I had not expected that Apple would give away iWork. Guess they're not getting enough traction on selling it. I see it in iCloud and haven't given it much of a spin yet. But it makes sense. Now is a good opportunity to take advantage of Microsoft's lumbering steps into the tablet world that already exists. It just showcases how far ahead the iOS ecosystem remains.
Microsoft Office isn't dominant because it's the best. It's because they're great at licensing and catering to the communities of developers, instructors and IT support staff. </soapbox>
Actually, iWork is very popular. It's been consistently at the top of the paid listings as individual apps.
What Apple is doing is cutting Microsoft off at the pass. iOS 7 has many more business features than ever, and is being looked at very favorably by business and government. iWork is also the most popular, by far, office suit in mobile for business and government.
Microsoft tried to play the Office card with Surface RT but has failed so far. A problem with it is that the license only allows personal use, and nothing for business. It's even being questioned whether the license allows one to write their personal business resume with it. Meanwhile, you can do anything with iWork that you need.
The Surface Pro, meanwhile, costs far more, and doesn't include a free Office, you need to pay for it.
This is a major push for Apple. When you include that in June Apple said that there would be a major upgrade to iWork coming, and introduced their cloud suite as well, we can see where Apple is going here. Apple sells hardware, and their software is there for the purpose of making the hardware more desirable. Apple is now big enough that the few hundreds of millions a year they may make from this has become inconsequential. But the extra hardware they may sell could end up in the billions. Remembering that they only sell apps once in iOS (free upgrades), the income from that software is trivial.
They are taking a page out of Microsoft's book, which is that of giving software away to hobble a competitor who can't afford to do the same. That's how Microsoft destroyed Netscape, and how they took the presentation program crown away from Adobe.
The shoe is now on the other foot, as the income from Office is not only important to Microsoft, but is their biggest division, with the greatest sales and profits. If Apple can prevent them from making money on it in mobile, they will cut off a major source of funding for Microsoft, and keep them out of the market, while solidifying Apple's strong hold on both business and government mobile markets.
Remembering the old adage about Office, which is that 80% of the users use 20% of the features, iWork doesn't need to be 100% equal to Office to take much of their user base away. They need to be mostly compatible, and that they are already. If the new upgrade to iWork moves them close enough, then many workers and home users will find the $99 per year license for those who already have Office giving them the ability to access it on their iPads and iPhones to be very overpriced, and unnecessary.
We now read that Ballmer has just said that they are working on Office for the iPad. But that it's waiting for their touch version which will first come out for the Windows suit in mid 2014, with the ipad version not debuting until the end of the year. That at least a year from now. It may already be too late, but a year?
And then what will it cost? If it's tied to the desktop versions in license as I just mentioned, it's way too much. If it's much cheaper, Microsoft will have an income problem, and people will demand they lower the price for the desktop version as well. Uh oh! That's a major problem for Microsoft.
So after that long winded explanation, I hope you get a better idea of what's really going on.
A convertible is a laptop that turns into a tablet, those are just cases. Maybe Apple will make a keyboard for the iPad someday, but I doubt it. And if they ever make a convertible, they've lost their way.
Apple will not release a product that is dramatically slower than its predecessor. An iPad convertible (what I've been calling an 'Ipad Pro' for the year or more that I've been talking about it) would not be a replacement for the MBA, but would rather be an extension of the iPad line. Although some people might buy it instead of the MBA, it wouldn't be marketed as an MBA.
You know, I've been giving this some thought myself. It's really an interesting question.
What would Apple need to do to come out with a notebook that used a more powerful version of the A7? Well, they would need all the software that most people would use a notebook for. That would be browsing, writing, homework, content watching and listening, games, finance, reading books, magazines, etc., photo editing, movie editing, painting, drawing, CAD, and other 3D work.
In other words, all the things that can be now done with iOS apps. Well, that's just amazing! What advantage would we get from this device though? That hard to say. Apple, so far, has been against the touch screen notebook format, so would that change? With Apple, no one will ever know unless it does. I would say; "The Shadow knows." But then, the Shadow is dead.
Could some version of OS X run on a more mature version of the A series chip? It's possible. Apple demonstrated, and Anandtech showed, that the A7 is equal to the mid Bay Trail Atom, and pretty close to the top chip there. Next year, if Apple can keep up their advances, it will push the mobile i3. And remember that it's still two core. If it were four core today, it would already be pushing the mobile i3.
