Inside iPhone 8: Apple's A11 Bionic introduces 5 new custom silicon engines

1246

Comments

  • Reply 61 of 119
    avon b7avon b7 Posts: 7,622member
    tmay said:
    avon b7 said:
    wizard69 said:
    avon b7 said:
    Blunt said:
    avon b7 said: 
    The idea that some have that Android is just some kind of ad platform, apart from being somewhat insulting to Android users, is just plain stupid and ignores the many good reasons people buy Android.


    Your reaction to the speed of the A11 this week was plain stupid and somewhat insulting and ignores the many good reasons people buy iPhones. When you buy an new iPhone it will last a long time and you will recieve updates for years. The iPhone 8 will be up speed for years and Android phones will not be as fast for maybe three years.
    There is also the fact that the A11 opens the door for exciting things to come. By the way we all know the main reason people buy Android: they are cheap just like Windows computers. If you want't the best quality Android is not an option.
    My reaction was in no way insulting to anybody. The opinion on speed was fine or has anything changed on that front in the last three years? Did it do anything big for Apple? Did it improve sales? Did it improve the user experience for the vast majority of users?

    As I have said before and detailed before, the industry hit a wall on features a couple of years ago and what we have seen is a fleshing out of those features not anything truly revolutionary. Processors are fast enough, cameras are good enough, screens are good enough etc, etc.

    The future is AI and possibly AR/VR. We might see some jump in battery tech. We might see some foldable screen tech etc but, today, smartphones largely do what most people currently ask of them and extremely well (and that includes those mid tier Androids).

    If your plan is to upgrade every year or two, how does having a phone that 'lasts' longer help?
    Well I'm interjecting here but I don't upgrade every two years.   That is a behavior that the gullible and pound foolish have engaged in for some time.   It bothers me to see comments about 2 year upgrades as it leads people to believe they have to do an upgrade.

    As for ipHone performance I don't think we will ever see an end to performance increases.    At some point cell phones will completely replace PC's for most user needs.    They will simply plug the phone into a monitor at homeland at work to get all the PC functionality they need.    Unplug the device and the phone reverts to cell phone mode.   Even if we don't get there anytime soon performance is still very important for user satisfaction.

    In any even your negativity is ether willful trolling or the result fo a complete misunderstanding what Apple has managed ot build into the A11 chip.   It is a major milestone.    Besides if you believe the the future is AI and AR/VR (I generally agree but have a hard time calling todays AI systems AI) then you apparently already realize that such features take a tremendous amount of processing power to implement. Frankly I see A11 as just broaching the idea of AI hardware in a computer for the common user.   To really get this tech to work we will need a massive amount of extra compute power just ot make it look half good.   I used the term milestone because that really is what we have in A11, it is the birth of a new generation of technologies and the only way to go from here is up.

    You can dismiss the new iPhones all you want, you are still free to do so.   The problem is that opinion doesn't seem to correlate with an objective look at what Apple debuted.  
    No one is obliged to upgrade at any given rate. The point is more a consideration at the time of the initial outlay. My wife has an iPhone 6 in the knowledge that the purchase was for three years. I have a 249€ Android in the knowledge that it would be for 18 months or more.

    AI is taking its first steps on mobile hardware. First with machine learning for a few things (handset optimisation, anticipation, camera etc) For me this was the first step so I know what you mean when you say you have a hard time calling it AI. But now we have dedicated NPUs finding space on the SoC. The results may seem a little lacking in substance at first because they need software to be written for them but the possibilities are vast going forward.

    Nearly a year ago the Honor Magic was released and it gave us a peek at what could be done on a 'simple' level (no NPU) just by incorporating subtle changes. For example, the idea that the phone could 'know' whose hand it was in and if the user was looking at the screen and react according. In the user's hand, the screen would use sensors in the frame and face recognition to wake the screen and reveal the text in notifications. The same gesture in anybody else's hand would not show the text in notifications. If the user stopped looking at the screen for 10 seconds it would switch to power saving mode. This was using its infrared and eye tracking tech. It would do different things depending on if the phone face was face up or down, in a dark room or not etc. Providing quick access to boarding passes at airports or package info when picking up packets.

    Much of this AI was 'invisible' to the user but is the kind of thing you somehow miss when you pick up a different phone that doesn't have it. And this was nearly a year ago. Expect major refinements very soon.

    I compare the sensation to the classic Mac OS Finder. It was a technological neanderthal but always seemed to be one step ahead of what the user was going to do next. In its own special way it was 'intelligent' even if it was just the result of nice design.

    Now, all this (iPhone 8, X and the Mate 10, next Honor Magic) is going 'local' to improve context, latency and privacy. The first areas of main development have been on image recognition and voice recognition but when you tie in information from on-board sensors, user habits, cloud support services etc, you begin to see the potential. I've always argued for onboard AI in some circumstances and now we will see the best of both worlds come into play. Imagine how text input and spelling could be improved (especially in multi language cases)

    Or imagine this: 

    You are diabetic type 1. Imagine having your continuous monitoring glucose sensor (hopefully non-invasive  at some point) tied in with your phone and being able to scan all the food at home (barcodes on packets and camera for food recognition) and for the phone to guide you through glucose changes (alerts) and give advice on which kind of insulin and how much to use, based both on the food you propose, the food you have eaten earlier, your history, and the exercise you have done, the sleep you have has AND learn from past events (hipoglucemia/hiperglucemia) to improve its suggestions. Now imagine if emergency services could upload pattern information (for FaceID for example) and users could store the information (relating to local personnel) locally on their phones and have only authorised medical personnel gain access - to the relevant information and that previously chosen a by the user - on their phones, and only in case of emergency. Information that would include up to the minute real health information.

    Some of this is available now but lacking the ability to learn from the situation and adapt to it (the intelligence). Imagine if it could know you have neuropatholigies, accomodate that information on your diabetes and keep all this info private (locally onboard the phone) until a time that it needs to be revealed (to a member of the emergency services for example and only in case of need).

    AR could take off universally or end up as a niche feature or fall somewhere in between but AI will truly change our lives even if we aren't quite aware of its presence.

    At some point sensor tech will become commonplace in glasses and then we will have even more options available to AI.

    There will be issues with privacy, legality and other problems but AI will have to be adapted to one way or another but, like I said, these are just the first steps.


    You could have shortened alt that to "just imagine what Apple could do with AI, health, wearables and specifically the Apple Watch.", but of course, you're afraid that you would explode in a ball of flame if you said anything nice about Apple.

    Who, other than Apple, controls all the pieces needed to put together what you are asking for? Sure, Google s attempting to be more like Apple as we speak, good luck with that. Otherwise, it will be the OEM's on their own, providing little bits and pieces, but no full solution, the achilles heel of Android OS.

    See how Apple rolled out ARKit, for example, and how Apple had the foresight to provide enough performance in their iPhones, even a few generations ago, to make this happen. So today's performance, that you ignore, is tomorrow's new capability, and we will see, again, Android OS fall behind.


    You lost me completely with that. We know where Apple is going but it isn't limited to Apple. Others have already started the journey.

    I mentioned the Honor Magic precisely to give some background info on what is already out there. Stuff that is working today and has been working for a while now. It's not the only one. There are phones with underscreen sensors to track where your finger is going to hit the screen and get apps ready to go even more quickly, that learn your habits, adapt to them and optimise the phone accordingly. This is done within the phone's hardware or on top of Android. It is definitely not behind Apple, with or without ARKit, but I really don't know what you want to say. Apple is there to a limited degree now and will move forward, along with others but that is stating the obvious. This is NOT something only Apple can do and even if it were, most of the world wouldn't have been able to use it.

    The NPU functionality of the Kirin970 is open to developers and they can use the Huawei AI API. It is also compatible with the TensorFlow/TensorFlow Lite or Caffe/Caffe2 frameworks.

