when the original ipad came out people looked at the A4 and saw that it was a Samsung SoC with some apple modifications. it has been well documented. the people at apple are fairly smart and know that you don't have to reinvent the wheel every single time to please the iCult with the NIH attitude
Yes, Apple is more pragmatic now and doesn't have NIH much. But you're misled. The Cortex-A8 CPU core in the Hummingbird is likely identical: based on an Intrinsity modified Cortex-A8 design to run at 1 GHz. Note, Inrinsity is not Samsung and was actually purchased by Apple sometime in Q1 2010. The two SoCs are not identical. The floorplan of the A4 and the Hummingbird are not the same with the Cortex-A8 cores not even in the same place. The GPUs are not the same. The A4 uses a PowerVR 535 while the Hummingbird uses a PowerVR 540.
Quote:
A4 doesn't have a GPU either. the RAM and GPU go into a single package kind of like the original intel i core CPU's
No. The GPU is on-die. The RAM is off-die but in the package. Both the A4 and A5 are 3 chip package-on-packages: the CPU-GPU-IE chip and 2 RAM chips.
Never said it was. DID say that it doesn't work in a traditional computing format.
I'd never buy one. That's not how anyone would want to use that OS. Apple even covered that.
Apple does sell a keyboard for use with iPad right? So what is different here? Nothing. Beyond that I think you miss many of the potential possibilities when it comes to new iOS devices. A iOS device with a keyboard does not need to look like a clam shell laptop.
It seems many here think Apple is retiring the MacBook line in favour of MacBook Air and MacBook Pro lines. I don't see how that makes sense.
For many the AIRs are a very good replacement for the Mac Book. The biggest issue is the truncated battery life.
Quote:
While I would expect them to offer x86_64-based MacBooks if they did reinvent the line with ARM chips running iOS with Aqua UI
That makes no sense. If you run the Aqua UI you are running Lion. IOS is an alternative GUI to the Mac OS GUI. Underneath they are running much of the same code. The iOS and Mac OS GUIs are just layers over the same basic OS.
Quote:
at a much lower price range that still yielded a healthy profit the one thing these new machines could be without is BootCamp access.
Such systems certainly could sell well. My point is that they will have to maintain X86 systems for sometime simply because there is a significant number of people that need that support.
This is ARM adding all their architecture references to LLVM/Clang. Clang 3.0 has all of these references included.
OpenCL 1.1 fully supported in Clang 3.0 is also part of the release and ARM has been working hard on making sure all their CPU designs are ready and supported.
I guess most of you don't work with LLVM/Clang and follow it's development.
This is not some big new thing.
Whomever Apple decides to stamp out it's designs I imagine will have some major minefields in legal action if it pulls a Samsung.
ARM-based MacBooks are absolutely inevitable. It's not a matter of "if." It's a matter of "when."
The advantages of a cool-running, energy-efficient ARM chip on, say, a MacBook Air far outweigh the disadvantages. Just a few of the many immediate advantages: longer battery life, no need for a cooling fan, smaller enclosures, lower cost.
The one disadvantage: no Windows (or possibly slow emulation.) Well so what. Apple could still sell legacy Intel-based desktops and laptops alongside its ARM-based MacBook Air. The vast majority of Mac users don't run windows on their Macs..
I don't think yoiu have any concept of why people buy computers.
The disadvantages of an ARM in a MBA are huge. It wouldn't have anywhere near the power to do even the basic stuff that people expect a laptop to do. Many people need Windows emulation, but even those who don't aren't going to be happy with 1/10 the performance.
Battery life? Not really very big a deal. You already get 7 hours out of an MBA. Getting 8 or 9 isn't that big of an advantage. Smaller enclosure? Hardly. The chip size isn't significantly different. You might save the fan's volume - again insignificant. Lower cost? Might reduce the cost of the system by $100 or so - again not sufficient to matter much.
People who can live with the power of ARM can get by with an iPad. People who want a laptop want more.
People are getting the whole ARM thing incorrect IMO.
ARM is in a fierce battle with Intel. Both are encroaching on each other's turf.
ARM rules the low power mobile device sector.
Intel rules the mid and high level X86 processors.
