cause you can have your research paper and your source material open side by side and be much more effective than having to switch back and forth.
Agreed, it's definitely not a one size fits all situation.
One of the interesting things shown in the Windows 8 Tablet (Currently x86) demo was the ability to split the display and run a separate app in each portion. I suspect that this could be a separate document of the same app.
I forget which, but some iPad app uses a custom on-screen keyboard when entering text -- the kb has arrow keys to granularly position the cursor within the text, and extending a text selection.
The Apple OS is kind of a circle of life:
Mac OS X ---> iOS ---> Mac OS X.
1) Many things that existed in Mac OS X were reimplemented in iOS (the way they should have been done in the first place) -- then ported back to Mac OS X.
2) Some things (like Location Services) were developed for iOS and have been ported to Mac OS X.
3) Some very powerful Mac OS X APIs have already been ported to iOS, even though there is no current support or usage for them.
4) With things like AirPlay, interconnection/interaction of iDevices and Macs is just beginning to exploit the interconnection and the advantages of each -- think iDevice game controllers, graphics tablets, Personal TVs, A/V editing control surfaces... interfacing other iDevices, TVs or Macs.
5) One major difference between iOS and Mac OS X is the UI -- there are, currently, two separate interfaces. However, that too, is starting to change: The iPhone uses a single-window-at-a-time to display and drill-down information. The additional screen real estate of the iPad provides side-by-side (or popup overlay) display of multiple columns (currently 2 columns: summary and detail).
I suspect that in future iOS releases we will see columnar tables (or spreadsheets) similar to the way Mac OS X displays tabular data, say in iTunes -- multiple columns that are resizable, movable, sortable...
This is an existing API in Mac OS X that can be readily ported to iOS -- when it makes sense.
6) There is a convergence in the Mac OS X and iOS development tools -- at some point, in the not too distant future we will have 'universal" apps that configure themselves upon installation (or execution) to optimize themselves (function, UI, etc.) to the device environment and the user.
I suspect that within 4 years we won't know or care which OS is running on which device...
We'll just satisfy our needs of the moment by choosing S-M-L-XL-XXX.
Firstly, they will never ship ARM based CPUs until they are at least 64 bit processors. By then the 128 bit market will be on the verge of hitting the server space.
Secondly, they will not remotely compete with the APU products coming out of AMD.
The Macbook Air will have to have a cluster of 64 bit ARM processors with up to 4 Cores per SoC to match the power of the current i5/i7 lineup.
The next major revision of CPUs from Intel and AMD will be battling it out with 8 and 10 core CPUs and their power dwarfs that of the embedded space.
The raw GFLOPS of the latest CPUs are surpassing 200 GFLOPS in their next revision. The Tegra from Nvidia's latest top GFLOP mark is 6.4. Sorry, but the power isn't remotely close.
Finally, the sheer volume of wafer production in the embedded space is already at it's limitations forcing 3rd parties to invest tens of billions just to keep up.
Apple isn't going to invest tens of billions for 3rd parties just to put 40 bit instruction set aware A15's into a Macbook Air.
OS X is a 64 bit through and through OS. The goal for Apple is to move iOS to a 64 bit code base. The ARM A8 is the first 64 bit OS ready design.
Your (flawed) assumption is that Macbook Air users really need the performance of an i5/i7 CPU.
If the iPad has shown us anything performance isn't everything.
90% of those buying computers rarely do anything beyond checking email, surfing the web, stream video, and word process/productivity.
x86 is like performing surgery with a chain saw.
First, bad analogy. x86 can perform the entire range of "cutting" from surgery to cutting wood. Is it overkill for many tasks, sure. But it's perfectly capably of performing them all, unlike your "surgery wtih a chain saw" analogy.
Second, even something as seemingly benign as iPhoto can take up a lot of processing power for some of it's functions, such as facial recognition. The point being that just because you don't always need the power of a "full-sized" CPU, it doesn't mean you never need the it. You wouldn't buy a car that can't go over 35 mph just because you mostly drive in the city, would you?