    I read somewhere that Huawei has thousands of developers writing for it.

    I'm the first to tell people to take Huawei marketing with a pinch of salt as much of it is just that, plain old marketing, but only a fool would not be able to see that the Honor Magic was just one of a host of their phones tackling machine learning and AI - in the field - and that any progress made in these areas will be integrated into the Mate 10 and P11 etc. Apple uses machine learning and AI but is it implementing AI as the Magic does?

    So when Huawei says that the 16th October will be the date for 'real' mobile AI to take the stage, I raise an eyebrow and am as sceptical as the next person, but if what they have had for almost a year now is pretty well thought through, and they now have the NPU to leverage as well, I'd say it will be an interesting presentation on the 16th at the very least.

    That, on top of what is already known about the Kirin 970 SoC:

    The dual ISPs and NPU together with the new Leica hardware to reduce things like motion blur and identify picture content in real time

    The new Cat18 hardware which can support dual 4G SIMS simultaneously along with improved design to manage connections at high speed (trains etc), difficult reception and data throughput scenarios (tunnels, basements etc) and playing off their networking backbone infrastructure which they say is used on half of the world's mobile communications hardware.

    If the AI and the hardware marry well, and are actually 'visibly' useful to consumers, I see the package being very attractive. It will be expensive. There is no doubt about it, but the technology will filter down to the lower tiers very, very quickly.

    And Android isn't really a big problem for them as they have already successfully implemented their own solutions and will continue to do so.

    As for Apple and me,  you clearly haven't been paying attention. I was probably the first person here to celebrate their new price spread which I had argued for previously and, unless I am mistaken, you said would never happen. I also spoke out in favour of the 'notch' and have defended some other aspects.

    I said it would be nice if they dumped black and was it you who piped in with something like that wouldn't happen? I don't pay much attention to names because I don't need to. I also said the NPU could appear on the A11 and celebrated that too.

    My guesses on Huawei are playing out well and I'm pretty sure the 16th October event will see an announcement on them landing in the US with a big carrier agreement (probably AT&T). 

    Why is that important? For the reason I have stated multiple times and which you appear to ignore: COMPETITION.

    Competition is the key.

    This isn't a place to say 'nice things' about Apple. It's a place to give opinions, whatever they may be, and no, I won't go up in flames, implode or otherwise suffer if I say something 'nice' about them.

    I celebrate competition over any other aspect because it is the only way consumers can get a relatively fair deal.

    I am glad that the hardware duopoly is over. I am glad that that there is so much choice on the market. I am glad that Huawei has competition and I am glad prices are so varied (and that now includes Apple).

    It's good for everyone.






  • Reply 62 of 119
    apple_evo said:
    melgross said:
    It will be interesting to see whether Apple decides to support Google’s codecs. It’s up to Apple to do that. Since it’s free to support, I can’t see why they wouldn’t, unless it just corporate rivalry on their part.

    remember that Apple never supported FLAC for music either, but now they do. So that’s a change.
    Why would Apple support Google codecs? Does Apple support Flash? No, but we can still use Flash thanks to the Flash plug-in on macOS. We use Microsoft's Silverlight plug-in to watch Netflix. Google must just provide its own plug-in for its codec. If it provides its own codec for tvOS and Apple refuses that then we may question that. But since Google doesn't provide a plug-in for macOS, it is most probably that it doesn't provide a codec for tvOS either. They might well implement it in their YouTube tvOS app. So the truth is not Apple does not support Google's codecs, the truth is Google's YouTube app does not support Google's own 4K codec on tvOS. If I'm wrong and it supports then correct me.
    It's not on google to support the AppleTV 4k, it's on apple to support the VP9 codec... 
    That’s a matter of opinion. Google could certainly serve .265 4k videos if they wanted, yes? They choose what to serve. It’s a choice, just as Apple’s position is a choice. 
    edited September 2017
  • Reply 63 of 119
    avon b7avon b7 Posts: 7,622member
    k2kw said:
    Folio said:

    Funny how the depth and breadth of 10th anniversary rollout— phenomenal to most of us partisans— so far eludes general appreciation. But that’s just a momentary hiccup, so I hope.

    Appreciate the insight on the processors. Look forward to a series: AI on AI. Perhaps also consider initiating a quarterly comparison: Sets of questions and answers to Siri, Alexa, Google Asst, Cortana, Bixby.

    I recently did 2400 mile road trip in California and Arizona. My travel buddy used Galaxy/Verizon/Google and I had iPhone5s/12.9iPadPro on T-Mobile. Verizon coverage noticeably superior outside of cities. Siri did fine, except for glaring misdirection to Sequoia National Park, taking us up by Edison Lake instead, so we missed seeing the gargantuan trees this trip, but at least hiked great lake with mountain air.

    Another time I asked Siri driving distance from Tucson to Phoenix. Instead of a simple quick answer, it demanded I turn on location services. Grrrr. In my dream, Siri would have traits like a concierge in a world class hotel and not talk back, let alone be impertinent. Note: I’ve not yet used new Siri.

    A recent brokerage report (BAML 31AUG2017 on Google Traffic Acquisition Costs) estimates Google pays Apple $4billion plus each year in TAC fees. That’s US$4,000,000,000 plus to tap each of us in Apple’s ecosystem. Google’s iOS recent assistant app, while it may help Apple retain any users who would otherwise defect to Android for the assistant, likely is a foot in the door to reduce annual TAC payments from Google to Apple.

    If Apple makes more Siri experiences superior or near frictionless, with consent tailored to its individual users, then it can not only protect TAC fees but gain more lucrative share in e-commerce from Google, Amazon, etc. Recent internal moves, though late in coming, are encouraging, and HomePod visibility means Siri development can’t languish anymore. Game on.

    Bionic chip processing power for AR/VR could be a differentiator for Apple mobile for next few years in buying fitted clothes, shower curtains, sofas, views from/of hotel rooms, maybe even head room in cars or leg room on airline seats. Question is how fast sellers ramp up.

    Is there a "new" SIRI that comes with the iPhone 8/8Plus or iPhoneX?    I use a iPhoneSE and 5SI.     SIRI has always worked like crap for me.    It's so frustrating that
    after all these years it still seems to be in beta.  Its because of SIRI that I'm not interested in CARPLAY , the apple watch, or AirPods.   I'm very curious and hopeful about the HomePod.   Hopefully there will be a SIRI in it that works as good as my Alexa (Amazon is beating Google on the hardware front-The Pixel 2 rumors make it sound like a pathetic Sophomore edition).

    Is there anything new (hardware or software wise) in the iPhone 8/8Plus/X that makes SIRI really work better.   Apple has added a notchful of sensors to the X for FaceID? Is there a new sensor to help SIRI work better.

    I've long thought that Apple aims for an A+ on display quality, A+ on core iOS Software, A+ on build quality, A- on Design,  B+ on audio, but a C on SIRI.   
    Looking forward to when SIRI has been improved by those aquisitions that DED has often pointed out.
    I don't use Siri enough to give a valid opinion on effectiveness but I'd say it could only get better.

    I tried it on a Hyundai Ionic 2017 and it was a little embarrassing but it could be that the Hyundai hardware was not helping. The activation button on the wheel seemed to be a little out of sync with the screen animation and as a result perhaps Siri was only getting bits of the comnands.

    I can confirm though that since iOS10, Siri's Spanish Spanish voice is horrible. Far worse than the old one.
  • Reply 64 of 119
    tmaytmay Posts: 6,309member
    avon b7 said:
    tmay said:
    avon b7 said:
    wizard69 said:
    avon b7 said:
    Blunt said:
    avon b7 said: 
    The idea that some have that Android is just some kind of ad platform, apart from being somewhat insulting to Android users, is just plain stupid and ignores the many good reasons people buy Android.