Intel is desirous of ARM's lucrative mobile business and have started their assault with Atom and some moving into more power efficient cores like Medfield.
ARM plans to go upstream and make processors suitable for Netbook and Tablets with more power. They have a A15 ARM core coming with quad core processing and more features that should be much more powerful than a quad core Cortex MP A9.
So the question really isn't "Will Apple use ARM in a MBA?" or attempts to link ARM to iOS devices. The question is how much power will larger ARM cores have in 2013 and to what effect could that have on a lightweight portable notebook?
Today's 11" Air has a 5hr battery life. What if an ARM based MBA 11" could last 15 hours? What if it was $699. I assure you that VM support or Boot Camp would not be that much of a deterrent to having such a device.
mdriftmeyer has already pointed towards the software availability with CLANG 3 and more. The tools exist to compile current Apple API into ARM.
It makes sense on many levels. The price point is there, the features are there we just need to know more about the performance of larger ARM cores.
Today's 11" Air has a 5hr battery life. What if an ARM based MBA 11" could last 15 hours?
Not even remotely possible. Even if the CPU used zero power, battery life wouldn't triple. Do you really believe that the CPU uses 2/3 of the entire energy budget of the machine? Not even close.
Quote:
Originally Posted by hmurchison
What if it was $699.
Not going to happen. The Sandy Bridge CPUs list for $200-300. Apple undoubtedly pays less.
The ARM CPUs would have to be free (at most) for Apple to be able to offer the MBA for $699. In all probability, the supplier would have to PAY Apple in order to get the price that low.
Quote:
Originally Posted by hmurchison
I assure you that VM support or Boot Camp would not be that much of a deterrent to having such a device.
Really? Who many Mac laptop owners do you know? Every single one I know uses some type of VM.
Quote:
Originally Posted by hmurchison
]mdriftmeyer has already pointed towards the software availability with CLANG 3 and more. The tools exist to compile current Apple API into ARM. .
The ABILITY to recompile was never the issue. The willingness to do so - and the performance degradation - is.
Alright I agree with that. I was reacting to the push that all of them are eventually going to be combined.
Not in the sense of UI but most definitely the underlying architecture will be highly unified in 5 years time. The differences between a Mac laptop and iPad will be like the differences between an iPad and iPhone.
With the trajectory of ARM one could even say only Mac desktops would use Intel in 2016. Even then ARM could be used in entry level iMacs by then.
The iPad form factor will definitely be around for a few years, it's just that awesome. For laptops, I imagine that it will become dual screen. That is, the "keyboard-trackpad" will be replaced by multitouch screen while the upper screen remains non-touch. So you have all the benefits of multitouch iPad-like wonder without sacrificing the benefits of a full laptop-like form factor. You gotta think for Apple, they want to do away with a Mac laptop and desktop having any kind of buttons (ie. keys) whatsoever. Two foldable glass panels is the ultimate evolution of a laptop as we currently know it. Just imagine. Two iPads with a simple hinge and Boom! There's your new MacBook Air. Tell me that isn't ultra sexy.
But these are just my predictions based on the path Steve may follow. With him not actively involved we'll just have to see what happens.
Not even remotely possible. Even if the CPU used zero power, battery life wouldn't triple. Do you really believe that the CPU uses 2/3 of the entire energy budget of the machine? Not even close
Not going to happen. The Sandy Bridge CPUs list for $200-300. Apple undoubtedly pays less.
The ARM CPUs would have to be free (at most) for Apple to be able to offer the MBA for $699. In all probability, the supplier would have to PAY Apple in order to get the price that low.
Really? Who many Mac laptop owners do you know? Every single one I know uses some type of VM.
The ABILITY to recompile was never the issue. The willingness to do so - and the performance degradation - is.
Sandy Bridge is a lot more expensive than a SoC. Now granted ARM simply isn't going to give Sandy Bridge performance in 2013 with a SoC but how many consumers are in need of a simple laptop for doing your basic word processing, internet and light media apps? I'm betting that in 2013 ARM will be fast enough to handle this just fine.
Biggest costs would be the display (11 inch), the SSD and the SoC. I'm not downplaying VM usage on a Mac because it's very powerful but I don't think every Mac notebook has to cater to the dual OS using crowd and I think there's a market for a sub $800 Mac notebook for people that iPads and touch interfaces just simply aren't going to work out as well.
