Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple to Move from Intel to ARM Processors in Future Laptops?
- Thread starter MacRumors
- Start date
- Sort by reaction score
- Status
Apple is gonna shove icloud down everyone's throats and we will eat it all up!!!
Average Joe will get a device(iphone,ipad,netbook,nettop) they will all have arm chips. Apple will still make macpros and macbookpros for people with those needs. The pro models could use 2 or 4 arm chips with many cores! People will have pics/videos uploaded to the cloud. If you need to edit your pictures or videos the cloud will do the work. In 4 or 5 years most gaming will be handled by streaming "onlive" type services... no need for high performance graphics cards.... This is the future! I see no reason apple wouldn't move to arm.
Average Joe will get a device(iphone,ipad,netbook,nettop) they will all have arm chips. Apple will still make macpros and macbookpros for people with those needs. The pro models could use 2 or 4 arm chips with many cores! People will have pics/videos uploaded to the cloud. If you need to edit your pictures or videos the cloud will do the work. In 4 or 5 years most gaming will be handled by streaming "onlive" type services... no need for high performance graphics cards.... This is the future! I see no reason apple wouldn't move to arm.
switching to ARM means no more boot camp
As others have mentioned, Windows 8 will be able to run on ARM. It is slated for release in 2012 or 2013.
As much as I do not want to see another architectural shift, there is a distinct possibility that this rumor is true. According to a recent article:
"ARM expects to ship dual-core Cortex-A15 SoCs in 2012 . . . the company predicts devices running on the 2.5GHz Cortex-A15 will actually reach store shelves before the end of 2012, with quad-core variants showing up in 2013."
The article mentions that the A15 is expected to perform at a level five fold that of the A9 chip and does state that the "quad-core SoCs expected for 2013 [are] geared towards tablets rather than phones," but I could see Apple moving toward convergence of OS X with iOS--the NeXT big thing. Again, in light of Apple's push toward tablets and iOS devices, this rumor makes sense. I just don't like it.
Last edited:
Why do people keep saying that ?
iOS and OS X already share the kernel and Foundation, not to mention Quartz. That's about the only parts you can converge. The UI itself (Appkit vs UIKit) use completely different paradigms, so there's no real possible merge there, nor would you want or wish to.
ARM does nothing to converge things. You're still linking to different frameworks for your desktop/laptop application than for your tablet/phone application.
If you guys want to talk convergence : Be precise what you mean.
yeah windows 8, what about windows xp or 7?
Even by 2013 those ARM chips are not likely to rival the quad core i series processors we have today, not to mention what intel comes out with in the next 2 years (haswell and beyond)
Apple is not going to turn osx/ios into one operating system... they have already said they won't bring touch to the desktop, it's awkward and doesn't make sense.
I think the chips themselves last longer than they are useful for already. From my understanding, power draw does go up, but the thermal output doesn't appear to change significantly (ie the stock intel coolers can handle the clocks just fine). At least this is the case the the Core iX line.
I can't imagine anyone being interested in running Windows XP in 2013. By then, XP will be as old as Windows 98 is today.
Hell, even assuming this happens, Microsoft will be into development for Windows 9 by then.
Hell, even assuming this happens, Microsoft will be into development for Windows 9 by then.
God.
People really don't understand how OS X operates.
Just like the Intel switch, the heavy lifting would all be done by Apple and developers who followed Apple guidelines wouldn't be impacted by the switch.
Everything you are running runs on top of API layers that run on top of the core system. As long as Apple doesn't break those API layers, you should be (more or less) good to go.
People really don't understand how OS X operates.
Just like the Intel switch, the heavy lifting would all be done by Apple and developers who followed Apple guidelines wouldn't be impacted by the switch.
Everything you are running runs on top of API layers that run on top of the core system. As long as Apple doesn't break those API layers, you should be (more or less) good to go.