    Your reaction to the speed of the A11 this week was plain stupid and somewhat insulting and ignores the many good reasons people buy iPhones. When you buy an new iPhone it will last a long time and you will recieve updates for years. The iPhone 8 will be up speed for years and Android phones will not be as fast for maybe three years.
    There is also the fact that the A11 opens the door for exciting things to come. By the way we all know the main reason people buy Android: they are cheap just like Windows computers. If you want't the best quality Android is not an option.
    My reaction was in no way insulting to anybody. The opinion on speed was fine or has anything changed on that front in the last three years? Did it do anything big for Apple? Did it improve sales? Did it improve the user experience for the vast majority of users?

    As I have said before and detailed before, the industry hit a wall on features a couple of years ago and what we have seen is a fleshing out of those features not anything truly revolutionary. Processors are fast enough, cameras are good enough, screens are good enough etc, etc.

    The future is AI and possibly AR/VR. We might see some jump in battery tech. We might see some foldable screen tech etc but, today, smartphones largely do what most people currently ask of them and extremely well (and that includes those mid tier Androids).

    If your plan is to upgrade every year or two, how does having a phone that 'lasts' longer help?
    Well I'm interjecting here but I don't upgrade every two years.   That is a behavior that the gullible and pound foolish have engaged in for some time.   It bothers me to see comments about 2 year upgrades as it leads people to believe they have to do an upgrade.

    As for ipHone performance I don't think we will ever see an end to performance increases.    At some point cell phones will completely replace PC's for most user needs.    They will simply plug the phone into a monitor at homeland at work to get all the PC functionality they need.    Unplug the device and the phone reverts to cell phone mode.   Even if we don't get there anytime soon performance is still very important for user satisfaction.

    In any even your negativity is ether willful trolling or the result fo a complete misunderstanding what Apple has managed ot build into the A11 chip.   It is a major milestone.    Besides if you believe the the future is AI and AR/VR (I generally agree but have a hard time calling todays AI systems AI) then you apparently already realize that such features take a tremendous amount of processing power to implement. Frankly I see A11 as just broaching the idea of AI hardware in a computer for the common user.   To really get this tech to work we will need a massive amount of extra compute power just ot make it look half good.   I used the term milestone because that really is what we have in A11, it is the birth of a new generation of technologies and the only way to go from here is up.

    You can dismiss the new iPhones all you want, you are still free to do so.   The problem is that opinion doesn't seem to correlate with an objective look at what Apple debuted.  
    No one is obliged to upgrade at any given rate. The point is more a consideration at the time of the initial outlay. My wife has an iPhone 6 in the knowledge that the purchase was for three years. I have a 249€ Android in the knowledge that it would be for 18 months or more.

    AI is taking its first steps on mobile hardware. First with machine learning for a few things (handset optimisation, anticipation, camera etc) For me this was the first step so I know what you mean when you say you have a hard time calling it AI. But now we have dedicated NPUs finding space on the SoC. The results may seem a little lacking in substance at first because they need software to be written for them but the possibilities are vast going forward.

    Nearly a year ago the Honor Magic was released and it gave us a peek at what could be done on a 'simple' level (no NPU) just by incorporating subtle changes. For example, the idea that the phone could 'know' whose hand it was in and if the user was looking at the screen and react according. In the user's hand, the screen would use sensors in the frame and face recognition to wake the screen and reveal the text in notifications. The same gesture in anybody else's hand would not show the text in notifications. If the user stopped looking at the screen for 10 seconds it would switch to power saving mode. This was using its infrared and eye tracking tech. It would do different things depending on if the phone face was face up or down, in a dark room or not etc. Providing quick access to boarding passes at airports or package info when picking up packets.

    Much of this AI was 'invisible' to the user but is the kind of thing you somehow miss when you pick up a different phone that doesn't have it. And this was nearly a year ago. Expect major refinements very soon.

    I compare the sensation to the classic Mac OS Finder. It was a technological neanderthal but always seemed to be one step ahead of what the user was going to do next. In its own special way it was 'intelligent' even if it was just the result of nice design.

    Now, all this (iPhone 8, X and the Mate 10, next Honor Magic) is going 'local' to improve context, latency and privacy. The first areas of main development have been on image recognition and voice recognition but when you tie in information from on-board sensors, user habits, cloud support services etc, you begin to see the potential. I've always argued for onboard AI in some circumstances and now we will see the best of both worlds come into play. Imagine how text input and spelling could be improved (especially in multi language cases)

    Or imagine this: 

    You are diabetic type 1. Imagine having your continuous monitoring glucose sensor (hopefully non-invasive  at some point) tied in with your phone and being able to scan all the food at home (barcodes on packets and camera for food recognition) and for the phone to guide you through glucose changes (alerts) and give advice on which kind of insulin and how much to use, based both on the food you propose, the food you have eaten earlier, your history, and the exercise you have done, the sleep you have has AND learn from past events (hipoglucemia/hiperglucemia) to improve its suggestions. Now imagine if emergency services could upload pattern information (for FaceID for example) and users could store the information (relating to local personnel) locally on their phones and have only authorised medical personnel gain access - to the relevant information and that previously chosen a by the user - on their phones, and only in case of emergency. Information that would include up to the minute real health information.

    Some of this is available now but lacking the ability to learn from the situation and adapt to it (the intelligence). Imagine if it could know you have neuropatholigies, accomodate that information on your diabetes and keep all this info private (locally onboard the phone) until a time that it needs to be revealed (to a member of the emergency services for example and only in case of need).

    AR could take off universally or end up as a niche feature or fall somewhere in between but AI will truly change our lives even if we aren't quite aware of its presence.

    At some point sensor tech will become commonplace in glasses and then we will have even more options available to AI.

    There will be issues with privacy, legality and other problems but AI will have to be adapted to one way or another but, like I said, these are just the first steps.


    You could have shortened alt that to "just imagine what Apple could do with AI, health, wearables and specifically the Apple Watch.", but of course, you're afraid that you would explode in a ball of flame if you said anything nice about Apple.

    Who, other than Apple, controls all the pieces needed to put together what you are asking for? Sure, Google s attempting to be more like Apple as we speak, good luck with that. Otherwise, it will be the OEM's on their own, providing little bits and pieces, but no full solution, the achilles heel of Android OS.

    See how Apple rolled out ARKit, for example, and how Apple had the foresight to provide enough performance in their iPhones, even a few generations ago, to make this happen. So today's performance, that you ignore, is tomorrow's new capability, and we will see, again, Android OS fall behind.


    You lost me completely with that. We know where Apple is going but it isn't limited to Apple. Others have already started the journey.

    I mentioned the Honor Magic precisely to give some background info on what is already out there. Stuff that is working today and has been working for a while now. It's not the only one. There are phones with underscreen sensors to track where your finger is going to hit the screen and get apps ready to go even more quickly, that learn your habits, adapt to them and optimise the phone accordingly. This is done within the phone's hardware or on top of Android. It is definitely not behind Apple, with or without ARKit, but I really don't know what you want to say. Apple is there to a limited degree now and will move forward, along with others but that is stating the obvious. This is NOT something only Apple can do and even if it were, most of the world wouldn't have been able to use it.

    The NPU functionality of the Kirin970 is open to developers and they can use the Huawei AI API. It is also compatible with the TensorFlow/TensorFlow Lite or Caffe/Caffe2 frameworks.

    I read somewhere that Huawei has thousands of developers writing for it.

    I'm the first to tell people to take Huawei marketing with a pinch of salt as much of it is just that, plain old marketing, but only a fool would not be able to see that the Honor Magic was just one of a host of their phones tackling machine learning and AI - in the field - and that any progress made in these areas will be integrated into the Mate 10 and P11 etc. Apple uses machine learning and AI but is it implementing AI as the Magic does?