First, bad analogy. x86 can perform the entire range of "cutting" from surgery to cutting wood. Is it overkill for many tasks, sure. But it's perfectly capably of performing them all, unlike your "surgery wtih a chain saw" analogy.
Second, even something as seemingly benign as iPhoto can take up a lot of processing power for some of it's functions, such as facial recognition. The point being that just because you don't always need the power of a "full-sized" CPU, it doesn't mean you never need the it. You wouldn't buy a car that can't go over 35 mph just because you mostly drive in the city, would you?
What you say is true!
As far as the Apple implementation of the ARM / GPU processing capabilities...
As I understand it, iOS can turn extra cores On or Off as needed and use them efficiently to maximize performance and minimize power.
I don't know that this level of granular control exists in x86 chips.
It would be neat to see Apple release an inexpensive hockey puck compute box similar in looks to the AppleTV 2.
This box would contain power supply, ARM SoCs, RAM, minimal SSD and 2 ThunderBolt ports.
You could just slap as many of these as needed on a ThunderBolt daisy-chain along with Computers, RAIDs, Displays.
The main Computer (maybe a Mini) would dispatch and monitor high power/performance compute tasks outboard to the hockey pucks.
Apple already has existing software that does this.
I see a future for an unified architecture in the computing space.
The situation for many(most) people at the moment is that we have a full-size computer on x86 and some sort of mobile computer on ARM, this is only becoming more prevelent, especially as our mobile computers(tablets, iPhones) become more powerful and common
But the software I have on one doesn't run on the other, i can't run msoffice on my iPad etc...
I think apple is moving this direction with features from ios making it to OSX, and with the testing of macs on ARM
The problem is that with ARM, it's just not a performance issue, but a compatability issue; none of the big Mac software will run without a recompile
What this new intel chip is; is a nearly competitive mobile processor that Could, theoretically run OSX and the existing Mac apps without needing to recompile, one that will get more competitive with the next generations
Imagine a iPad that could be your only PC, pair it with a dock and it's a desktop and the os switches to keyboard/mouse mode with multiple windows remove it from the dock and all your apps stay open and switch to touchscreen mode, full screen apps and gesture control
Apple already has the same os on both, just different interfaces (and as of now, different binaries for ARM)
Now, apple could go either way, switch OSX to ARM or switch IOS to x86
I think the latter is easier.
1. Most of the BIG expensive apps like photoshop, etc are x86 right now
2. Having users buy a new version to use on their new MacBook is cost prohibitive
3. Most iOS apps are free/less expensive
4. This intel chip, as well as the next generations can run arm binaries( through built in emulation)
A. This would mean most of the low-performance iOS apps, like utilities, basic games, etc... could run as-is on this(or the next generation) chip
5. the current and next generation arm chips are LESS powerful than intels desktop chips(though they are catching up)
A. You generally cannot emulate a higher performance system on a lower performance chip
B. as stated before, many of the popular apps on the Mac are bigger, higher performance apps, unsuitable for emulation
These points, assuming that In the next couple of generations intel can match Arm performance in the mobile arena, tell me that an iOS switch to x86 is more likley than a Mac switch to ARM
The proof will be when we are able to compare contemporaneously shipping products from both companies. Right now there are no shipping products with Intel's new chip and Intel is making performance claims relative to ARM chips that are in currently shipping products. If we compared ARMs future products to Intel's currently shipping products ARM would look a lot better.
Yeah in some cases they compare to year and a half old devices that are supposedly low power. Intel appears to be running about twice the power levels of these old devices.
Quote:
We've seen this dance play out between Intel and AMD for years, and it's always the same -- you don't really know until you've got shipping versions of both products in front of you.
Obviously the GUI is different as is process switching but the core bits are the same. In effect iOS gives you a UNIX like OS in your pocket.
Quote:
Originally Posted by cvaldes1831
Yes, that would be a more likely scenario.
