By: Doug S (foo.delete@this.bar.bar), February 2, 2013 2:38 pm
Room: Moderated Discussions
anon (anon.delete@this.anon.com) on February 1, 2013 7:08 pm wrote:
> > That's a fair criticism, but look at it this way. Apple and Samsung produce
> > at least half the smartphones and tablets shipping worldwide.
>
> So what? Intel could fab for Apple, and Android market is very fluid. Samsung does
> very well because they have good devices, not because they have some kind of monopoly
> on consumers. Only a few years ago, the Anroid market looked very different.
I think Intel should fab for Apple, this would be a win for both companies. I would think Apple would jump at the chance if they could get a similar deal to whatever they have currently with Samsung. Obviously Intel is the reason it hasn't happened yet. Probably because they still have crazy dreams of getting Apple to switch to x86 for iOS. Perhaps also because they aren't sure they can handle the capacity/schedule to take on Apple as customer.
If they do fab for Apple I would think they'd take on a lower volume product rather than trying to take on everything. Maybe make the A*X for the full sized iPad, since it doesn't sell as much as iPhone, and doesn't have as big of peaks and valleys in sales.
As for Samsung, where did I suggest they had a monopoly? I only pointed out they have a large market share, and both Apple and Samsung are effectively closed to Intel as far as selling them x86 SoCs. While it's true things can change, Samsung's share of the Android market is still growing, so it is more likely for them to become more dominant in the near term, not less.
I think you underestimate Samsung's mindshare, or are overestimating how many people actually compare all Android phones out there and choose the "best" and have no brand loyalty or bias. IMHO more people buy Samsung phones which just happen to be Android, rather than buy Android phones which just happen to be Samsung. If the GS4 came out and was a bit disappointing (which could certainly happen, given the Apple-like hype and overly optimistic guesses about what it will do) then HTC released a new phone a few weeks later that was generally considered to be better, and cost $50 less, the GS4 would still handily outsell that HTC phone. Because the GS4 would say Samsung on it.
> > The window of opportunity for Android devices to sell on performance is probably closing.
>
> Better manufacturing is more than just about performance. Cost and energy efficiency.
>
> > Ditto for battery life. Beyond a day's use for a smartphone
> > it is nice to have but not a critical factor for
> > very many people (if it was, there would be more phones like
> > the Razr Maxx) The main benefit of better battery
> > life is that it lets you make a thinner, lighter phone.
>
> And cheaper. Exactly. What is not to like about that?
>
> The screen and cellular/GPS/wifi/bluetooth radios take
> > a larger and larger portion of the battery as screens get
> > bigger and we go from 3G to LTE to LTE Advanced and
> > so on, making CPU efficiency less and less important to overall device battery life as time goes on.
>
> Although important enough for Apple to buy a design team and start doing their
> own CPUs. Clearly they believe there is room for long term differentiation.
We still don't know Apple's end game for bringing chip design in house. Was it just because Steve Jobs was a control freak? If so, why didn't they buy Imagination and take the GPU in house, too? Maybe COO Cook figured that cutting extra stuff they don't need from standard SoCs sold onto the market (like the one or two too many cores on quad core phone SoCs) could save enough sq mm per SoC across hundreds of millions of SoCs to pay for itself in savings from their foundry partner. That would be exactly the kind of calculations he would have done back then.
Maybe they had something else in mind, something that the standard ARM designs could not get them. Something they need to add additional instructions for, like emulating x86 code or accessing x86 little endian data. If they wanted to use it in laptops like Charlie has claimed. Or they intend to put them into phones to allow them to be capable of running the OS X GUI and possibly even x86 OS X software when connected to a monitor/keyboard/mouse to really disrupt the PC market (both their own, and Microsoft's)
Or they have needs for some other product they're working on like the rumored TV or something further in the future that's somehow been kept a secret so far. Maybe they didn't like the long timeline for 64 bit ARM, and figured they could do it faster themselves, and get exactly what they need for some use they foresaw. There's no point to wasting the die area and cache/RAM for bigger pointers on a 64 bit CPU in a phone, unless it needs to be capable of running a full desktop GUI.