Even if Apple must slow speed advances over the next few years, they will still be advancing much faster than intel can, and will have a chip, even with the small batteries iPad must use, that will be getting close to serious performance in the form of Intel's mobile i5 line.
But I see much more interesting things happening here. While we never know what's happening in Apple's labs, I get a feeling that Apple is doing more with the A line than just making them better iOS engines.
Apple, being unique in that they are the only ones today making both the OS and the chips, has an opportunity to do something special. What they can do is to look at the way x86 processes and the way ARM processes, and modify, or even change some of ARM's instructions to mimic x86. Those could be the areas in which an OS and its apps would have the most work in an emulation situation, and would therefor slow that emulation down the most. Remember that Apple has a lot of experience in this area.
If they could work out the worst performing areas, emulation might not be nearly as much of a chore. Giving the chip more power could allow Apple to raise the speed at which the chip is running. Right now, the A7 is running at the same 1.3Gz as the A6. But is twice as powerful. Many other top ARM designs are already running well beyond 2Gz, up to 2.7Gz. Imagine if Apple doubled that speed to 2.6Gz! Around twice the performance. In a notebook, that could be done because cooling would be far better. Of course, going by past experience, Apple will raise that to maybe 1.5Gz in the iPad, so we can talk about 3 Gz.
Now figure a four core version, with a newer design as well. I could easily see an A8x, if Apple were to do it, that could run at 3Gz in a notebook that would be at least ten times as powerful as today's A7, and more so in graphics, if they use the iPad's extra graphics abilities. That's already pushing the mobile i7 line.
It could be done. Whether Apple wants to do it is another affair, but the fact that they said the A7 had desktop performance shows that they may be thinking it. It could have been a tease.
Not sure you followed me. I'm suggesting that it would have been easier for Apple to increase clock speed and/or the number of cores on the A6X without the major upgrade to 64 bit. However, if this SOC is intended for uses on other devices with greater processing AND RAM requirements, then this would explain why Apple chose the more complex path.
I don't know much about PA Semi and Intrinsity, just remember reading some high level announcement when they were bought by Apple. Didn't realise they were that good!
Easier... nope.
In fact, the 64bit solution is amazingly elegant from a hardware and compiler level. and the dual 32/64 bit SoC made the upgrade 'less major.' The major issue is purely in the OS, and apple has solved that a few times already...
In short, this solved a problem now. I don't see this chip or the next turn of it being 'workstation worthy.' But a MacBook Air is probably sporting an A7 as we speak, and being shown to Intel as a 'We expect you to keep your price performance in line, as we do have options";-)
As noted elsewhere, the key benefit in 64 bit A7 is the handling of object creation and deletion events 50% faster (primarily taking advantage of 64bit register to register 1 clock MOVs... to do the same with a 32 bit would have required doubling clock speed, unlikely in today's power envelope). And having multiple cores on a non-compute intensive device (really, how many threads can my fingers have in RUN state at a time) wouldn't have sped stuff up much if this is the perceived bottleneck to the user (responsiveness 'feeling' as I drag or touch, all those events are being created, acted on and destroyed [I saw a finger touch[highlight], it moved[drag], it moved again[drag it some more], again[move it more], again[ditto... a couple hundred times a second], it stopped[what's it doing?], it lifted off[unhighlight] and the hundreds of thousands of graphics request messages that were fired off to display this simple action]
Make that 50% faster (handle 3 events in the CPU time it used to take to do 2]. That's like turning your clock from 2ghz to 3ghz(or more...).
The problem is, Qualcomm doesn't see the problem as 'make dragging seem less 'draggy'', just 'you need a faster processor' or you have a RAM bottle neck, or your GPU isn't polygonning fast enough.... Nope... it's just needing a wider data path.
The issue is less about how great the chip guys are, and more about 'old school' computer design. Back in the day, chip guys and compiler guys and OS guys were in same company, if not the same division or same building. Today, Apple and IBM are the only ones I know of that do it now (anyone build a better Zseries Mainframe lately?). Compilers, and systems developers often overlapped teams. in 'open systems' solutions, now you got to interface waste, as each interface is generalized, instead of optimized.