    So when Huawei says that the 16th October will be the date for 'real' mobile AI to take the stage, I raise an eyebrow and am as sceptical as the next person, but if what they have had for almost a year now is pretty well thought through, and they now have the NPU to leverage as well, I'd say it will be an interesting presentation on the 16th at the very least.

    That, on top of what is already known about the Kirin 970 SoC:

    The dual ISPs and NPU together with the new Leica hardware to reduce things like motion blur and identify picture content in real time

    The new Cat18 hardware which can support dual 4G SIMS simultaneously along with improved design to manage connections at high speed (trains etc), difficult reception and data throughput scenarios (tunnels, basements etc) and playing off their networking backbone infrastructure which they say is used on half of the world's mobile communications hardware.

    If the AI and the hardware marry well, and are actually 'visibly' useful to consumers, I see the package being very attractive. It will be expensive. There is no doubt about it, but the technology will filter down to the lower tiers very, very quickly.

    And Android isn't really a big problem for them as they have already successfully implemented their own solutions and will continue to do so.

    As for Apple and me,  you clearly haven't been paying attention. I was probably the first person here to celebrate their new price spread which I had argued for previously and, unless I am mistaken, you said would never happen. I also spoke out in favour of the 'notch' and have defended some other aspects.

    I said it would be nice if they dumped black and was it you who piped in with something like that wouldn't happen? I don't pay much attention to names because I don't need to. I also said the NPU could appear on the A11 and celebrated that too.

    My guesses on Huawei are playing out well and I'm pretty sure the 16th October event will see an announcement on them landing in the US with a big carrier agreement (probably AT&T). 

    Why is that important? For the reason I have stated multiple times and which you appear to ignore: COMPETITION.

    Competition is the key.

    This isn't a place to say 'nice things' about Apple. It's a place to give opinions, whatever they may be, and no, I won't go up in flames, implode or otherwise suffer if I say something 'nice' about them.

    I celebrate competition over any other aspect because it is the only way consumers can get a relatively fair deal.

    I am glad that the hardware duopoly is over. I am glad that that there is so much choice on the market. I am glad that Huawei has competition and I am glad prices are so varied (and that now includes Apple).

    It's good for everyone.






    You, lost me, at "we know where Apple is going".

    No, we don't know where Apple is going. Apple doesn't actually tell us where they are going, unlike most of its "competition", that are only to happy to paint an optimistic roadmap. With Apple, we only know where they might be heading because of lower order information, like hirings, meetings with the FDA, Tim Cook wearing a medical sensor of unknown origin.

    You're like the guy who sees an iceberg, and goes, "what's the big deal?"
    edited September 2017 ericthehalfbeecaliargonaut
  • Reply 65 of 119
    asdasdasdasd Posts: 5,686member
    tmay said:
    crosslad said:
    "Google  - aimed at hitting an average selling price of less than $300—Android One phones have an aggressive price target of $100." 

    Not entirely true nowadays; the Pixel phones were priced the same as the iPhones and the new MotoX4, running Android One is around $400. This has got to be good for Apple. I would not pay the same price for an android phone as an iPhone. Now that Apple has phones at every price point  I can only see Apple sales increasing.

    Average selling price of Androids, including Pixel, is actually below $200. Even Samsung’s ASP, which includes its premium-priced Galaxy phones that sell in the 10s of millions, is now at $227. 

    http://www.androidauthority.com/price-gap-samsung-apple-smartphones-769772/

    This isn’t a new development. iPhone ASP has been pretty constant at or above $650 while Androids have have been below $300 since 2013. 

    Having a vanity model that sells in tiny quantities has little impact on an average, whether Pixel or Virtu. 



    The few members of the Church of Marketshare on AI will not acknowledge those facts.
    When Tim Cook joins that church, as he is clearly indicating, you will all rush to be members. 
  • Reply 66 of 119
    tmaytmay Posts: 6,309member
    asdasd said:
    tmay said:
    crosslad said:
    "Google  - aimed at hitting an average selling price of less than $300—Android One phones have an aggressive price target of $100." 

    Not entirely true nowadays; the Pixel phones were priced the same as the iPhones and the new MotoX4, running Android One is around $400. This has got to be good for Apple. I would not pay the same price for an android phone as an iPhone. Now that Apple has phones at every price point  I can only see Apple sales increasing.

    Average selling price of Androids, including Pixel, is actually below $200. Even Samsung’s ASP, which includes its premium-priced Galaxy phones that sell in the 10s of millions, is now at $227. 

    http://www.androidauthority.com/price-gap-samsung-apple-smartphones-769772/

    This isn’t a new development. iPhone ASP has been pretty constant at or above $650 while Androids have have been below $300 since 2013. 

    Having a vanity model that sells in tiny quantities has little impact on an average, whether Pixel or Virtu. 



    The few members of the Church of Marketshare on AI will not acknowledge those facts.
    When Tim Cook joins that church, as he is clearly indicating, you will all rush to be members. 
    The Church of Marketshare is all about maximizing share, not about profits, so, I'm not seeing what you are wrt Tim Cook.
  • Reply 67 of 119
    iqatedoiqatedo Posts: 1,822member
    haar said:
    iqatedo said:
    Does anyone here know on what systems Apple designs their hardware, at chip level and final device level?
    While you are at it... ask what software does INTEL use to design its CPU’s... or AMD uses... or NVIDIA uses, or qualcomm...

    My guess... some custom million dollar program that is specfic to the fab process.

    IMO in otherwards, there is no need to know... but i guess if you had to know start using the google web browser to see what software is compatible with ASML (the makers of fab machines) ... etc...
    Not interested in Intel, AMD or NVIDIA. Nor am I interested in the rest of your stupid, banal post.
    caliargonaut
  • Reply 68 of 119
    Wasn’t Huawei bragging about their Kirin 970 - trying to steal some thunder just before Apples event?

    Sorry, Huawei, you’re not the first to put in a neural processor. You were the first to “announce” it, but it’s all vaporware until you ship devices (which Apple is).

    Even so, it’s not that big of a deal - neural processors are not that difficult to design (compared to a CPU or GPU) since they are optimized to perform a few basic tasks really well. The 970 still uses inferior ARM cores (instead of designing their own) or Mali GPUs (again, instead of designing their own). If they really wanted to impress people they should have made a completely custom SoC with their own CPU/GPU cores.

    They took the easy way out by designing a much simpler component (nueral procesdor) and trying to make it appear more relevant than it is to make up for their lack of ability to design their own CPU/GPU.
    tmaySoli
  • Reply 69 of 119
    tmay said:
    avon b7 said:
    wizard69 said:
    avon b7 said:
    Blunt said:
    avon b7 said: 
    The idea that some have that Android is just some kind of ad platform, apart from being somewhat insulting to Android users, is just plain stupid and ignores the many good reasons people buy Android.


    Your reaction to the speed of the A11 this week was plain stupid and somewhat insulting and ignores the many good reasons people buy iPhones. When you buy an new iPhone it will last a long time and you will recieve updates for years. The iPhone 8 will be up speed for years and Android phones will not be as fast for maybe three years.
    There is also the fact that the A11 opens the door for exciting things to come. By the way we all know the main reason people buy Android: they are cheap just like Windows computers. If you want't the best quality Android is not an option.
    My reaction was in no way insulting to anybody. The opinion on speed was fine or has anything changed on that front in the last three years? Did it do anything big for Apple? Did it improve sales? Did it improve the user experience for the vast majority of users?

    As I have said before and detailed before, the industry hit a wall on features a couple of years ago and what we have seen is a fleshing out of those features not anything truly revolutionary. Processors are fast enough, cameras are good enough, screens are good enough etc, etc.

    The future is AI and possibly AR/VR. We might see some jump in battery tech. We might see some foldable screen tech etc but, today, smartphones largely do what most people currently ask of them and extremely well (and that includes those mid tier Androids).