I give it 60% odds at this time. There's too much code sharing now between iOS and OS X to rule it out completely. iOS now has Core Image, Core Data, AV Foundation and more matching OS X. The API are coming close enough that any developer using the newer API would be able to bring their applications to ARM based machines without too much trouble.
Irrelevant. You're claiming that switching from Sandy Bridge to ARM would extend MBA battery life from 5 hours to 15 hours. That is only possible if ARM uses effectively zero power AND if Sandy Bridge accounts for 67% of the total power usage of the computer. Do the math.
There's no way that the CPU accounts for 67% of the power usage, so it's impossible for you to go from 5 hours to 15 hours simply by switching the CPU.
Of course, if you're going to claim that you gain all those extra hours by using advanced battery technology, then switching the CPU is unnecessary, anyway.
Quote:
Originally Posted by hmurchison
Sandy Bridge is a lot more expensive than a SoC. Now granted ARM simply isn't going to give Sandy Bridge performance in 2013 with a SoC but how many consumers are in need of a simple laptop for doing your basic word processing, internet and light media apps? I'm betting that in 2013 ARM will be fast enough to handle this just fine.
So your magic handwaving is supposed to be a logical argument? ARM is currently something like 1/10 the performance of the CPUs that Apple is using. Do you really think that ARM is going to increase its performance 1000% in a year and a half - AND that Intel won't advance their chips?
Quote:
Originally Posted by hmurchison
Biggest costs would be the display (11 inch), the SSD and the SoC. I'm not downplaying VM usage on a Mac because it's very powerful but I don't think every Mac notebook has to cater to the dual OS using crowd and I think there's a market for a sub $800 Mac notebook for people that iPads and touch interfaces just simply aren't going to work out as well.
Sure there's a market for a sub-$800 Mac notebook. That doesn't mean that Apple can supply it without crippling the product. And so far, there's absolutely no evidence that Apple can do what you're suggesting (your vigorous hand-waving aside).
Quote:
Originally Posted by hmurchison
I give it 60% odds at this time. There's too much code sharing now between iOS and OS X to rule it out completely. iOS now has Core Image, Core Data, AV Foundation and more matching OS X. The API are coming close enough that any developer using the newer API would be able to bring their applications to ARM based machines without too much trouble.
So ARM is getting faster. Big deal - no one ever disputed it. You're ignoring the important facts, though:
1. Intel is getting faster, too. Ivy Bridge will provide significant gains - along the lines of the 40% you're claiming above.
2. ARM's performance is so far behind Intel's performance that for a general purpose laptop, ARM will take years to catch up - if ever
3. For most people, the software availability (including the ability to run Windows) is far more important than getting 6 or 7 hours out of the battery instead of 5. Very, very few people are going to run a laptop that won't run Office, for example.
I wouldn't rule out the possibility of Apple creating a new platform - sort of a super-iPad - which operates more like an iPad (targeted for media consumption), but I can't see any way that they're going to switch the MBA to ARM any time soon.
Did take you for so much of an Intel fan. No hand waving here ...I'm pleased as punch with my Core i5 MBA and i'm hoping to mate it with an Ivy Bridge 27" iMac next year.
However I do see a market for a basic computer. Something that can be functional for "run of the mill" tasks and light gaming.
Intel might have opportunities in this space with ULV processor and more but it'll be fun to see what ARM has for upscale chips from the A15 and on.
You're right the CPU doesn't account for 67% of the power but we cannot take the just the CPU figures alone. We've got to account for the chipset as well. Even the MBA have a thermal envelope of 25 watts.
I'm in no way saying ARM is going to be competitive with Intel but consider me impressed that my iPad 2 still retains the 10hr battery life in a lighter package with a dual core A5.
I really don't think the MBA line would go ARM. I'm more like thinking about a MBA 11" Plus. A notebook built to last as long as a trip across the Atlantic. It would only come in say two portable models an 11" and 13" and be the lowest cost option. The rest would remain Intel
However...to show i'm not merely hand waving I'll include David Kanter's excellent summation of the "Apple's moving to ARM" hype. It certainly throws water on the ideal happening soon but I've got no money on Intel or ARM in this race as of yet.