Apple can build a processor _for_themselves_, for One OS, One compiler, One set of frameworks. And provide a canonical profile of how iOS COULD operate if in a chip had these characteristics. This bi-directional (SW guys telling chip guys how to build, and chip guys telling compiler guys how to exploit) synergy of a organically integrated and optimized compute model (from the silicon up to the UI) is the Strong Sauce that Apple is exploiting.
You know, I've been giving this some thought myself. It's really an interesting question.
.....
It could be done. Whether Apple wants to do it is another affair, but the fact that they said the A7 had desktop performance shows that they may be thinking it. It could have been a tease.
I've posted before that the next Rosetta will be between OS X and iOS, thus allowing Apple to have the Ax chip in the MBA and then beyond which would increase margins as well as keep the competition at bay. You can easily and quickly copy software, you can't do the same with hardware. Apple could have a preprocessor core that quickly translates x86 to ARM.
Then it is only a matter of time when devs offer both the x86 and iOS versions until such time that x86 is no longer viable on the platform. And why not? If Apple can maintain momentum with the ARM, then they control one more piece of the puzzle.
Even if Apple must slow speed advances over the next few years, they will still be advancing much faster than intel can, and will have a chip, even with the small batteries iPad must use, that will be getting close to serious performance in the form of Intel's mobile i5 line.
But I see much more interesting things happening here. While we never know what's happening in Apple's labs, I get a feeling that Apple is doing more with the A line than just making them better iOS engines.
Apple, being unique in that they are the only ones today making both the OS and the chips, has an opportunity to do something special. What they can do is to look at the way x86 processes and the way ARM processes, and modify, or even change some of ARM's instructions to mimic x86. Those could be the areas in which an OS and its apps would have the most work in an emulation situation, and would therefor slow that emulation down the most. Remember that Apple has a lot of experience in this area.
If they could work out the worst performing areas, emulation might not be nearly as much of a chore. Giving the chip more power could allow Apple to raise the speed at which the chip is running. Right now, the A7 is running at the same 1.3Gz as the A6. But is twice as powerful. Many other top ARM designs are already running well beyond 2Gz, up to 2.7Gz. Imagine if Apple doubled that speed to 2.6Gz! Around twice the performance. In a notebook, that could be done because cooling would be far better. Of course, going by past experience, Apple will raise that to maybe 1.5Gz in the iPad, so we can talk about 3 Gz.
Now figure a four core version, with a newer design as well. I could easily see an A8x, if Apple were to do it, that could run at 3Gz in a notebook that would be at least ten times as powerful as today's A7, and more so in graphics, if they use the iPad's extra graphics abilities. That's already pushing the mobile i7 line.
It could be done. Whether Apple wants to do it is another affair, but the fact that they said the A7 had desktop performance shows that they may be thinking it. It could have been a tease.
No, it's not a tease... it's an over the bow shot at Intel.
Intel is driving power demands down as quickly as Apple is driving AxX speeds up. my guess is that Intel will maintain a slight lead with the x86 line for a while. Any posturing will only be for price and functionality concessions (ie, optimizing parts of the chip to better handle OSX unique and high use instruction pipelining).
It will be 4-5 years before Apple can evolve the A series into a significant desktop compute platform, but in 4-5 years will there be a desktop?
I think it's more impressive the 'other stuff' apple integrates into their platforms. When the Laptops start getting LTE integrated in, and M7 chips, or gyros or touchID, then you'll start looking at the 'family' of apple, not iOS and OSX. A corporation that commits to touchID as it's auth standard (or someone like the DoD), and apple is able to be the laptop and the iPad, and the phone, and the iPod, and the Thumb Drive manufacturer for them.... that's leverage of the apple difference... the CPU... that's just an implementation detail
I've posted before that the next Rosetta will be between OS X and iOS, thus allowing Apple to have the Ax chip in the MBA and then beyond which would increase margins as well as keep the competition at bay. You can easily and quickly copy software, you can't do the same with hardware. Apple could have a preprocessor core that quickly translates x86 to ARM.
Then it is only a matter of time when devs offer both the x86 and iOS versions until such time that x86 is no longer viable on the platform. And why not? If Apple can maintain momentum with the ARM, then they control one more piece of the puzzle.