The original release of Mac OS X was delayed substantially until its release in early 2001. There have been persistent rumors (never confirmed) that much of the delay was caused by Apple insisting that the code run on x86 architecture on secret Intel-based prototypes. When Apple announced their switch from PowerPC to x86 architecture in the middle of the last decade, they were able to make a relatively smooth transition to the Intel processors.
Mac OS X 10.6 Snow Leopard was described by Apple as a total under-the-hood rewrite of the operating system. It is possible that ARM support was included in a parallel source tree and that each major revision of the code much run on secret ARM-based Mac prototypes.
Of course, we will never know for sure based on Apple's penchant for secrecy, but if Apple ever does release a Mac with an ARM processor, it will likely run OS X. At that point, it is likely that Apple will give developers to generate Universal binaries (Intel and ARM code combined) or to generate thin binaries that only support one architecture. With a few tweaks to the Mac App Store, it would probably be rather simple to deliver the correct thin binary to the target machine.
Not so easily, I don't think. Apple was able to make the jump from Motorola 68K to PowerPC because the PowerPC was powerful enough to emulate the 68K instruction set and still have acceptable performance. Same holds for the PowerPC to Intel jump. The Intel chips had the extra horsepower to run PowerPC code in emmulation. During both transitions most applications and even large chunks for the operating system itself were running in emmulation for months and years after the transition.
To allow a similar transition, the ARM chip would need to be sufficiently more powerful than the Intel chip it is replacing to run code for that Intel chip in emmulation at acceptable speeds. The only alternative would be to get the entire OS and a good representation of the application library 100% ARM native prior to the transition. I would not put that in the category of "easily" dumping Intel.
It may happen someday. But it won't be anytime soon and it won't be easy.
One thing Apple could do is abandon the market for high powered computers. If ARM chips could eventually produce the computing power of an i5 mid level chip, forget about the few i7's PC's sold and focus on what most people buy.
Performance without power considerations is meaningless, especially in the smartphone world. Luckily for Intel, Medfield seems very competitive there as well. Intel provided some power and performance data for Medfield based on its reference platform. I still haven't been able to verify any of this for myself, but I was able to see some power tests run in person on the reference platform and competitive devices.
The Intel provided values are pretty astonishing . Sub 20mW idle, sub 750mW during a call on 3G and although not pictured here, Intel's internal data suggests ~1W power consumption while browsing the web compared to ~1.3W on the iPhone 4S and Galaxy S 2. I've done my own measurements on 4S web browsing and came up with a very similar value.
Intel Measured Smartphone Power Consumption (Identical Display Brightness)
So from the looks of it, it looks like Intel has a significant performance lead with comparable power consumption. Now keep in mind that this is Intel-provided data, so take that however you will.
Also remember that phones using Medfield will not appear until at least Q2 2012; by then ARM A15 should be out, which should decrease the performance gap and may actually have better power consumption with big.LITTLE.
Finally remember that Medfield is built on 32nm while ARM designs are pretty much all 40/45nm right now. We're expected to see 32/28nm ARM designs this year so it will be interesting to see how much Intel is relying on its process node advantages when that happens.
While I'm still somewhat skeptical at this point, with the data released thus far, and with Intel finally throwing its weight behind Android and partnering up with Motorola and Lenovo, they definitely have potential here.
I am not sure I understand this hard-on for ARM in laptops. It's not like ARM is the good guys and intel is the bad guys. if ARM performance gets there, sure, but that is a long way off, and they would always be chasing the puck rather than getting to where intel will be sending it.
I could see how Medfield fits in though, if you were chasing ultra long battery life in a laptop. Medfield would still be a performance hit, but as it is x86 compatible it would not be as much of a dog as an ARM chip requiring emulation.
So the chip range could have:
ARM for IOS and Android devices;
Medfield for long life, low power home servers, desktops and laptops running OSX or Win8, and some Win8* or Android tablets*;
i7 etc for higher performance products running OSX or Win8.
* these will not be successful, as in the case of very low power mobile devices, ARM, not intel would be the architecture determining where the performance envelope puck is going to be for a long time to come.