Long term, ARM could become a viable notebook option. Apple has invested a lot of time and effort in portability (e.g. LLVM, OpenCL, OpenGL), so it is clearly possible. Apple has few attachments to x86 and would not hesitate to shift to a better alternative. However, there are no ARM designs that will meet Apple?s needs for performance and efficiency in the next two or three years. Even if there was, it is hard to see how such a design would be substantially better than AMD's offerings. Over 5-10 years though, many of the technical and business hurdles may change. The ARM ecosystem is moving forward at a rapid pace, and Apple is watching carefully.
Apple is very likely exploring ARM based Mac Book Airs - for some simple reasons....
- MBA buyers are already used to the performance compromise and accept it - in return for better portability. Over the next 12-18 months, Quad core ARM SoCs will make it comparable to the Intel chips used in the MBA currently. Also remember, Apple does not use the cutting edge Intel chips in the MBA, it uses much slower chips than cutting edge.
- Performance wise, the iPad already does most tasks as fast as regular OSX based computers do. So, performance might not be a serious factor in this debate. Availability of apps, drivers, etc would be a much more significant issue. No one uses MBA for Video editing, or similar heavy duty applications. Mostly for web browsing, media playing, mail, etc.
- If nothing else, Apple will want to keep Intel on its toes. Intel has been operating against Apple's interest of late - for instance, they sided with Google in the Nortel Auctions - and they came up with the Ultrabook specification to increase competition for the MBA.
- While Intel also has a roadmap to lower power consumption, it does not look like Intel can achieve ARM levels of power consumption any time soon. There is far too much legacy software that is Windows based, because of which Intel cannot give up its X86 architecture any time soon - and it is tough for Intel to compete with ARM in terms of power efficiency, as long as it retains the X86 architecture. Of course, if MS makes a clean break with Metro, and stops supporting Legacy software, this problem will be reduced for Intel.
- With the launch of Lion, Apple has clearly shown that they are looking to converge iOS and Mac OS X. In the future, we are likely to see more similarities between these products. In any case, it is only the UI portions of iOS that differ from Mac OS X. All other Core components are already present in both.
- The likely scenario is that Apple retains Intel based MBA for the users that need it, and move to an ARM based solution for a new product line. This covers both bases for Apple.
- In any case, whether Apple does this or not, I think at some point we will start seeing Linux based solutions that use ARM to lower cost, weight (because of lower battery needs). So Apple might as well develop such a product.
An ARM based MBA will certainly be able to run windows. Since W8 is arm compatible it's inevitable. Furthermore, Apple is not going to let MSFT be the only major OS on ARM. Future ARM MBAs (though possibly not the first) will run os x and W8 just fine.
- If nothing else, Apple will want to keep Intel on its toes. Intel has been operating against Apple's interest of late - for instance, they sided with Google in the Nortel Auctions - and they came up with the Ultrabook specification to increase competition for the MBA.
- While Intel also has a roadmap to lower power consumption, it does not look like Intel can achieve ARM levels of power consumption any time soon. There is far too much legacy software that is Windows based, because of which Intel cannot give up its X86 architecture any time soon - and it is tough for Intel to compete with ARM in terms of power efficiency, as long as it retains the X86 architecture. Of course, if MS makes a clean break with Metro, and stops supporting Legacy software, this problem will be reduced for Intel.
You're missing a couple things here. Intel expanding on the ultrabook specification doesn't really work against Apple at all. Pushing for a larger market for the processor class used there could really help advance things for the macbook air and potentially bring down manufacturing costs. That could turn into a win for Apple if Intel devotes more resources to that processor class.
Regarding quad ARM processors vs Intel and X86, what do predictions look like in terms of ARM based cpus matching Intel /x86 in terms of performance while retaining a significant advantage on power efficiency? Everyone else has suggested that this won't bring them up to similar performance levels. I have yet to see any sources referenced for these raw performance speculations from anyone.
Edit: Also consider the nightmare of emulating x86 software on slower hardware. The intel core duos fell behind the powerbooks on some applications when run through rosetta (I can try to find the info on it, but it's quite old at this point). The point being that could be a rough transition with existing software.
