going to be even more difficult for nVidia to catch up.

     

    Apple obviously has a roadmap for development of the AX series, and Metal supersedes OpenGL for games; there won't be any PC games being ported for the simple reason that the tablet gaming market isn't the same as the PC gaming market. 

     

    Here's a link to a Tegra X1 vs A8X comparison from January of this year:

     

    http://techcrunch.com/2015/01/04/heres-how-nvidias-new-tegra-x1-stacks-up-to-the-k1-and-apple-a8x-on-paper/

     

    All of those numbers are based on a development system benchmarked by nVidia against an iPad Air 2 with and A8X back in January. Shield TV is the other device using the Tegra X1. With luck, Google will have the Pixel C out in December, eleven months after the Tegra X1 was announced, and now you are counting on the next generation of Tegra to both exceed the performance of the A9X, and ship in a tablet product before Apple drops the A10 or A10X?

  • Reply 80 of 108
    Quote:

    Originally Posted by jameskatt2 View Post

     

    What you don't understand is that Android Phone manufacturers compete against each other for customers. In order to compete they have to differentiate their product from one another. 

     

    Differentiation means having features such as processor speed and capability that gives them a competitive advantage over their competitor. 

     

    So Android Phone Manufacturers purchase different chips from Qualcomm, MediaTek, or Samsung to give them that competitive advantage. These companies create their own designs just like Intel competes against AMD in creating chips for PC manufacturers.

     

    What Google wants is for all the chip design firms to create the SAME design.  This would be like Microsoft asking Intel to stop competing against AMD and for both of them to create the same x86 or x64 chip.

     

    The problem for PC companies is that if Intel and AMD did this, they won't be able to create product differentiation.  The analogous problem for Android Phone Manufacturers is that they also won't be able to create product differentiation.

     

    Google seeks to force every Android Phone Manufacturer to create the same hardware using the same software.  Sony, Samsung, HTC, LG, etc. would use the same hardware - and the same software without customization.

     

    So how are Android Phone Manufacturers going to attract customers if their competitors create the same phone?  They cannot.  it forces them in a race to the bottom. It will kill off several of them since profits are already razor thin.

     

    This is why there is resistance to what Google wants to do.

     

    After all - Android is about CHOICE. 


    You make some good points.

     

    But I could see Google wanting to settle on something like 3 streamlined chip designs, low end, medium, and high end in order to make Android run better.

     

    Manufacturers can still differentiate with their device materials and designs, screen sizes and resolutions, battery life/size/removable, micro SD slot, dual sim slots, wireless charging, camera quality, etc.

     

    Thanks for the response.

     

    Quote:

    Originally Posted by DavidW View Post

     

    Because Google would be using their dominance in one market (mobile device OS) to dictate what another market can do. Since Android is on over 75% of the mobile phones, if Google change Android to work only with their new chip designs, this would force chip makers to obtain a license from Google to make compatible chips. And if Google actually contract out to fab their own chip design, then other chip makers are screwed if Google don't license out their proprietary codes for other makers chips to work with Android or Google supplies their own chip at a steep discount. Plus mobile phone makers would be forced to use Google designed chips if they want to sell mobile phones with Android. Samsung is the biggest Android mobile phone maker and is also capable of designing and fabbing their own chips. If Google start controlling the chips that Android can run on, Samsung would lose a big advantage in their market as they would be forced to use the same Google designed chips that all their competitors must use.     

     

    Apple is not in this predicament as Apple design their chips for their own phone and their own OS. Since iOS is not available to any other mobile phone makers, Apple is not forcing any other mobile phone makers to use Apple designed chips.

     

    If Google wants to design a chip that is optimized for a version of Android, contract the fabbing of it and use it for their own phone (Nexus) or sell it to other phone makers without predatory pricing, I don't see a problem with that. If chip designers design a better chip for Android and Google then add the changes so that Android will work with that chip. There's nothing wrong with that either. And there would be nothing wrong with Google adding changes to Android that chip makers can take advantage of when designing their own chips for Android. But for Google to enter the chip business with the control they have with Android in the mobile device OS market would be frown upon by anti-trust regulators as this would give Google a big advantage in the chip business and force all Android phone makers into using Google designed chips.  

     

    This would be like if Microsoft got into the chip business and then started changing the codes in Windows so that it would be optimized for their own chip design. Microsoft would end up controlling the computer chip market because of a monopoly they have in the computer OS market. That's would be a violation of anti-trust laws. Just like it was a violation of anti-trust when Microsoft used their monopoly in Windows to optimized IE for Windows. This handed Microsoft a monopoly in the browser market as other browsers could not take advantage of Windows coding. 


    You make some fair points and I understand now what I was missing.

     

    It seems that Google may need to walk a bit of a tightrope using your line of thinking. Which very well will be what they would need to do.

     

    Again, thanks for the response.

Sign In or Register to comment.