If Apple chooses to do what I've outlined, then emulation could be fairly easy. Not easy, but fairly easy. In the old days, it was considered necessary to have the machine doing the emulation about 5 times as powerful as he one being emulated for the feel of speed to be the same. From my experiences with emulation, I would agree.
If all Apple did was to increase the lower of their SoC, it wouldn't be enough. ARM would need to be as powerful as the most powerful mobile i7 to emulate a top mobile i3, at best. Not acceptable!
If Apple can eliminate some of the worst bottlenecks, then it could be done. Considering how expensive Intel's solutions are, and how cheap Apple's are (that is, the cost of the chips for a complete system), then Apple may have a major cost advantage. We can look to why Surface Pro is so expensive, including the new, slightly less expensive model. It's because the chips are so expensive. If Microsoft could cut the cost of the chips to a quarter, the surface Pro could cost $495 instead of $895, and who knows what profit Microsoft is getting on that price?
So could we see an arm powered OS X notebook at some point? It's not impossible. There is a way. Whether there is the will is something else.
No, it's not a tease... it's an over the bow shot at Intel.
Intel is driving power demands down as quickly as Apple is driving AxX speeds up. my guess is that Intel will maintain a slight lead with the x86 line for a while. Any posturing will only be for price and functionality concessions (ie, optimizing parts of the chip to better handle OSX unique and high use instruction pipelining).
It will be 4-5 years before Apple can evolve the A series into a significant desktop compute platform, but in 4-5 years will there be a desktop?
I think it's more impressive the 'other stuff' apple integrates into their platforms. When the Laptops start getting LTE integrated in, and M7 chips, or gyros or touchID, then you'll start looking at the 'family' of apple, not iOS and OSX. A corporation that commits to touchID as it's auth standard (or someone like the DoD), and apple is able to be the laptop and the iPad, and the phone, and the iPod, and the Thumb Drive manufacturer for them.... that's leverage of the apple difference... the CPU... that's just an implementation detail
When I said tease, I meant to the developers sitting in the hall listening to the presentation, and to Apple users, and prospective Apple users watching the video of it later on.
I don't see it needing four or five years. And as I said, it's already performing at the top of the brand new Bay Trail Atom line.
Apple can do things to their OS and chips that neither Microsoft nor Intel can do, because Intel needs to have their chips perform well with a number of OS's, from the cheapest tablets to the most expensive supercomputers.
Apple can break whatever they want to, and have, if they find a major increase in performance warrants it. This can lead to much greater advances that Intel can manage with their need for backward compatibility.
Just to add to this, Apple could also retain the old instructions. Detect OS X, run this, detect iOS, run that.
“The comments made by Anand Chandrasekher, Qualcomm CMO, about 64-bit computing were inaccurate,” said a Qualcomm spokesperson in an email. “The mobile hardware and software ecosystem is already moving in the direction of 64-bit; and, the evolution to 64-bit brings desktop class capabilities and user experiences to mobile, as well as enabling mobile processors and software to run new classes of computing devices.”
Agam Shah, reporting for IDG:
“The comments made by Anand Chandrasekher, Qualcomm CMO, about 64-bit computing were inaccurate,” said a Qualcomm spokesperson in an email. “The mobile hardware and software ecosystem is already moving in the direction of 64-bit; and, the evolution to 64-bit brings desktop class capabilities and user experiences to mobile, as well as enabling mobile processors and software to run new classes of computing devices.”
Qualcomm did not provide further comment.
I'm happy to see Qualcomm again, is being honest about something. They're a pretty good company. They resisted both Motorola and Samsung's efforts to get them to improperly sue Apple over patents Qualcomm licensed from them.
Since we are wishing for our Dream Machine, I wish for a 13" Mac that has a touch screen and is capable of running OS X, iOS and a virtual machine. A Mac that has the ability to switch between processors or virtual iOS & Windows capability. I just don't think iOS is very close to being what we want in a Mac, but I sure would like to occasionally run some iOS apps on my Mac, either as some type of switched processor task or virtual type of machine.
Just wishing to have a Mac that can run A7 iOS tasks and OS X.