    If your plan is to upgrade every year or two, how does having a phone that 'lasts' longer help?
    Well I'm interjecting here but I don't upgrade every two years.   That is a behavior that the gullible and pound foolish have engaged in for some time.   It bothers me to see comments about 2 year upgrades as it leads people to believe they have to do an upgrade.

    As for ipHone performance I don't think we will ever see an end to performance increases.    At some point cell phones will completely replace PC's for most user needs.    They will simply plug the phone into a monitor at homeland at work to get all the PC functionality they need.    Unplug the device and the phone reverts to cell phone mode.   Even if we don't get there anytime soon performance is still very important for user satisfaction.

    In any even your negativity is ether willful trolling or the result fo a complete misunderstanding what Apple has managed ot build into the A11 chip.   It is a major milestone.    Besides if you believe the the future is AI and AR/VR (I generally agree but have a hard time calling todays AI systems AI) then you apparently already realize that such features take a tremendous amount of processing power to implement. Frankly I see A11 as just broaching the idea of AI hardware in a computer for the common user.   To really get this tech to work we will need a massive amount of extra compute power just ot make it look half good.   I used the term milestone because that really is what we have in A11, it is the birth of a new generation of technologies and the only way to go from here is up.

    You can dismiss the new iPhones all you want, you are still free to do so.   The problem is that opinion doesn't seem to correlate with an objective look at what Apple debuted.  
    No one is obliged to upgrade at any given rate. The point is more a consideration at the time of the initial outlay. My wife has an iPhone 6 in the knowledge that the purchase was for three years. I have a 249€ Android in the knowledge that it would be for 18 months or more.

    AI is taking its first steps on mobile hardware. First with machine learning for a few things (handset optimisation, anticipation, camera etc) For me this was the first step so I know what you mean when you say you have a hard time calling it AI. But now we have dedicated NPUs finding space on the SoC. The results may seem a little lacking in substance at first because they need software to be written for them but the possibilities are vast going forward.

    Nearly a year ago the Honor Magic was released and it gave us a peek at what could be done on a 'simple' level (no NPU) just by incorporating subtle changes. For example, the idea that the phone could 'know' whose hand it was in and if the user was looking at the screen and react according. In the user's hand, the screen would use sensors in the frame and face recognition to wake the screen and reveal the text in notifications. The same gesture in anybody else's hand would not show the text in notifications. If the user stopped looking at the screen for 10 seconds it would switch to power saving mode. This was using its infrared and eye tracking tech. It would do different things depending on if the phone face was face up or down, in a dark room or not etc. Providing quick access to boarding passes at airports or package info when picking up packets.

    Much of this AI was 'invisible' to the user but is the kind of thing you somehow miss when you pick up a different phone that doesn't have it. And this was nearly a year ago. Expect major refinements very soon.

    I compare the sensation to the classic Mac OS Finder. It was a technological neanderthal but always seemed to be one step ahead of what the user was going to do next. In its own special way it was 'intelligent' even if it was just the result of nice design.

    Now, all this (iPhone 8, X and the Mate 10, next Honor Magic) is going 'local' to improve context, latency and privacy. The first areas of main development have been on image recognition and voice recognition but when you tie in information from on-board sensors, user habits, cloud support services etc, you begin to see the potential. I've always argued for onboard AI in some circumstances and now we will see the best of both worlds come into play. Imagine how text input and spelling could be improved (especially in multi language cases)

    Or imagine this: 

    You are diabetic type 1. Imagine having your continuous monitoring glucose sensor (hopefully non-invasive  at some point) tied in with your phone and being able to scan all the food at home (barcodes on packets and camera for food recognition) and for the phone to guide you through glucose changes (alerts) and give advice on which kind of insulin and how much to use, based both on the food you propose, the food you have eaten earlier, your history, and the exercise you have done, the sleep you have has AND learn from past events (hipoglucemia/hiperglucemia) to improve its suggestions. Now imagine if emergency services could upload pattern information (for FaceID for example) and users could store the information (relating to local personnel) locally on their phones and have only authorised medical personnel gain access - to the relevant information and that previously chosen a by the user - on their phones, and only in case of emergency. Information that would include up to the minute real health information.

    Some of this is available now but lacking the ability to learn from the situation and adapt to it (the intelligence). Imagine if it could know you have neuropatholigies, accomodate that information on your diabetes and keep all this info private (locally onboard the phone) until a time that it needs to be revealed (to a member of the emergency services for example and only in case of need).

    AR could take off universally or end up as a niche feature or fall somewhere in between but AI will truly change our lives even if we aren't quite aware of its presence.

    At some point sensor tech will become commonplace in glasses and then we will have even more options available to AI.

    There will be issues with privacy, legality and other problems but AI will have to be adapted to one way or another but, like I said, these are just the first steps.


    You could have shortened alt that to "just imagine what Apple could do with AI, health, wearables and specifically the Apple Watch.", but of course, you're afraid that you would explode in a ball of flame if you said anything nice about Apple.

    Who, other than Apple, controls all the pieces needed to put together what you are asking for? Sure, Google s attempting to be more like Apple as we speak, good luck with that. Otherwise, it will be the OEM's on their own, providing little bits and pieces, but no full solution, the achilles heel of Android OS.

    See how Apple rolled out ARKit, for example, and how Apple had the foresight to provide enough performance in their iPhones, even a few generations ago, to make this happen. So today's performance, that you ignore, is tomorrow's new capability, and we will see, again, Android OS fall behind.


    My life has been a lot happier since I put Avon b7 in my ignore list.
    edited September 2017 ericthehalfbeebrucemc
  • Reply 70 of 119
    wizard69wizard69 Posts: 13,377member
    mizhou said:
    wizard69 said:
    I'd like to know how they lowered power while increasing performance in the GPU for one.
    Avieshek said:
    From TSMC's higher Fabrication Processes.

    Avieshek has a valid point there, and also the article mentioned that Apple earlier used graphics hardware from other (Imagination Tech) that had support for decoding WMV and other formats. Even if Apple didn't use these features in earlier hardware, I guess there were a lot of transistors on those chips that consumed some power, even though they weren't in use.

    If Apple just builds a chip with exactly what is needed, then they can also make that chip more power efficient.

    It is the same process as the A10X uses yet outclasses that processor isn several ways.

    As for transistors this response relates to questions about the 32 bit support.   Apple could have completely deleted all support for the old ARM 32 bit instructions set in favor of the new 64 bit architecture.   If they did that could explain a lot with respect to power savings and performance improvements.   The GPU of course is completely new so that is also an area ripe for power savings.
  • Reply 71 of 119
    wizard69wizard69 Posts: 13,377member

    tipoo said:
    AMD and now Nvidia also licence out their patents, and they both manage tile based rendering without Imagination, so I wonder if one of them got a royalty deal from Apple to use TBR. 
    Could be.   Apple did hire a massive number of AMD engineers back when AMD was really on the ropes.
  • Reply 72 of 119
    avon b7avon b7 Posts: 7,622member
    tmay said:
    asdasd said:
    tmay said:
    crosslad said:
    "Google  - aimed at hitting an average selling price of less than $300—Android One phones have an aggressive price target of $100." 

    Not entirely true nowadays; the Pixel phones were priced the same as the iPhones and the new MotoX4, running Android One is around $400. This has got to be good for Apple. I would not pay the same price for an android phone as an iPhone. Now that Apple has phones at every price point  I can only see Apple sales increasing.

    Average selling price of Androids, including Pixel, is actually below $200. Even Samsung’s ASP, which includes its premium-priced Galaxy phones that sell in the 10s of millions, is now at $227. 

    http://www.androidauthority.com/price-gap-samsung-apple-smartphones-769772/

    This isn’t a new development. iPhone ASP has been pretty constant at or above $650 while Androids have have been below $300 since 2013. 