Intel has the best fabs in the planet. Can produce chips cheaper than anyone else.
Probably, but ARM don't produce chips. It's hard to poach business from a company not in you business. How do you undercut zero?
Intel will have to try a different tactic, like making a better product. Better performance, power consumption, graphics, anything... They can't compete on price.
Do you have any data to support this contention? We know that the Wintel OEMs are seeing their sales lag because customers are buying tablets rather than laptops and desktops.
This can also mean they are delaying their purchase of a replacement computer because they purchased a tablet. Not that they have chosen a tablet and not having a computer.
I'm still waiting for you to name some real applications. All I see there are two terrible programs that have much better alternatives already in tablet form.
It was said that Apple officials were impressed by the results, as the test machine performed "better than expected."
Every time I read a rumor about this, I feel like people are distorting what was likely originally conveyed. As has been said many a time, the ARM is an excellent chip, but in raw power it pales beside the Core2 chip inside the Air right now. A thing can perform quite poorly but still be "better than expected." So were Apple officials actually described as being impressed, or just not as depressed as they'd expected to be? Seriously. Someone should go back to their original source (assuming such a source was real and not a peyote dream) and ask that, specifically.
Comments
Originally Posted by sleepy3
...
cause you can have your research paper and your source material open side by side and be much more effective than having to switch back and forth.
Agreed, it's definitely not a one size fits all situation.
One of the interesting things shown in the Windows 8 Tablet (Currently x86) demo was the ability to split the display and run a separate app in each portion. I suspect that this could be a separate document of the same app.
I forget which, but some iPad app uses a custom on-screen keyboard when entering text -- the kb has arrow keys to granularly position the cursor within the text, and extending a text selection.
The Apple OS is kind of a circle of life:
Mac OS X ---> iOS ---> Mac OS X.
1) Many things that existed in Mac OS X were reimplemented in iOS (the way they should have been done in the first place) -- then ported back to Mac OS X.
2) Some things (like Location Services) were developed for iOS and have been ported to Mac OS X.
3) Some very powerful Mac OS X APIs have already been ported to iOS, even though there is no current support or usage for them.
4) With things like AirPlay, interconnection/interaction of iDevices and Macs is just beginning to exploit the interconnection and the advantages of each -- think iDevice game controllers, graphics tablets, Personal TVs, A/V editing control surfaces... interfacing other iDevices, TVs or Macs.
5) One major difference between iOS and Mac OS X is the UI -- there are, currently, two separate interfaces. However, that too, is starting to change: The iPhone uses a single-window-at-a-time to display and drill-down information. The additional screen real estate of the iPad provides side-by-side (or popup overlay) display of multiple columns (currently 2 columns: summary and detail).
I suspect that in future iOS releases we will see columnar tables (or spreadsheets) similar to the way Mac OS X displays tabular data, say in iTunes -- multiple columns that are resizable, movable, sortable...
This is an existing API in Mac OS X that can be readily ported to iOS -- when it makes sense.
6) There is a convergence in the Mac OS X and iOS development tools -- at some point, in the not too distant future we will have 'universal" apps that configure themselves upon installation (or execution) to optimize themselves (function, UI, etc.) to the device environment and the user.
I suspect that within 4 years we won't know or care which OS is running on which device...
We'll just satisfy our needs of the moment by choosing S-M-L-XL-XXX.
ARM doesnt manufacture the chips, it just licences the designs to chip makers, makes it difficult for Intel to undercut them on price.
Intel has the best fabs in the planet. Can produce chips cheaper than anyone else.
Firstly, they will never ship ARM based CPUs until they are at least 64 bit processors. By then the 128 bit market will be on the verge of hitting the server space.
Secondly, they will not remotely compete with the APU products coming out of AMD.
The Macbook Air will have to have a cluster of 64 bit ARM processors with up to 4 Cores per SoC to match the power of the current i5/i7 lineup.