Since we are wishing for our Dream Machine, I wish for a 13" Mac that has a touch screen and is capable of running OS X, iOS and a virtual machine. A Mac that has the ability to switch between processors or virtual iOS & Windows capability. I just don't think iOS is very close to being what we want in a Mac, but I sure would like to occasionally run some iOS apps on my Mac, either as some type of switched processor task or virtual type of machine.
Just wishing to have a Mac that can run A7 iOS tasks and OS X.
Since the iOS SoC is so cheap, they could add one to a Mac. I keep wondering if they ever will.
A larger 13" iPad would be cool, I have no objections against that, but screw the notebook part. Leave it as an iPad, large and super thin, with an even bigger screen! I think that there is definitely a market of people who would be interested in a larger iPad, and who would be willing to pay for such an iPad.
If people want lame, failed, frankenstein contraptions, like the Surface tablets, then go buy one of those. I hope that Apple doesn't make any so-called convertibles or hybrid messes. Those already exist, and why should Apple make such a failed and terrible item?
If you want a laptop, go buy a Macbook Pro or Macbook Air.
These clueless analysts should just STFU. Apple is doing just fine without their braindead ideas and dumb predictions. A hybrid piece of junk? Forget it, Apple has better things to do with their time.
For you and all the other people vehemently trashing this idea - maybe you should start by realizing that not everyone has the same needs.
I can see a great demand for the product. When I'm traveling, an extra pound or two is a pain. An iPad does about 90% of what I need to do when I'm traveling, but I still end up carrying an iPad PLUS my MBP. With an iPad Pro (10-13" screen, keyboard, but otherwise an iPad), I'd only need one device and could leave the MBP at home. For that matter, I'd have an iMac on my desk rather than an MBP in the first place. An iMac and iPad Pro would be the ideal computing environment for me - and a lot of business executives I know.
Comments
That doesn't mean that the concept is bad. Dell have just announced a new range of hybrids which seem to be a better proposition than the Surface 2.
Besides, and iPad with a Bluetooth keyboard IS a sort of hybrid. It's just that it wasn't designed as a hybrid from the outset.
Stop. This is not even a rumor. It is fantasy.
Oh, the ‘hybrid’, you mean? Yeah, the concept is bad. You’re attaching mechanical legs to a car.
What are you going to do when Apple makes a hybrid?
So, what you are you saying is that adding a keyboard to an iPad is like attaching mechanical legs to a car?
Actually, iWork is very popular. It's been consistently at the top of the paid listings as individual apps.
What Apple is doing is cutting Microsoft off at the pass. iOS 7 has many more business features than ever, and is being looked at very favorably by business and government. iWork is also the most popular, by far, office suit in mobile for business and government.
Microsoft tried to play the Office card with Surface RT but has failed so far. A problem with it is that the license only allows personal use, and nothing for business. It's even being questioned whether the license allows one to write their personal business resume with it. Meanwhile, you can do anything with iWork that you need.
The Surface Pro, meanwhile, costs far more, and doesn't include a free Office, you need to pay for it.
This is a major push for Apple. When you include that in June Apple said that there would be a major upgrade to iWork coming, and introduced their cloud suite as well, we can see where Apple is going here. Apple sells hardware, and their software is there for the purpose of making the hardware more desirable. Apple is now big enough that the few hundreds of millions a year they may make from this has become inconsequential. But the extra hardware they may sell could end up in the billions. Remembering that they only sell apps once in iOS (free upgrades), the income from that software is trivial.
They are taking a page out of Microsoft's book, which is that of giving software away to hobble a competitor who can't afford to do the same. That's how Microsoft destroyed Netscape, and how they took the presentation program crown away from Adobe.
The shoe is now on the other foot, as the income from Office is not only important to Microsoft, but is their biggest division, with the greatest sales and profits. If Apple can prevent them from making money on it in mobile, they will cut off a major source of funding for Microsoft, and keep them out of the market, while solidifying Apple's strong hold on both business and government mobile markets.
Remembering the old adage about Office, which is that 80% of the users use 20% of the features, iWork doesn't need to be 100% equal to Office to take much of their user base away. They need to be mostly compatible, and that they are already. If the new upgrade to iWork moves them close enough, then many workers and home users will find the $99 per year license for those who already have Office giving them the ability to access it on their iPads and iPhones to be very overpriced, and unnecessary.