Comments
when the original ipad came out people looked at the A4 and saw that it was a Samsung SoC with some apple modifications. it has been well documented. the people at apple are fairly smart and know that you don't have to reinvent the wheel every single time to please the iCult with the NIH attitude
Yes, Apple is more pragmatic now and doesn't have NIH much. But you're misled. The Cortex-A8 CPU core in the Hummingbird is likely identical: based on an Intrinsity modified Cortex-A8 design to run at 1 GHz. Note, Inrinsity is not Samsung and was actually purchased by Apple sometime in Q1 2010. The two SoCs are not identical. The floorplan of the A4 and the Hummingbird are not the same with the Cortex-A8 cores not even in the same place. The GPUs are not the same. The A4 uses a PowerVR 535 while the Hummingbird uses a PowerVR 540.
A4 doesn't have a GPU either. the RAM and GPU go into a single package kind of like the original intel i core CPU's
No. The GPU is on-die. The RAM is off-die but in the package. Both the A4 and A5 are 3 chip package-on-packages: the CPU-GPU-IE chip and 2 RAM chips.
Never said it was. DID say that it doesn't work in a traditional computing format.
I'd never buy one. That's not how anyone would want to use that OS. Apple even covered that.
Apple does sell a keyboard for use with iPad right? So what is different here? Nothing. Beyond that I think you miss many of the potential possibilities when it comes to new iOS devices. A iOS device with a keyboard does not need to look like a clam shell laptop.
Apples MAC's will not move to ARM.
*Mac. All caps would means it's an acronym, it isn't. Mac is an abbreviation of Macintosh.
It seems many here think Apple is retiring the MacBook line in favour of MacBook Air and MacBook Pro lines. I don't see how that makes sense.
For many the AIRs are a very good replacement for the Mac Book. The biggest issue is the truncated battery life.
While I would expect them to offer x86_64-based MacBooks if they did reinvent the line with ARM chips running iOS with Aqua UI
That makes no sense. If you run the Aqua UI you are running Lion. IOS is an alternative GUI to the Mac OS GUI. Underneath they are running much of the same code. The iOS and Mac OS GUIs are just layers over the same basic OS.
at a much lower price range that still yielded a healthy profit the one thing these new machines could be without is BootCamp access.
Such systems certainly could sell well. My point is that they will have to maintain X86 systems for sometime simply because there is a significant number of people that need that support.
iOS also doesn't mean touch screen, it simply means ARM based, not arm based.
IOS runs fine on i86.
OpenCL 1.1 fully supported in Clang 3.0 is also part of the release and ARM has been working hard on making sure all their CPU designs are ready and supported.
I guess most of you don't work with LLVM/Clang and follow it's development.
This is not some big new thing.
Whomever Apple decides to stamp out it's designs I imagine will have some major minefields in legal action if it pulls a Samsung.
ARM-based MacBooks are absolutely inevitable. It's not a matter of "if." It's a matter of "when."
The advantages of a cool-running, energy-efficient ARM chip on, say, a MacBook Air far outweigh the disadvantages. Just a few of the many immediate advantages: longer battery life, no need for a cooling fan, smaller enclosures, lower cost.
The one disadvantage: no Windows (or possibly slow emulation.) Well so what. Apple could still sell legacy Intel-based desktops and laptops alongside its ARM-based MacBook Air. The vast majority of Mac users don't run windows on their Macs..
I don't think yoiu have any concept of why people buy computers.
The disadvantages of an ARM in a MBA are huge. It wouldn't have anywhere near the power to do even the basic stuff that people expect a laptop to do. Many people need Windows emulation, but even those who don't aren't going to be happy with 1/10 the performance.
Battery life? Not really very big a deal. You already get 7 hours out of an MBA. Getting 8 or 9 isn't that big of an advantage. Smaller enclosure? Hardly. The chip size isn't significantly different. You might save the fan's volume - again insignificant. Lower cost? Might reduce the cost of the system by $100 or so - again not sufficient to matter much.
People who can live with the power of ARM can get by with an iPad. People who want a laptop want more.
ARM is in a fierce battle with Intel. Both are encroaching on each other's turf.
ARM rules the low power mobile device sector.
Intel rules the mid and high level X86 processors.