    Having a vanity model that sells in tiny quantities has little impact on an average, whether Pixel or Virtu. 



    The few members of the Church of Marketshare on AI will not acknowledge those facts.
    When Tim Cook joins that church, as he is clearly indicating, you will all rush to be members. 
    The Church of Marketshare is all about maximizing share, not about profits, so, I'm not seeing what you are wrt Tim Cook.
    You have been told over and over that that isn't the case. If Apple wants to pull more revenues (services etc) off its iDevices and ship more accessories, it will need more share.

    More share also gives them more leverage in negotiations with the likes of banks, content providers etc.

    They have been moving in that direction for a couple of years now, just very slowly. The new product/price spread is another step in that direction.

    When smartphones in their current guise hit a wall on features and build quality, more and more sales went into affordable premium and away from ultra premium. There was a risk of stagnation as everyone in the developed world who needed a phone, already had one. That meant that manufacturers had to fight to attract users from competitors. The same thing happened with TVs but the upgrade cycle for smartphones is far shorter so competition is fierce.

    This year, is without a shadow of a doubt, the fiercest yet for Apple in smartphones. Even with the A11. A little ironic perhaps but there you go.

    What can Apple do?

    Firstly, what it is doing now. Move its offer down to where the real growth seems to be, which isn't (or at least wasn't) ultra premium. Should they abandon ultra premium? Of course not. Even Huawei has a 1,400€ phone.

    You are in business to make money. There are hundreds of smaller brands making money from even low end Android phones but if you want to ship in volume you need a spread of phones.

    If you want each phone to be a potential source of after sale revenues, you should be looking for share - even if it risks eating into your upper lines. Apple is doing this but hedging it's bets a little by offering older hardware a lower price points or slightly better hardware (the case of the SE at launch) but in a very small screen phone (knowing that the higher demand was in the larger screen phones).

    I am not going to criticise this move because I think it is right given the market circumstances - today.

    edited September 2017
  • Reply 73 of 119

    The tech world is in for a surprise once they find out what really lies beneath the hood of the iPhone 8 and iPhone 8 Plus like the High Efficiency decoder that greatly reduce the space taken up by high-resolution photos and videos and the new Neural Engine that is credited with giving the A11 Bionic its name. "Bionic" generally refers to a human having electromechanical enhancements, and suggests the idea of superhuman abilities due to those enhancements. All this and more provides a user interface that is geared towards performance and sophisticated multiprocessing support. 

    Like Daniel said ‘One of the remaining interesting facets of the A11 Bionic is that if you stack together the parts of the chip that Apple detailed, there's a whole lot of surface area on it that remains completely mysterious.’

    Thanks Daniel for your peak below the hood. 

  • Reply 74 of 119
    avon b7avon b7 Posts: 7,622member
    Wasn’t Huawei bragging about their Kirin 970 - trying to steal some thunder just before Apples event?

    Sorry, Huawei, you’re not the first to put in a neural processor. You were the first to “announce” it, but it’s all vaporware until you ship devices (which Apple is).

    Even so, it’s not that big of a deal - neural processors are not that difficult to design (compared to a CPU or GPU) since they are optimized to perform a few basic tasks really well. The 970 still uses inferior ARM cores (instead of designing their own) or Mali GPUs (again, instead of designing their own). If they really wanted to impress people they should have made a completely custom SoC with their own CPU/GPU cores.

    They took the easy way out by designing a much simpler component (nueral procesdor) and trying to make it appear more relevant than it is to make up for their lack of ability to design their own CPU/GPU.
    Vapourware? 

    They didn't announce it and say, it's coming at some point.

    They announced it, they showed it on stage and in the hands of the president, they had boards running the chip right there on the show floor, they had it in the hands of independent testers to test things like the CAT18 modem.

    Vapourware? Yeah, whatever you say.

    They even gave the date for the announcement of the first phone that will use it. It wasn't earlier because the Mate series is announced in October/November. If they wanted to be first with it shipping in a phone, they could. The processor has been in mass production for a while now and they need far less than Apple. If it isn't available today it's because they set a date and don't need to change it.

    So you think neural processors are easy to design? They've been around for a while - but on mobile? No. If it were so easy, everybody would have one. They don't.

    This is major step for whoever gets them onto a SoC. Major, but only the first one. Now software is needed.

    Huawei develops their own SoCs. They could easily design their own CPU/GPUs too. They choose not too.

    Be careful what you wish for. When the US government scuppered a deal for intel to ship Xeons to China for use in supercomputers the Chinese said, OK we'll brew our own. The result was the SW26010 and look what happened then. 

    Do you know why Apple depends on other people's designs (not only manufacturing capacity) so much for key elements of its hardware? Because it chooses to. Just like Huawei. Just like everyone that has the resources to go it alone but chooses not to. There are things you prefer to do alone and others you don't.

    Huawei is not using the latest ARM cores because it chose not to use them on the Kirin 970. It will use the latest cores on the following design. Using them on the Kirin 970 would have delayed the release. There are other reasons too. With this decision they still have a far better modem than anyone else, dual ISPs, a better sensor controller, the NPU, better efficiencies, far better GPU etc.

    And even with the 960 (and below) they recently became the world's second largest handset manufacturer, moving past Apple. They even decided to pull out of the low end.

    Apple designs its own chips because it wants differentiation with the rest of the market and is still using external IP on the GPUs on most of its phones and in many other areas. Let's face it, Apple has its biggest eggs in one basket. If the bottom falls out of that basket it wouldn't be a disaster for users, just Apple and Wall Street. The company would still be a billion dollar company. The same cannot be said of Huawei. Until relatively recently it wasn't even in the handset business. Huawei invests a lot in R&D but still has many agreements with other companies. Its phones are the fruit of many different IP agreements and of all kinds. Some are exclusive and some are not. Some involve shared IP and some don't.

    Perhaps you are focusing too much on benchmarks and not seeing the bigger picture. Don't forget that Apple has had to play catch-up in many areas with their new phones. 

    edited September 2017
  • Reply 75 of 119
    tmaytmay Posts: 6,309member
    avon b7 said:
    Wasn’t Huawei bragging about their Kirin 970 - trying to steal some thunder just before Apples event?

    Sorry, Huawei, you’re not the first to put in a neural processor. You were the first to “announce” it, but it’s all vaporware until you ship devices (which Apple is).

    Even so, it’s not that big of a deal - neural processors are not that difficult to design (compared to a CPU or GPU) since they are optimized to perform a few basic tasks really well. The 970 still uses inferior ARM cores (instead of designing their own) or Mali GPUs (again, instead of designing their own). If they really wanted to impress people they should have made a completely custom SoC with their own CPU/GPU cores.

    They took the easy way out by designing a much simpler component (nueral procesdor) and trying to make it appear more relevant than it is to make up for their lack of ability to design their own CPU/GPU.
    Vapourware? 

    They didn't announce it and say, it's coming at some point.

    They announced it, they showed it on stage and in the hands of the president, they had boards running the chip right there on the show floor, they had it in the hands of independent testers to test things like the CAT18 modem.

    Vapourware? Yeah, whatever you say.

    They even gave the date for the announcement of the first phone that will use it. It wasn't earlier because the Mate series is announced in October/November. If they wanted to be first with it shipping in a phone, they could. The processor has been in mass production for a while now and they need far less than Apple. If it isn't available today it's because they set a date and don't need to change it.

    So you think neural processors are easy to design? They've been around for a while - but on mobile? No. If it were so easy, everybody would have one. They don't.

    This is major step for whoever gets them onto a SoC. Major, but only the first one. Now software is needed.

    Huawei develops their own SoCs. They could easily design their own CPU/GPUs too. They choose not too.