The next major revision of CPUs from Intel and AMD will be battling it out with 8 and 10 core CPUs and their power dwarfs that of the embedded space.
The raw GFLOPS of the latest CPUs are surpassing 200 GFLOPS in their next revision. The Tegra from Nvidia's latest top GFLOP mark is 6.4. Sorry, but the power isn't remotely close.
Finally, the sheer volume of wafer production in the embedded space is already at it's limitations forcing 3rd parties to invest tens of billions just to keep up.
Apple isn't going to invest tens of billions for 3rd parties just to put 40 bit instruction set aware A15's into a Macbook Air.
OS X is a 64 bit through and through OS. The goal for Apple is to move iOS to a 64 bit code base. The ARM A8 is the first 64 bit OS ready design.
http://www.theregister.co.uk/2011/10...ldings_arm_v8/
They are a good 24-30 months away from it being in general use.
Good post!
I think it would be mutually beneficial for Apple to contract Intel as the foundry for its next gen ARM chips
I know, I know... what's done is done, the past is the bast -- but moving forward...
Which would you rather have doing your surgery?
The Macbook Air will have to have a cluster of 64 bit ARM processors with up to 4 Cores per SoC to match the power of the current i5/i7 lineup.
Your (flawed) assumption is that Macbook Air users really need the performance of an i5/i7 CPU.
If the iPad has shown us anything performance isn't everything.
90% of those buying computers rarely do anything beyond checking email, surfing the web, stream video, and word process/productivity.
x86 is like performing surgery with a chain saw.
Your (flawed) assumption is that Macbook Air users really need the performance of an i5/i7 CPU.
If the iPad has shown us anything performance isn't everything.
90% of those buying computers rarely do anything beyond checking email, surfing the web, stream video, and word process/productivity.
x86 is like performing surgery with a chain saw.
First, bad analogy. x86 can perform the entire range of "cutting" from surgery to cutting wood. Is it overkill for many tasks, sure. But it's perfectly capably of performing them all, unlike your "surgery wtih a chain saw" analogy.
Second, even something as seemingly benign as iPhoto can take up a lot of processing power for some of it's functions, such as facial recognition. The point being that just because you don't always need the power of a "full-sized" CPU, it doesn't mean you never need the it. You wouldn't buy a car that can't go over 35 mph just because you mostly drive in the city, would you?
First, bad analogy. x86 can perform the entire range of "cutting" from surgery to cutting wood. Is it overkill for many tasks, sure. But it's perfectly capably of performing them all, unlike your "surgery wtih a chain saw" analogy.
Second, even something as seemingly benign as iPhoto can take up a lot of processing power for some of it's functions, such as facial recognition. The point being that just because you don't always need the power of a "full-sized" CPU, it doesn't mean you never need the it. You wouldn't buy a car that can't go over 35 mph just because you mostly drive in the city, would you?
What you say is true!
As far as the Apple implementation of the ARM / GPU processing capabilities...
As I understand it, iOS can turn extra cores On or Off as needed and use them efficiently to maximize performance and minimize power.
I don't know that this level of granular control exists in x86 chips.
It would be neat to see Apple release an inexpensive hockey puck compute box similar in looks to the AppleTV 2.
This box would contain power supply, ARM SoCs, RAM, minimal SSD and 2 ThunderBolt ports.
You could just slap as many of these as needed on a ThunderBolt daisy-chain along with Computers, RAIDs, Displays.
The main Computer (maybe a Mini) would dispatch and monitor high power/performance compute tasks outboard to the hockey pucks.
Apple already has existing software that does this.
I see a future for an unified architecture in the computing space.
The situation for many(most) people at the moment is that we have a full-size computer on x86 and some sort of mobile computer on ARM, this is only becoming more prevelent, especially as our mobile computers(tablets, iPhones) become more powerful and common
But the software I have on one doesn't run on the other, i can't run msoffice on my iPad etc...