> > That's a fair criticism, but look at it this way. Apple and Samsung produce
> > at least half the smartphones and tablets shipping worldwide.
>
> So what? Intel could fab for Apple, and Android market is very fluid. Samsung does
> very well because they have good devices, not because they have some kind of monopoly
> on consumers. Only a few years ago, the Anroid market looked very different.
I think Intel should fab for Apple, this would be a win for both companies. I would think Apple would jump at the chance if they could get a similar deal to whatever they have currently with Samsung. Obviously Intel is the reason it hasn't happened yet. Probably because they still have crazy dreams of getting Apple to switch to x86 for iOS. Perhaps also because they aren't sure they can handle the capacity/schedule to take on Apple as customer.
If they do fab for Apple I would think they'd take on a lower volume product rather than trying to take on everything. Maybe make the A*X for the full sized iPad, since it doesn't sell as much as iPhone, and doesn't have as big of peaks and valleys in sales.
As for Samsung, where did I suggest they had a monopoly? I only pointed out they have a large market share, and both Apple and Samsung are effectively closed to Intel as far as selling them x86 SoCs. While it's true things can change, Samsung's share of the Android market is still growing, so it is more likely for them to become more dominant in the near term, not less.
I think you underestimate Samsung's mindshare, or are overestimating how many people actually compare all Android phones out there and choose the "best" and have no brand loyalty or bias. IMHO more people buy Samsung phones which just happen to be Android, rather than buy Android phones which just happen to be Samsung. If the GS4 came out and was a bit disappointing (which could certainly happen, given the Apple-like hype and overly optimistic guesses about what it will do) then HTC released a new phone a few weeks later that was generally considered to be better, and cost $50 less, the GS4 would still handily outsell that HTC phone. Because the GS4 would say Samsung on it.
> > The window of opportunity for Android devices to sell on performance is probably closing.
>
> Better manufacturing is more than just about performance. Cost and energy efficiency.
>
> > Ditto for battery life. Beyond a day's use for a smartphone
> > it is nice to have but not a critical factor for
> > very many people (if it was, there would be more phones like
> > the Razr Maxx) The main benefit of better battery
> > life is that it lets you make a thinner, lighter phone.
>
> And cheaper. Exactly. What is not to like about that?
>
> The screen and cellular/GPS/wifi/bluetooth radios take
> > a larger and larger portion of the battery as screens get
> > bigger and we go from 3G to LTE to LTE Advanced and
> > so on, making CPU efficiency less and less important to overall device battery life as time goes on.
>
> Although important enough for Apple to buy a design team and start doing their
> own CPUs. Clearly they believe there is room for long term differentiation.
We still don't know Apple's end game for bringing chip design in house. Was it just because Steve Jobs was a control freak? If so, why didn't they buy Imagination and take the GPU in house, too? Maybe COO Cook figured that cutting extra stuff they don't need from standard SoCs sold onto the market (like the one or two too many cores on quad core phone SoCs) could save enough sq mm per SoC across hundreds of millions of SoCs to pay for itself in savings from their foundry partner. That would be exactly the kind of calculations he would have done back then.
Maybe they had something else in mind, something that the standard ARM designs could not get them. Something they need to add additional instructions for, like emulating x86 code or accessing x86 little endian data. If they wanted to use it in laptops like Charlie has claimed. Or they intend to put them into phones to allow them to be capable of running the OS X GUI and possibly even x86 OS X software when connected to a monitor/keyboard/mouse to really disrupt the PC market (both their own, and Microsoft's)
Or they have needs for some other product they're working on like the rumored TV or something further in the future that's somehow been kept a secret so far. Maybe they didn't like the long timeline for 64 bit ARM, and figured they could do it faster themselves, and get exactly what they need for some use they foresaw. There's no point to wasting the die area and cache/RAM for bigger pointers on a 64 bit CPU in a phone, unless it needs to be capable of running a full desktop GUI.