Intel is desirous of ARM's lucrative mobile business and have started their assault with Atom and some moving into more power efficient cores like Medfield.
ARM plans to go upstream and make processors suitable for Netbook and Tablets with more power. They have a A15 ARM core coming with quad core processing and more features that should be much more powerful than a quad core Cortex MP A9.
So the question really isn't "Will Apple use ARM in a MBA?" or attempts to link ARM to iOS devices. The question is how much power will larger ARM cores have in 2013 and to what effect could that have on a lightweight portable notebook?
Today's 11" Air has a 5hr battery life. What if an ARM based MBA 11" could last 15 hours? What if it was $699. I assure you that VM support or Boot Camp would not be that much of a deterrent to having such a device.
mdriftmeyer has already pointed towards the software availability with CLANG 3 and more. The tools exist to compile current Apple API into ARM.
It makes sense on many levels. The price point is there, the features are there we just need to know more about the performance of larger ARM cores.
Today's 11" Air has a 5hr battery life. What if an ARM based MBA 11" could last 15 hours?
Not even remotely possible. Even if the CPU used zero power, battery life wouldn't triple. Do you really believe that the CPU uses 2/3 of the entire energy budget of the machine? Not even close.
What if it was $699.
Not going to happen. The Sandy Bridge CPUs list for $200-300. Apple undoubtedly pays less.
The ARM CPUs would have to be free (at most) for Apple to be able to offer the MBA for $699. In all probability, the supplier would have to PAY Apple in order to get the price that low.
I assure you that VM support or Boot Camp would not be that much of a deterrent to having such a device.
Really? Who many Mac laptop owners do you know? Every single one I know uses some type of VM.
]mdriftmeyer has already pointed towards the software availability with CLANG 3 and more. The tools exist to compile current Apple API into ARM. .
The ABILITY to recompile was never the issue. The willingness to do so - and the performance degradation - is.
Alright I agree with that. I was reacting to the push that all of them are eventually going to be combined.
Not in the sense of UI but most definitely the underlying architecture will be highly unified in 5 years time. The differences between a Mac laptop and iPad will be like the differences between an iPad and iPhone.
With the trajectory of ARM one could even say only Mac desktops would use Intel in 2016. Even then ARM could be used in entry level iMacs by then.
The iPad form factor will definitely be around for a few years, it's just that awesome. For laptops, I imagine that it will become dual screen. That is, the "keyboard-trackpad" will be replaced by multitouch screen while the upper screen remains non-touch. So you have all the benefits of multitouch iPad-like wonder without sacrificing the benefits of a full laptop-like form factor. You gotta think for Apple, they want to do away with a Mac laptop and desktop having any kind of buttons (ie. keys) whatsoever. Two foldable glass panels is the ultimate evolution of a laptop as we currently know it. Just imagine. Two iPads with a simple hinge and Boom! There's your new MacBook Air. Tell me that isn't ultra sexy.
But these are just my predictions based on the path Steve may follow. With him not actively involved we'll just have to see what happens.
Not even remotely possible. Even if the CPU used zero power, battery life wouldn't triple. Do you really believe that the CPU uses 2/3 of the entire energy budget of the machine? Not even close
Not going to happen. The Sandy Bridge CPUs list for $200-300. Apple undoubtedly pays less.
The ARM CPUs would have to be free (at most) for Apple to be able to offer the MBA for $699. In all probability, the supplier would have to PAY Apple in order to get the price that low.
Really? Who many Mac laptop owners do you know? Every single one I know uses some type of VM.
The ABILITY to recompile was never the issue. The willingness to do so - and the performance degradation - is.
Of course it's possible given the power efficiency of ARM SoC and improvements in battery technology (Improved Battery Tech - Patently Apple ) and more
Sandy Bridge is a lot more expensive than a SoC. Now granted ARM simply isn't going to give Sandy Bridge performance in 2013 with a SoC but how many consumers are in need of a simple laptop for doing your basic word processing, internet and light media apps? I'm betting that in 2013 ARM will be fast enough to handle this just fine.
Biggest costs would be the display (11 inch), the SSD and the SoC. I'm not downplaying VM usage on a Mac because it's very powerful but I don't think every Mac notebook has to cater to the dual OS using crowd and I think there's a market for a sub $800 Mac notebook for people that iPads and touch interfaces just simply aren't going to work out as well.