We now read that Ballmer has just said that they are working on Office for the iPad. But that it's waiting for their touch version which will first come out for the Windows suit in mid 2014, with the ipad version not debuting until the end of the year. That at least a year from now. It may already be too late, but a year?
And then what will it cost? If it's tied to the desktop versions in license as I just mentioned, it's way too much. If it's much cheaper, Microsoft will have an income problem, and people will demand they lower the price for the desktop version as well. Uh oh! That's a major problem for Microsoft.
So after that long winded explanation, I hope you get a better idea of what's really going on.
A convertible is a laptop that turns into a tablet, those are just cases. Maybe Apple will make a keyboard for the iPad someday, but I doubt it. And if they ever make a convertible, they've lost their way.
You know, I've been giving this some thought myself. It's really an interesting question.
What would Apple need to do to come out with a notebook that used a more powerful version of the A7? Well, they would need all the software that most people would use a notebook for. That would be browsing, writing, homework, content watching and listening, games, finance, reading books, magazines, etc., photo editing, movie editing, painting, drawing, CAD, and other 3D work.
In other words, all the things that can be now done with iOS apps. Well, that's just amazing! What advantage would we get from this device though? That hard to say. Apple, so far, has been against the touch screen notebook format, so would that change? With Apple, no one will ever know unless it does. I would say; "The Shadow knows." But then, the Shadow is dead.
Could some version of OS X run on a more mature version of the A series chip? It's possible. Apple demonstrated, and Anandtech showed, that the A7 is equal to the mid Bay Trail Atom, and pretty close to the top chip there. Next year, if Apple can keep up their advances, it will push the mobile i3. And remember that it's still two core. If it were four core today, it would already be pushing the mobile i3.
Even if Apple must slow speed advances over the next few years, they will still be advancing much faster than intel can, and will have a chip, even with the small batteries iPad must use, that will be getting close to serious performance in the form of Intel's mobile i5 line.
But I see much more interesting things happening here. While we never know what's happening in Apple's labs, I get a feeling that Apple is doing more with the A line than just making them better iOS engines.
Apple, being unique in that they are the only ones today making both the OS and the chips, has an opportunity to do something special. What they can do is to look at the way x86 processes and the way ARM processes, and modify, or even change some of ARM's instructions to mimic x86. Those could be the areas in which an OS and its apps would have the most work in an emulation situation, and would therefor slow that emulation down the most. Remember that Apple has a lot of experience in this area.
If they could work out the worst performing areas, emulation might not be nearly as much of a chore. Giving the chip more power could allow Apple to raise the speed at which the chip is running. Right now, the A7 is running at the same 1.3Gz as the A6. But is twice as powerful. Many other top ARM designs are already running well beyond 2Gz, up to 2.7Gz. Imagine if Apple doubled that speed to 2.6Gz! Around twice the performance. In a notebook, that could be done because cooling would be far better. Of course, going by past experience, Apple will raise that to maybe 1.5Gz in the iPad, so we can talk about 3 Gz.
Now figure a four core version, with a newer design as well. I could easily see an A8x, if Apple were to do it, that could run at 3Gz in a notebook that would be at least ten times as powerful as today's A7, and more so in graphics, if they use the iPad's extra graphics abilities. That's already pushing the mobile i7 line.
It could be done. Whether Apple wants to do it is another affair, but the fact that they said the A7 had desktop performance shows that they may be thinking it. It could have been a tease.
Not sure you followed me. I'm suggesting that it would have been easier for Apple to increase clock speed and/or the number of cores on the A6X without the major upgrade to 64 bit. However, if this SOC is intended for uses on other devices with greater processing AND RAM requirements, then this would explain why Apple chose the more complex path.
I don't know much about PA Semi and Intrinsity, just remember reading some high level announcement when they were bought by Apple. Didn't realise they were that good!
Easier... nope.
In fact, the 64bit solution is amazingly elegant from a hardware and compiler level. and the dual 32/64 bit SoC made the upgrade 'less major.' The major issue is purely in the OS, and apple has solved that a few times already...