    Be careful what you wish for. When the US government scuppered a deal for intel to ship Xeons to China for use in supercomputers the Chinese said, OK we'll brew our own. The result was the SW26010 and look what happened then. 

    Do you know why Apple depends on other people's designs (not only manufacturing capacity) so much for key elements of its hardware? Because it chooses to. Just like Huawei. Just like everyone that has the resources to go it alone but chooses not to. There are things you prefer to do alone and others you don't.

    Huawei is not using the latest ARM cores because it chose not to use them on the Kirin 970. It will use the latest cores on the following design. Using them on the Kirin 970 would have delayed the release. There are other reasons too. With this decision they still have a far better modem than anyone else, dual ISPs, a better sensor controller, the NPU, better efficiencies, far better GPU etc.

    And even with the 960 (and below) they recently became the world's second largest handset manufacturer, moving past Apple. They even decided to pull out of the low end.

    Apple designs its own chips because it wants differentiation with the rest of the market and is still using external IP on the GPUs on most of its phones and in many other areas. Let's face it, Apple has its biggest eggs in one basket. If the bottom falls out of that basket it wouldn't be a disaster for users, just Apple and Wall Street. The company would still be a billion dollar company. The same cannot be said of Huawei. Until relatively recently it wasn't even in the handset business. Huawei invests a lot in R&D but still has many agreements with other companies. Its phones are the fruit of many different IP agreements and of all kinds. Some are exclusive and some are not. Some involve shared IP and some don't.

    Perhaps you are focusing too much on benchmarks and not seeing the bigger picture. Don't forget that Apple has had to play catch-up in many areas with their new phones. 

    Bold move coming out as a Huawei shill three weeks before they deliver the "knockout blow" to Apple. 

    Me, I'll be making popcorn. 
    ericthehalfbeewilliamlondonargonaut
  • Reply 76 of 119
    asdasdasdasd Posts: 5,686member
    tmay said:
    asdasd said:
    tmay said:
    crosslad said:
    "Google  - aimed at hitting an average selling price of less than $300—Android One phones have an aggressive price target of $100." 

    Not entirely true nowadays; the Pixel phones were priced the same as the iPhones and the new MotoX4, running Android One is around $400. This has got to be good for Apple. I would not pay the same price for an android phone as an iPhone. Now that Apple has phones at every price point  I can only see Apple sales increasing.

    Average selling price of Androids, including Pixel, is actually below $200. Even Samsung’s ASP, which includes its premium-priced Galaxy phones that sell in the 10s of millions, is now at $227. 

    http://www.androidauthority.com/price-gap-samsung-apple-smartphones-769772/

    This isn’t a new development. iPhone ASP has been pretty constant at or above $650 while Androids have have been below $300 since 2013. 

    Having a vanity model that sells in tiny quantities has little impact on an average, whether Pixel or Virtu. 



    The few members of the Church of Marketshare on AI will not acknowledge those facts.
    When Tim Cook joins that church, as he is clearly indicating, you will all rush to be members. 
    The Church of Marketshare is all about maximizing share, not about profits, so, I'm not seeing what you are wrt Tim Cook.
    Apple never say anything about being a profit seeking company. They deny it in fact. 

    However they are increasingly getting more revenue from services, for that to continue unit sales must increase. 
  • Reply 77 of 119
    asdasdasdasd Posts: 5,686member

    avon b7 said:
    Wasn’t Huawei bragging about their Kirin 970 - trying to steal some thunder just before Apples event?

    Sorry, Huawei, you’re not the first to put in a neural processor. You were the first to “announce” it, but it’s all vaporware until you ship devices (which Apple is).

    Even so, it’s not that big of a deal - neural processors are not that difficult to design (compared to a CPU or GPU) since they are optimized to perform a few basic tasks really well. The 970 still uses inferior ARM cores (instead of designing their own) or Mali GPUs (again, instead of designing their own). If they really wanted to impress people they should have made a completely custom SoC with their own CPU/GPU cores.

    They took the easy way out by designing a much simpler component (nueral procesdor) and trying to make it appear more relevant than it is to make up for their lack of ability to design their own CPU/GPU.
    Vapourware? 

    They didn't announce it and say, it's coming at some point.

    They announced it, they showed it on stage and in the hands of the president, they had boards running the chip right there on the show floor, they had it in the hands of independent testers to test things like the CAT18 modem.

    Vapourware? Yeah, whatever you say.

    They even gave the date for the announcement of the first phone that will use it. It wasn't earlier because the Mate series is announced in October/November. If they wanted to be first with it shipping in a phone, they could. The processor has been in mass production for a while now and they need far less than Apple. If it isn't available today it's because they set a date and don't need to change it.

    So you think neural processors are easy to design? They've been around for a while - but on mobile? No. If it were so easy, everybody would have one. They don't.

    This is major step for whoever gets them onto a SoC. Major, but only the first one. Now software is needed.

    Huawei develops their own SoCs. They could easily design their own CPU/GPUs too. They choose not too.

    Be careful what you wish for. When the US government scuppered a deal for intel to ship Xeons to China for use in supercomputers the Chinese said, OK we'll brew our own. The result was the SW26010 and look what happened then. 

    Do you know why Apple depends on other people's designs (not only manufacturing capacity) so much for key elements of its hardware? Because it chooses to. Just like Huawei. Just like everyone that has the resources to go it alone but chooses not to. There are things you prefer to do alone and others you don't.

    Huawei is not using the latest ARM cores because it chose not to use them on the Kirin 970. It will use the latest cores on the following design. Using them on the Kirin 970 would have delayed the release. There are other reasons too. With this decision they still have a far better modem than anyone else, dual ISPs, a better sensor controller, the NPU, better efficiencies, far better GPU etc.

    And even with the 960 (and below) they recently became the world's second largest handset manufacturer, moving past Apple. They even decided to pull out of the low end.

    Apple designs its own chips because it wants differentiation with the rest of the market and is still using external IP on the GPUs on most of its phones and in many other areas. Let's face it, Apple has its biggest eggs in one basket. If the bottom falls out of that basket it wouldn't be a disaster for users, just Apple and Wall Street. The company would still be a billion dollar company. The same cannot be said of Huawei. Until relatively recently it wasn't even in the handset business. Huawei invests a lot in R&D but still has many agreements with other companies. Its phones are the fruit of many different IP agreements and of all kinds. Some are exclusive and some are not. Some involve shared IP and some don't.

    Perhaps you are focusing too much on benchmarks and not seeing the bigger picture. Don't forget that Apple has had to play catch-up in many areas with their new phones. 

    Is there a HuaweiInsider where you can tell them all about the A11 Chip?
    tmay
  • Reply 78 of 119
    tmaytmay Posts: 6,309member
    asdasd said:
    tmay said:
    asdasd said:
    tmay said:
    crosslad said:
    "Google  - aimed at hitting an average selling price of less than $300—Android One phones have an aggressive price target of $100." 

    Not entirely true nowadays; the Pixel phones were priced the same as the iPhones and the new MotoX4, running Android One is around $400. This has got to be good for Apple. I would not pay the same price for an android phone as an iPhone. Now that Apple has phones at every price point  I can only see Apple sales increasing.

    Average selling price of Androids, including Pixel, is actually below $200. Even Samsung’s ASP, which includes its premium-priced Galaxy phones that sell in the 10s of millions, is now at $227. 

    http://www.androidauthority.com/price-gap-samsung-apple-smartphones-769772/

    This isn’t a new development. iPhone ASP has been pretty constant at or above $650 while Androids have have been below $300 since 2013. 

    Having a vanity model that sells in tiny quantities has little impact on an average, whether Pixel or Virtu. 



    The few members of the Church of Marketshare on AI will not acknowledge those facts.
    When Tim Cook joins that church, as he is clearly indicating, you will all rush to be members. 
    The Church of Marketshare is all about maximizing share, not about profits, so, I'm not seeing what you are wrt Tim Cook.
    Apple never say anything about being a profit seeking company. They deny it in fact. 