I give it 60% odds at this time. There's too much code sharing now between iOS and OS X to rule it out completely. iOS now has Core Image, Core Data, AV Foundation and more matching OS X. The API are coming close enough that any developer using the newer API would be able to bring their applications to ARM based machines without too much trouble.
ARM Cortex A15
40% faster than A9
plus faster PowerVR graphics
PowerVR Rogue
We'll see what happens but the option is there.
Of course it's possible given the power efficiency of ARM SoC and improvements in battery technology (Improved Battery Tech - Patently Apple ) and more
Irrelevant. You're claiming that switching from Sandy Bridge to ARM would extend MBA battery life from 5 hours to 15 hours. That is only possible if ARM uses effectively zero power AND if Sandy Bridge accounts for 67% of the total power usage of the computer. Do the math.
There's no way that the CPU accounts for 67% of the power usage, so it's impossible for you to go from 5 hours to 15 hours simply by switching the CPU.
Of course, if you're going to claim that you gain all those extra hours by using advanced battery technology, then switching the CPU is unnecessary, anyway.
Sandy Bridge is a lot more expensive than a SoC. Now granted ARM simply isn't going to give Sandy Bridge performance in 2013 with a SoC but how many consumers are in need of a simple laptop for doing your basic word processing, internet and light media apps? I'm betting that in 2013 ARM will be fast enough to handle this just fine.
So your magic handwaving is supposed to be a logical argument? ARM is currently something like 1/10 the performance of the CPUs that Apple is using. Do you really think that ARM is going to increase its performance 1000% in a year and a half - AND that Intel won't advance their chips?
Biggest costs would be the display (11 inch), the SSD and the SoC. I'm not downplaying VM usage on a Mac because it's very powerful but I don't think every Mac notebook has to cater to the dual OS using crowd and I think there's a market for a sub $800 Mac notebook for people that iPads and touch interfaces just simply aren't going to work out as well.
Sure there's a market for a sub-$800 Mac notebook. That doesn't mean that Apple can supply it without crippling the product. And so far, there's absolutely no evidence that Apple can do what you're suggesting (your vigorous hand-waving aside).
I give it 60% odds at this time. There's too much code sharing now between iOS and OS X to rule it out completely. iOS now has Core Image, Core Data, AV Foundation and more matching OS X. The API are coming close enough that any developer using the newer API would be able to bring their applications to ARM based machines without too much trouble.
ARM Cortex A15
40% faster than A9
plus faster PowerVR graphics
PowerVR Rogue
We'll see what happens but the option is there.
So ARM is getting faster. Big deal - no one ever disputed it. You're ignoring the important facts, though:
1. Intel is getting faster, too. Ivy Bridge will provide significant gains - along the lines of the 40% you're claiming above.
2. ARM's performance is so far behind Intel's performance that for a general purpose laptop, ARM will take years to catch up - if ever
3. For most people, the software availability (including the ability to run Windows) is far more important than getting 6 or 7 hours out of the battery instead of 5. Very, very few people are going to run a laptop that won't run Office, for example.
I wouldn't rule out the possibility of Apple creating a new platform - sort of a super-iPad - which operates more like an iPad (targeted for media consumption), but I can't see any way that they're going to switch the MBA to ARM any time soon.
Did take you for so much of an Intel fan. No hand waving here ...I'm pleased as punch with my Core i5 MBA and i'm hoping to mate it with an Ivy Bridge 27" iMac next year.
However I do see a market for a basic computer. Something that can be functional for "run of the mill" tasks and light gaming.
Intel might have opportunities in this space with ULV processor and more but it'll be fun to see what ARM has for upscale chips from the A15 and on.
You're right the CPU doesn't account for 67% of the power but we cannot take the just the CPU figures alone. We've got to account for the chipset as well. Even the MBA have a thermal envelope of 25 watts.
I'm in no way saying ARM is going to be competitive with Intel but consider me impressed that my iPad 2 still retains the 10hr battery life in a lighter package with a dual core A5.