I think apple is moving this direction with features from ios making it to OSX, and with the testing of macs on ARM
The problem is that with ARM, it's just not a performance issue, but a compatability issue; none of the big Mac software will run without a recompile
What this new intel chip is; is a nearly competitive mobile processor that Could, theoretically run OSX and the existing Mac apps without needing to recompile, one that will get more competitive with the next generations
Imagine a iPad that could be your only PC, pair it with a dock and it's a desktop and the os switches to keyboard/mouse mode with multiple windows remove it from the dock and all your apps stay open and switch to touchscreen mode, full screen apps and gesture control
Apple already has the same os on both, just different interfaces (and as of now, different binaries for ARM)
Now, apple could go either way, switch OSX to ARM or switch IOS to x86
I think the latter is easier.
1. Most of the BIG expensive apps like photoshop, etc are x86 right now
2. Having users buy a new version to use on their new MacBook is cost prohibitive
3. Most iOS apps are free/less expensive
4. This intel chip, as well as the next generations can run arm binaries( through built in emulation)
A. This would mean most of the low-performance iOS apps, like utilities, basic games, etc... could run as-is on this(or the next generation) chip
5. the current and next generation arm chips are LESS powerful than intels desktop chips(though they are catching up)
A. You generally cannot emulate a higher performance system on a lower performance chip
B. as stated before, many of the popular apps on the Mac are bigger, higher performance apps, unsuitable for emulation
These points, assuming that In the next couple of generations intel can match Arm performance in the mobile arena, tell me that an iOS switch to x86 is more likley than a Mac switch to ARM
The proof will be when we are able to compare contemporaneously shipping products from both companies. Right now there are no shipping products with Intel's new chip and Intel is making performance claims relative to ARM chips that are in currently shipping products. If we compared ARMs future products to Intel's currently shipping products ARM would look a lot better.
Yeah in some cases they compare to year and a half old devices that are supposedly low power. Intel appears to be running about twice the power levels of these old devices.
We've seen this dance play out between Intel and AMD for years, and it's always the same -- you don't really know until you've got shipping versions of both products in front of you.
Exactly.
Yes, that would be a more likely scenario.
The original release of Mac OS X was delayed substantially until its release in early 2001. There have been persistent rumors (never confirmed) that much of the delay was caused by Apple insisting that the code run on x86 architecture on secret Intel-based prototypes. When Apple announced their switch from PowerPC to x86 architecture in the middle of the last decade, they were able to make a relatively smooth transition to the Intel processors.
Mac OS X 10.6 Snow Leopard was described by Apple as a total under-the-hood rewrite of the operating system. It is possible that ARM support was included in a parallel source tree and that each major revision of the code much run on secret ARM-based Mac prototypes.
Of course, we will never know for sure based on Apple's penchant for secrecy, but if Apple ever does release a Mac with an ARM processor, it will likely run OS X. At that point, it is likely that Apple will give developers to generate Universal binaries (Intel and ARM code combined) or to generate thin binaries that only support one architecture. With a few tweaks to the Mac App Store, it would probably be rather simple to deliver the correct thin binary to the target machine.
X86 is like a chain saw while ARM is like a scapule.
Which would you rather have doing your surgery?
Well for tree surgery a scalple would take a hell of a long time
Some things just need raw power
Not so easily, I don't think. Apple was able to make the jump from Motorola 68K to PowerPC because the PowerPC was powerful enough to emulate the 68K instruction set and still have acceptable performance. Same holds for the PowerPC to Intel jump. The Intel chips had the extra horsepower to run PowerPC code in emmulation. During both transitions most applications and even large chunks for the operating system itself were running in emmulation for months and years after the transition.
To allow a similar transition, the ARM chip would need to be sufficiently more powerful than the Intel chip it is replacing to run code for that Intel chip in emmulation at acceptable speeds. The only alternative would be to get the entire OS and a good representation of the application library 100% ARM native prior to the transition. I would not put that in the category of "easily" dumping Intel.
It may happen someday. But it won't be anytime soon and it won't be easy.