I really hope this is just a rumor. Like someone posted earlier, if apple does switch, intel is not going to be far behind. The ARM architecture is simply not powerful enough, and doesn't have the muscle of an intel proc.
the temperature definitely goes up the more you tax the processor...why do you think the macbook gets hot when you are doing cpu intensive tasks (encoding a movie for example)?
Agree with the above but I'll add more. Where will we be in ten years? In 20? I think maybe we might be talking about "kilo-cores" in 20 years.
I've hear so many people say "I don't need more than four." Yes you don't need more than four to run the software you run today but what software will you be using in 20 years. I suspect you wil be able to
(1) Point with fingers, like you do with a multi touch pad but only in the air in 3D space in front of your screen. Stereo web-cab will track all 10 fingers and your eyes. as you move a 3d object that floats in front of the screen.
(2) voice will be used to some extent. Most likely to do a google-like search of voice messages or other audio/video recordings. (A student might ask his tablet to find the place where his teacher talked about some subject.)
(3) the "Find my car keys" app lets you wave the camera around your house and it will find some object for you.
(4) Searching a text book by "idea" rather than text string. "Find where it talks about that civil war battle, the one that had Gen McClellan and that bridge". (Yes I purposely messed up the question, software should figure it out.)
(5) Automated assistant app for phone: you say to the phone: "call fred, tell him I can't make the meeting this afternoon." The phone figures out which of ten Freds and places a voice mesage in his phone for you.
OK you get the idea, NONE of the above is going to happen on an 8-core Intel CPU. Those are kilo-core problems and all those 1,000 cores need to run on battery power.
More or less.... regardless of what api and language a developer uses, they still need to compile different machine code at some point. Contrary to what Apple would have you believe, sometimes it's a lot more than a click of the checkbox in Xcode. Especially applications that are heavily optimized (like multimedia, graphics, or any heavy lifting), require some custom processor routines.
A lot can change in a few years' time ....
I don't think this was "ridiculous" at all. Rather, it's just a little premature and in the realm of "an interesting, educated guess at a possibility a lot of people probably hadn't even considered yet".
I was just discussing something similar with a few people over on Slashdot's web site. They brought up the fact that ARM processors are currently outselling Intel processors, to the extent that there are now more ARM processors in circulation than ALL the CPU's Intel ever sold, put together!
THAT is why Windows 8 plans on having ARM support. Microsoft sees the numbers and the trend, and doesn't want to be left out of it.
Now, at the present time, these ARM processors are finding their way into cellphones and other dedicated gadgets/devices, primarily, because they're not designed to compete directly with higher-powered CPUs like Intel or AMD sell. But if (and really, this is a big "if") Apple can come up with a line of ARM CPUs that are on parity with Intel's offerings, in the next 2-3 years? Then yeah - I can see a transition happening.
The problem for Apple with being tied to using Intel CPUs is that they can't really differentiate their hardware from everyone else's, beyond just the styling and the details (like the mag-safe charging connection on their laptops, or the backlit keys on their Macbook Pro keyboards). OS X remains the only really important reason people would pay more for a Mac system.
I have a feeling that all other things being equal, Apple would feel a lot more comfortable if they offered a machine based around a CPU that was actually made in facilities they owned, by a company they owned. Then they could do hardware updates on a cycle of their own choosing, instead of being pushed to release a new model based around whenever Intel had a new processor model ready. (Right now, even when Intel gives them "first dibs" on something new, that means Apple has a short time window to get something out the door that uses it, before it goes mainstream and Dell, HP, and everyone else has it too.)
I don't think this was "ridiculous" at all. Rather, it's just a little premature and in the realm of "an interesting, educated guess at a possibility a lot of people probably hadn't even considered yet".
I was just discussing something similar with a few people over on Slashdot's web site. They brought up the fact that ARM processors are currently outselling Intel processors, to the extent that there are now more ARM processors in circulation than ALL the CPU's Intel ever sold, put together!
THAT is why Windows 8 plans on having ARM support. Microsoft sees the numbers and the trend, and doesn't want to be left out of it.