I really don't think the MBA line would go ARM. I'm more like thinking about a MBA 11" Plus. A notebook built to last as long as a trip across the Atlantic. It would only come in say two portable models an 11" and 13" and be the lowest cost option. The rest would remain Intel
However...to show i'm not merely hand waving I'll include David Kanter's excellent summation of the "Apple's moving to ARM" hype. It certainly throws water on the ideal happening soon but I've got no money on Intel or ARM in this race as of yet.
Why Apple won't ARM the Macbook
Long term, ARM could become a viable notebook option. Apple has invested a lot of time and effort in portability (e.g. LLVM, OpenCL, OpenGL), so it is clearly possible. Apple has few attachments to x86 and would not hesitate to shift to a better alternative. However, there are no ARM designs that will meet Apple?s needs for performance and efficiency in the next two or three years. Even if there was, it is hard to see how such a design would be substantially better than AMD's offerings. Over 5-10 years though, many of the technical and business hurdles may change. The ARM ecosystem is moving forward at a rapid pace, and Apple is watching carefully.
- MBA buyers are already used to the performance compromise and accept it - in return for better portability. Over the next 12-18 months, Quad core ARM SoCs will make it comparable to the Intel chips used in the MBA currently. Also remember, Apple does not use the cutting edge Intel chips in the MBA, it uses much slower chips than cutting edge.
- Performance wise, the iPad already does most tasks as fast as regular OSX based computers do. So, performance might not be a serious factor in this debate. Availability of apps, drivers, etc would be a much more significant issue. No one uses MBA for Video editing, or similar heavy duty applications. Mostly for web browsing, media playing, mail, etc.
- If nothing else, Apple will want to keep Intel on its toes. Intel has been operating against Apple's interest of late - for instance, they sided with Google in the Nortel Auctions - and they came up with the Ultrabook specification to increase competition for the MBA.
- While Intel also has a roadmap to lower power consumption, it does not look like Intel can achieve ARM levels of power consumption any time soon. There is far too much legacy software that is Windows based, because of which Intel cannot give up its X86 architecture any time soon - and it is tough for Intel to compete with ARM in terms of power efficiency, as long as it retains the X86 architecture. Of course, if MS makes a clean break with Metro, and stops supporting Legacy software, this problem will be reduced for Intel.
- With the launch of Lion, Apple has clearly shown that they are looking to converge iOS and Mac OS X. In the future, we are likely to see more similarities between these products. In any case, it is only the UI portions of iOS that differ from Mac OS X. All other Core components are already present in both.
- The likely scenario is that Apple retains Intel based MBA for the users that need it, and move to an ARM based solution for a new product line. This covers both bases for Apple.
- In any case, whether Apple does this or not, I think at some point we will start seeing Linux based solutions that use ARM to lower cost, weight (because of lower battery needs). So Apple might as well develop such a product.
- If nothing else, Apple will want to keep Intel on its toes. Intel has been operating against Apple's interest of late - for instance, they sided with Google in the Nortel Auctions - and they came up with the Ultrabook specification to increase competition for the MBA.
- While Intel also has a roadmap to lower power consumption, it does not look like Intel can achieve ARM levels of power consumption any time soon. There is far too much legacy software that is Windows based, because of which Intel cannot give up its X86 architecture any time soon - and it is tough for Intel to compete with ARM in terms of power efficiency, as long as it retains the X86 architecture. Of course, if MS makes a clean break with Metro, and stops supporting Legacy software, this problem will be reduced for Intel.
You're missing a couple things here. Intel expanding on the ultrabook specification doesn't really work against Apple at all. Pushing for a larger market for the processor class used there could really help advance things for the macbook air and potentially bring down manufacturing costs. That could turn into a win for Apple if Intel devotes more resources to that processor class.
Regarding quad ARM processors vs Intel and X86, what do predictions look like in terms of ARM based cpus matching Intel /x86 in terms of performance while retaining a significant advantage on power efficiency? Everyone else has suggested that this won't bring them up to similar performance levels. I have yet to see any sources referenced for these raw performance speculations from anyone.
Edit: Also consider the nightmare of emulating x86 software on slower hardware. The intel core duos fell behind the powerbooks on some applications when run through rosetta (I can try to find the info on it, but it's quite old at this point). The point being that could be a rough transition with existing software.