In short, this solved a problem now. I don't see this chip or the next turn of it being 'workstation worthy.' But a MacBook Air is probably sporting an A7 as we speak, and being shown to Intel as a 'We expect you to keep your price performance in line, as we do have options";-)
As noted elsewhere, the key benefit in 64 bit A7 is the handling of object creation and deletion events 50% faster (primarily taking advantage of 64bit register to register 1 clock MOVs... to do the same with a 32 bit would have required doubling clock speed, unlikely in today's power envelope). And having multiple cores on a non-compute intensive device (really, how many threads can my fingers have in RUN state at a time) wouldn't have sped stuff up much if this is the perceived bottleneck to the user (responsiveness 'feeling' as I drag or touch, all those events are being created, acted on and destroyed [I saw a finger touch[highlight], it moved[drag], it moved again[drag it some more], again[move it more], again[ditto... a couple hundred times a second], it stopped[what's it doing?], it lifted off[unhighlight] and the hundreds of thousands of graphics request messages that were fired off to display this simple action]
Make that 50% faster (handle 3 events in the CPU time it used to take to do 2]. That's like turning your clock from 2ghz to 3ghz(or more...).
The problem is, Qualcomm doesn't see the problem as 'make dragging seem less 'draggy'', just 'you need a faster processor' or you have a RAM bottle neck, or your GPU isn't polygonning fast enough.... Nope... it's just needing a wider data path.
The issue is less about how great the chip guys are, and more about 'old school' computer design. Back in the day, chip guys and compiler guys and OS guys were in same company, if not the same division or same building. Today, Apple and IBM are the only ones I know of that do it now (anyone build a better Zseries Mainframe lately?). Compilers, and systems developers often overlapped teams. in 'open systems' solutions, now you got to interface waste, as each interface is generalized, instead of optimized.
Apple can build a processor _for_themselves_, for One OS, One compiler, One set of frameworks. And provide a canonical profile of how iOS COULD operate if in a chip had these characteristics. This bi-directional (SW guys telling chip guys how to build, and chip guys telling compiler guys how to exploit) synergy of a organically integrated and optimized compute model (from the silicon up to the UI) is the Strong Sauce that Apple is exploiting.
What is old is new again.
You know, I've been giving this some thought myself. It's really an interesting question.
.....
I've posted before that the next Rosetta will be between OS X and iOS, thus allowing Apple to have the Ax chip in the MBA and then beyond which would increase margins as well as keep the competition at bay. You can easily and quickly copy software, you can't do the same with hardware. Apple could have a preprocessor core that quickly translates x86 to ARM.
Then it is only a matter of time when devs offer both the x86 and iOS versions until such time that x86 is no longer viable on the platform. And why not? If Apple can maintain momentum with the ARM, then they control one more piece of the puzzle.
[...]
Even if Apple must slow speed advances over the next few years, they will still be advancing much faster than intel can, and will have a chip, even with the small batteries iPad must use, that will be getting close to serious performance in the form of Intel's mobile i5 line.
But I see much more interesting things happening here. While we never know what's happening in Apple's labs, I get a feeling that Apple is doing more with the A line than just making them better iOS engines.
Apple, being unique in that they are the only ones today making both the OS and the chips, has an opportunity to do something special. What they can do is to look at the way x86 processes and the way ARM processes, and modify, or even change some of ARM's instructions to mimic x86. Those could be the areas in which an OS and its apps would have the most work in an emulation situation, and would therefor slow that emulation down the most. Remember that Apple has a lot of experience in this area.
If they could work out the worst performing areas, emulation might not be nearly as much of a chore. Giving the chip more power could allow Apple to raise the speed at which the chip is running. Right now, the A7 is running at the same 1.3Gz as the A6. But is twice as powerful. Many other top ARM designs are already running well beyond 2Gz, up to 2.7Gz. Imagine if Apple doubled that speed to 2.6Gz! Around twice the performance. In a notebook, that could be done because cooling would be far better. Of course, going by past experience, Apple will raise that to maybe 1.5Gz in the iPad, so we can talk about 3 Gz.
Now figure a four core version, with a newer design as well. I could easily see an A8x, if Apple were to do it, that could run at 3Gz in a notebook that would be at least ten times as powerful as today's A7, and more so in graphics, if they use the iPad's extra graphics abilities. That's already pushing the mobile i7 line.
It could be done. Whether Apple wants to do it is another affair, but the fact that they said the A7 had desktop performance shows that they may be thinking it. It could have been a tease.