Now, at the present time, these ARM processors are finding their way into cellphones and other dedicated gadgets/devices, primarily, because they're not designed to compete directly with higher-powered CPUs like Intel or AMD sell. But if (and really, this is a big "if") Apple can come up with a line of ARM CPUs that are on parity with Intel's offerings, in the next 2-3 years? Then yeah - I can see a transition happening.
The problem for Apple with being tied to using Intel CPUs is that they can't really differentiate their hardware from everyone else's, beyond just the styling and the details (like the mag-safe charging connection on their laptops, or the backlit keys on their Macbook Pro keyboards). OS X remains the only really important reason people would pay more for a Mac system.
I have a feeling that all other things being equal, Apple would feel a lot more comfortable if they offered a machine based around a CPU that was actually made in facilities they owned, by a company they owned. Then they could do hardware updates on a cycle of their own choosing, instead of being pushed to release a new model based around whenever Intel had a new processor model ready. (Right now, even when Intel gives them "first dibs" on something new, that means Apple has a short time window to get something out the door that uses it, before it goes mainstream and Dell, HP, and everyone else has it too.)
I think you are right on the money. I've brought up a few times now my prediction that Apple is eventually planning to fully merge OSX into iOS (instead of the other way around) and the evidence just keeps piling up. It makes zero sense to move to ARM processors (and kill all software compatibility with current and past software in the process, save iOS stuff), especially given Intel (and also technically AMD) is THE prime CPU market. It will almost undoubtedly be the fastest for ages to come (after all, isn't a lack of updates/progress/speed the main reason Apple switched to Intel from PPC in the first place? Isn't the compatibility with Windows also very helpful in bringing in new customers and letting Mac users operate in areas that would not otherwise want Apple hardware?)
Put simply, it would be sheer suicide for the Mac lines to move to ARM...unless of course by then a good chunk of the software is already merged to iOS and the App store. Thus, when the ARM MBP finally comes out, you won't NEED boxed/open software. The App store will have 100% of the available software and it will be the only place you are allowed to buy software. Apple will collect its 30% share of every single developers' hard work in the process and Steve Jobs will officially be the richest man in the Universe (assuming he's still alive and 'God' hasn't struck him down for being the greediest bugger on the face of the Earth).
Think about it. It's not hard to believe. Apple already regularly makes changes to the operating system that leaves software incompatible (bye bye carbon; bye bye 32-bit next and finally bye bye compiled Intel code last along with all open software to go with it). It's the Jobs Master Plan. You will put up with it or you will find yourself moving to Windows or Linux once and for all. I know I will happily move there at that point as I have no interest in a closed system with overpriced software running on painfully slower computers than Windows machines have (already painful enough in GPU terms).
Believe it or not, it's going to happen. It's only a matter of time. The only thing that could possibly stop it is if Mr. Jobs is no longer the CEO of Apple in the next year or two (for whatever reason). Maybe then someone with more sense could prevent the merging of the two lines. But otherwise, forget about it. Intel is history. Open software will be history and the Mac and Windows machines will never compete directly again. It will be like Wii Versus XBox 360. They're just not the same thing at all, even if some find themselves having to choose between them.
You might find some hybridization in the desktop lines (or at least the Mac Pro) for awhile given its need for bigger CPU and professional applications. More than likely, though, the Mac Pro will simply be abandoned in the future (like the Xserver line already) and Apple will stop the charade of pretending they still care about the professional market at all (the "Pro" models are already mostly consumer gear with a 'pro' name).
It's not more or less. His comment was spot on. If you stick to APIs and Apple ports those APIs, you are source compatible and a compile away from shipping off to Q&A.
These days, applications don't use hand written assembly. Optimizing compilers are much more efficient than humans, even in heavy lifting/graphics/multimedia. No one writes graphics/multimedia code in assembly anyway, since modern operating systems don't expose GPUs in a way that permits it nor is anyone knowledgeable enough about all the different GPUs to do it.