One thing Apple could do is abandon the market for high powered computers. If ARM chips could eventually produce the computing power of an i5 mid level chip, forget about the few i7's PC's sold and focus on what most people buy.
Performance without power considerations is meaningless, especially in the smartphone world. Luckily for Intel, Medfield seems very competitive there as well. Intel provided some power and performance data for Medfield based on its reference platform. I still haven't been able to verify any of this for myself, but I was able to see some power tests run in person on the reference platform and competitive devices.
The Intel provided values are pretty astonishing . Sub 20mW idle, sub 750mW during a call on 3G and although not pictured here, Intel's internal data suggests ~1W power consumption while browsing the web compared to ~1.3W on the iPhone 4S and Galaxy S 2. I've done my own measurements on 4S web browsing and came up with a very similar value.
Intel Measured Smartphone Power Consumption (Identical Display Brightness)
\tStandby (3G) \tTalk (3G) \tBrowsing (3G) \tVideo Playback 720p
Apple iPhone 4S \t~38mW \t~800mW \t~1.3W \t~500mW
Intel Medfield Reference \t~18mW \t~700mW \t~1.0W \t~850mW
Samsung Galaxy S II \t~19mW \t~675mW \t~1.2W \t~650mW
source: http://www.anandtech.com/show/5365/i...-smartphones/1
So from the looks of it, it looks like Intel has a significant performance lead with comparable power consumption. Now keep in mind that this is Intel-provided data, so take that however you will.
Also remember that phones using Medfield will not appear until at least Q2 2012; by then ARM A15 should be out, which should decrease the performance gap and may actually have better power consumption with big.LITTLE.
Finally remember that Medfield is built on 32nm while ARM designs are pretty much all 40/45nm right now. We're expected to see 32/28nm ARM designs this year so it will be interesting to see how much Intel is relying on its process node advantages when that happens.
While I'm still somewhat skeptical at this point, with the data released thus far, and with Intel finally throwing its weight behind Android and partnering up with Motorola and Lenovo, they definitely have potential here.
I could see how Medfield fits in though, if you were chasing ultra long battery life in a laptop. Medfield would still be a performance hit, but as it is x86 compatible it would not be as much of a dog as an ARM chip requiring emulation.
So the chip range could have:
ARM for IOS and Android devices;
Medfield for long life, low power home servers, desktops and laptops running OSX or Win8, and some Win8* or Android tablets*;
i7 etc for higher performance products running OSX or Win8.
* these will not be successful, as in the case of very low power mobile devices, ARM, not intel would be the architecture determining where the performance envelope puck is going to be for a long time to come.
Intel has the best fabs in the planet. Can produce chips cheaper than anyone else.
Probably, but ARM don't produce chips. It's hard to poach business from a company not in you business. How do you undercut zero?
Intel will have to try a different tactic, like making a better product. Better performance, power consumption, graphics, anything... They can't compete on price.
Do you have any data to support this contention? We know that the Wintel OEMs are seeing their sales lag because customers are buying tablets rather than laptops and desktops.
This can also mean they are delaying their purchase of a replacement computer because they purchased a tablet. Not that they have chosen a tablet and not having a computer.
I'm still waiting for you to name some real applications. All I see there are two terrible programs that have much better alternatives already in tablet form.
Could you name the better alternative?
Well for tree surgery a scalple would take a hell of a long time
Some things just need raw power
Try trimming a bonsai with a chainsaw!
It was said that Apple officials were impressed by the results, as the test machine performed "better than expected."
Every time I read a rumor about this, I feel like people are distorting what was likely originally conveyed. As has been said many a time, the ARM is an excellent chip, but in raw power it pales beside the Core2 chip inside the Air right now. A thing can perform quite poorly but still be "better than expected." So were Apple officials actually described as being impressed, or just not as depressed as they'd expected to be? Seriously. Someone should go back to their original source (assuming such a source was real and not a peyote dream) and ask that, specifically.
Could you name the better alternative?
Pages, Numbers, and obviously Keynote. That goes for the desktop, too.