No, it's not a tease... it's an over the bow shot at Intel.
Intel is driving power demands down as quickly as Apple is driving AxX speeds up. my guess is that Intel will maintain a slight lead with the x86 line for a while. Any posturing will only be for price and functionality concessions (ie, optimizing parts of the chip to better handle OSX unique and high use instruction pipelining).
It will be 4-5 years before Apple can evolve the A series into a significant desktop compute platform, but in 4-5 years will there be a desktop?
I think it's more impressive the 'other stuff' apple integrates into their platforms. When the Laptops start getting LTE integrated in, and M7 chips, or gyros or touchID, then you'll start looking at the 'family' of apple, not iOS and OSX. A corporation that commits to touchID as it's auth standard (or someone like the DoD), and apple is able to be the laptop and the iPad, and the phone, and the iPod, and the Thumb Drive manufacturer for them.... that's leverage of the apple difference... the CPU... that's just an implementation detail
If Apple chooses to do what I've outlined, then emulation could be fairly easy. Not easy, but fairly easy. In the old days, it was considered necessary to have the machine doing the emulation about 5 times as powerful as he one being emulated for the feel of speed to be the same. From my experiences with emulation, I would agree.
If all Apple did was to increase the lower of their SoC, it wouldn't be enough. ARM would need to be as powerful as the most powerful mobile i7 to emulate a top mobile i3, at best. Not acceptable!
If Apple can eliminate some of the worst bottlenecks, then it could be done. Considering how expensive Intel's solutions are, and how cheap Apple's are (that is, the cost of the chips for a complete system), then Apple may have a major cost advantage. We can look to why Surface Pro is so expensive, including the new, slightly less expensive model. It's because the chips are so expensive. If Microsoft could cut the cost of the chips to a quarter, the surface Pro could cost $495 instead of $895, and who knows what profit Microsoft is getting on that price?
So could we see an arm powered OS X notebook at some point? It's not impossible. There is a way. Whether there is the will is something else.
When I said tease, I meant to the developers sitting in the hall listening to the presentation, and to Apple users, and prospective Apple users watching the video of it later on.
I don't see it needing four or five years. And as I said, it's already performing at the top of the brand new Bay Trail Atom line.
Apple can do things to their OS and chips that neither Microsoft nor Intel can do, because Intel needs to have their chips perform well with a number of OS's, from the cheapest tablets to the most expensive supercomputers.
Apple can break whatever they want to, and have, if they find a major increase in performance warrants it. This can lead to much greater advances that Intel can manage with their need for backward compatibility.
Just to add to this, Apple could also retain the old instructions. Detect OS X, run this, detect iOS, run that.
I'll repeat this post in this newer thread:
From Daring Fireball:
Agam Shah, reporting for IDG:
“The comments made by Anand Chandrasekher, Qualcomm CMO, about 64-bit computing were inaccurate,” said a Qualcomm spokesperson in an email. “The mobile hardware and software ecosystem is already moving in the direction of 64-bit; and, the evolution to 64-bit brings desktop class capabilities and user experiences to mobile, as well as enabling mobile processors and software to run new classes of computing devices.”
Qualcomm did not provide further comment.
http://daringfireball.net/linked/2013/10/08/qualcomm
I'm happy to see Qualcomm again, is being honest about something. They're a pretty good company. They resisted both Motorola and Samsung's efforts to get them to improperly sue Apple over patents Qualcomm licensed from them.
Define "best."
Just wishing to have a Mac that can run A7 iOS tasks and OS X.
Since the iOS SoC is so cheap, they could add one to a Mac. I keep wondering if they ever will.
For you and all the other people vehemently trashing this idea - maybe you should start by realizing that not everyone has the same needs.
I can see a great demand for the product. When I'm traveling, an extra pound or two is a pain. An iPad does about 90% of what I need to do when I'm traveling, but I still end up carrying an iPad PLUS my MBP. With an iPad Pro (10-13" screen, keyboard, but otherwise an iPad), I'd only need one device and could leave the MBP at home. For that matter, I'd have an iMac on my desk rather than an MBP in the first place. An iMac and iPad Pro would be the ideal computing environment for me - and a lot of business executives I know.
Called it.