So the only point left is endianess, if you mess around with code that is dependant on the byte order (and bit order) in memory. This is very few and far between, though I do have 1 or 2 routines that still depend somewhat on this (bad hacks I use in my iOS stuff). ARM is bi-endian and thus Apple could simply use the Intel little-endian configuration on their ARM stuff to make this point moot.
why do you think the macbook gets hot when you are doing cpu intensive tasks (encoding a movie for example)?
A 4ghz machine at idle shouldn't have much of a different profile than a 3Ghz machine at idle. Under load, with proper cooling there will be a difference but not a whole lot.
To be honest it does depend on what binning you get. I haven't messed with OCing in a few years so there appears to be more voltage tricks that one needs to run CPU's at a higher rate. I have noticed some enterprising individuals on overclockers.com have gotten the i7-2600k up to 5ghz. Not sure if that is on Air though (it appears to be).
number 1 could be done today with kinect
the rest sounds more like it is a software development issue. Just look at how hard it was for IBM to create watson (the jeopardy computer)
Rosetta
Another thing that people aren't thinking about is the App store. I seem to recall on the transition to intel Rosetta worked by converting executable PPC code to intel code. Once the conversion was complete the generated Intel executable just ran. This added overhead because Rosetta had to perform a translation before the code could execute.
With the App store this Rosetta conversion could be done on Apple's server farms. Think about it, on day 1 every item in the App store could be available in a native binary format with no work by developers.
The other thing that Apple would gain from moving to ARM is being able to create their own SOC with their own logic decoders/encoders. ARM does not create any chips, they design chips, and building blocks for chips then let others build them. This gives Apple ultimate flexibility. If Apple wanted a 16-core A15, they could create it.
Another thing that people aren't thinking about is the App store. I seem to recall on the transition to intel Rosetta worked by converting executable PPC code to intel code. Once the conversion was complete the generated Intel executable just ran. This added overhead because Rosetta had to perform a translation before the code could execute.
With the App store this Rosetta conversion could be done on Apple's server farms. Think about it, on day 1 every item in the App store could be available in a native binary format with no work by developers.
The other thing that Apple would gain from moving to ARM is being able to create their own SOC with their own logic decoders/encoders. ARM does not create any chips, they design chips, and building blocks for chips then let others build them. This gives Apple ultimate flexibility. If Apple wanted a 16-core A15, they could create it.
Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_3_2 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8H7 Safari/6533.18.5)
As evidenced by the Atom processor? Intel and AMD have backed themselves into a corner with X86. Granted Intel has the resources to make it in a post X86 world but they haven't shown any real commitment to do so. I agree that it's a bit premature to write X86 off but I don't think the ARM scenario is unrealistic. Mobile OS's won't replace X86 OS's completely but I do think mobile OS's are going to dominate the computer market in the next 5 to ten years.
lPHONE said:
As evidenced by the Atom processor? Intel and AMD have backed themselves into a corner with X86. Granted Intel has the resources to make it in a post X86 world but they haven't shown any real commitment to do so. I agree that it's a bit premature to write X86 off but I don't think the ARM scenario is unrealistic. Mobile OS's won't replace X86 OS's completely but I do think mobile OS's are going to dominate the computer market in the next 5 to ten years.
atom is doing fairly well. maybe netbooks are going away, but intel's TV chip business which is based on Atom is doing pretty good. it's just not publicized and like most Mac fans unless it's a consumer product with advertising they don't know it exists
and Intel's recent announcement combined with 22nm tech will help a lot. most ARM CPU's are made on much older tech which is more power hungry and more expensive
unlike previous new CPU's where only one factory made it, intel plans on having 4 factories make 22nm CPU's at launch. including the Atom
x86 is just the instruction set. the way the CPU's work now is very different than when intel first came up with the instruction set
It will take three years for Atom to sync with the leading tech. The next iteration will be 32nm.
Oh man that is priceless. Thanks for digging that old thread up.
--C. Alan
- Status