    However they are increasingly getting more revenue from services, for that to continue unit sales must increase. 
    Unit sales are increasing along the same trend line they have been on for years . To imply or state otherwise is an absolute falsehood. 

    http://www.businessinsider.com/apple-iphone-sales-yearly-chart-2017-2
  • Reply 79 of 119
    tmay said:
    avon b7 said:
    Wasn’t Huawei bragging about their Kirin 970 - trying to steal some thunder just before Apples event?

    Sorry, Huawei, you’re not the first to put in a neural processor. You were the first to “announce” it, but it’s all vaporware until you ship devices (which Apple is).

    Even so, it’s not that big of a deal - neural processors are not that difficult to design (compared to a CPU or GPU) since they are optimized to perform a few basic tasks really well. The 970 still uses inferior ARM cores (instead of designing their own) or Mali GPUs (again, instead of designing their own). If they really wanted to impress people they should have made a completely custom SoC with their own CPU/GPU cores.

    They took the easy way out by designing a much simpler component (nueral procesdor) and trying to make it appear more relevant than it is to make up for their lack of ability to design their own CPU/GPU.
    Vapourware? 

    They didn't announce it and say, it's coming at some point.

    They announced it, they showed it on stage and in the hands of the president, they had boards running the chip right there on the show floor, they had it in the hands of independent testers to test things like the CAT18 modem.

    Vapourware? Yeah, whatever you say.

    They even gave the date for the announcement of the first phone that will use it. It wasn't earlier because the Mate series is announced in October/November. If they wanted to be first with it shipping in a phone, they could. The processor has been in mass production for a while now and they need far less than Apple. If it isn't available today it's because they set a date and don't need to change it.

    So you think neural processors are easy to design? They've been around for a while - but on mobile? No. If it were so easy, everybody would have one. They don't.

    This is major step for whoever gets them onto a SoC. Major, but only the first one. Now software is needed.

    Huawei develops their own SoCs. They could easily design their own CPU/GPUs too. They choose not too.

    Be careful what you wish for. When the US government scuppered a deal for intel to ship Xeons to China for use in supercomputers the Chinese said, OK we'll brew our own. The result was the SW26010 and look what happened then. 

    Do you know why Apple depends on other people's designs (not only manufacturing capacity) so much for key elements of its hardware? Because it chooses to. Just like Huawei. Just like everyone that has the resources to go it alone but chooses not to. There are things you prefer to do alone and others you don't.

    Huawei is not using the latest ARM cores because it chose not to use them on the Kirin 970. It will use the latest cores on the following design. Using them on the Kirin 970 would have delayed the release. There are other reasons too. With this decision they still have a far better modem than anyone else, dual ISPs, a better sensor controller, the NPU, better efficiencies, far better GPU etc.

    And even with the 960 (and below) they recently became the world's second largest handset manufacturer, moving past Apple. They even decided to pull out of the low end.

    Apple designs its own chips because it wants differentiation with the rest of the market and is still using external IP on the GPUs on most of its phones and in many other areas. Let's face it, Apple has its biggest eggs in one basket. If the bottom falls out of that basket it wouldn't be a disaster for users, just Apple and Wall Street. The company would still be a billion dollar company. The same cannot be said of Huawei. Until relatively recently it wasn't even in the handset business. Huawei invests a lot in R&D but still has many agreements with other companies. Its phones are the fruit of many different IP agreements and of all kinds. Some are exclusive and some are not. Some involve shared IP and some don't.

    Perhaps you are focusing too much on benchmarks and not seeing the bigger picture. Don't forget that Apple has had to play catch-up in many areas with their new phones. 

    Bold move coming out as a Huawei shill three weeks before they deliver the "knockout blow" to Apple. 

    Me, I'll be making popcorn. 

    Didn't you have this troll on ignore? Might to time to add him back. Nothing upsets a troll more than watching their posts wither away without a response.
    williamlondon
  • Reply 80 of 119
    tmaytmay Posts: 6,309member
    tmay said:
    avon b7 said:
    Wasn’t Huawei bragging about their Kirin 970 - trying to steal some thunder just before Apples event?

    Sorry, Huawei, you’re not the first to put in a neural processor. You were the first to “announce” it, but it’s all vaporware until you ship devices (which Apple is).

    Even so, it’s not that big of a deal - neural processors are not that difficult to design (compared to a CPU or GPU) since they are optimized to perform a few basic tasks really well. The 970 still uses inferior ARM cores (instead of designing their own) or Mali GPUs (again, instead of designing their own). If they really wanted to impress people they should have made a completely custom SoC with their own CPU/GPU cores.

    They took the easy way out by designing a much simpler component (nueral procesdor) and trying to make it appear more relevant than it is to make up for their lack of ability to design their own CPU/GPU.
    Vapourware? 

    They didn't announce it and say, it's coming at some point.

    They announced it, they showed it on stage and in the hands of the president, they had boards running the chip right there on the show floor, they had it in the hands of independent testers to test things like the CAT18 modem.

    Vapourware? Yeah, whatever you say.

    They even gave the date for the announcement of the first phone that will use it. It wasn't earlier because the Mate series is announced in October/November. If they wanted to be first with it shipping in a phone, they could. The processor has been in mass production for a while now and they need far less than Apple. If it isn't available today it's because they set a date and don't need to change it.

    So you think neural processors are easy to design? They've been around for a while - but on mobile? No. If it were so easy, everybody would have one. They don't.

    This is major step for whoever gets them onto a SoC. Major, but only the first one. Now software is needed.

    Huawei develops their own SoCs. They could easily design their own CPU/GPUs too. They choose not too.

    Be careful what you wish for. When the US government scuppered a deal for intel to ship Xeons to China for use in supercomputers the Chinese said, OK we'll brew our own. The result was the SW26010 and look what happened then. 

    Do you know why Apple depends on other people's designs (not only manufacturing capacity) so much for key elements of its hardware? Because it chooses to. Just like Huawei. Just like everyone that has the resources to go it alone but chooses not to. There are things you prefer to do alone and others you don't.

    Huawei is not using the latest ARM cores because it chose not to use them on the Kirin 970. It will use the latest cores on the following design. Using them on the Kirin 970 would have delayed the release. There are other reasons too. With this decision they still have a far better modem than anyone else, dual ISPs, a better sensor controller, the NPU, better efficiencies, far better GPU etc.

    And even with the 960 (and below) they recently became the world's second largest handset manufacturer, moving past Apple. They even decided to pull out of the low end.

    Apple designs its own chips because it wants differentiation with the rest of the market and is still using external IP on the GPUs on most of its phones and in many other areas. Let's face it, Apple has its biggest eggs in one basket. If the bottom falls out of that basket it wouldn't be a disaster for users, just Apple and Wall Street. The company would still be a billion dollar company. The same cannot be said of Huawei. Until relatively recently it wasn't even in the handset business. Huawei invests a lot in R&D but still has many agreements with other companies. Its phones are the fruit of many different IP agreements and of all kinds. Some are exclusive and some are not. Some involve shared IP and some don't.

    Perhaps you are focusing too much on benchmarks and not seeing the bigger picture. Don't forget that Apple has had to play catch-up in many areas with their new phones. 

    Bold move coming out as a Huawei shill three weeks before they deliver the "knockout blow" to Apple. 

    Me, I'll be making popcorn. 

    Didn't you have this troll on ignore? Might to time to add him back. Nothing upsets a troll more than watching their posts wither away without a response.
    I guessing that now would be a good time to add him back on the list.

    Done.

    I'm actually happy that he exposed himself as a shill on AI. Makes it easier now to disregard him.
    edited September 2017 williamlondon
Sign In or Register to comment.