I have little doubt there are ARM-powered Macs running OS X in a lab somewhere in Cupertino, possibly for years.
While we'll never know, it's possible they started doing this in 2012.
ARM unveiled their new 64-bit architecture in October 2011. Most likely Apple taped out one or two 64-bit chips that never saw service in a shipping product, but remained in the lab for study. Finally, in 2013, Apple's 64-bit design was ready for primetime, in the form of the Apple A7 processor. By then, there have surely been ARM-powered Macs in a lab.
If Intel has demonstrated something in its history is that it can innovate. Remember the Apple "PowerPC trounces Intel" FIASCO. Hopefully, Apple learned the lesson once and for all.
This is true if you credit Intel for creating the Athlon, Athlon 64 (first 64-bit desktop CPU), and Athlon X2 (first dual-core desktop CPU), and forget the Itanium FIASCO. Hopefully, Intel learned the lesson once and for all.
Intel (and everyone else) knows that Apple is the trend setter. If ARM has reached a point that the power per watt is more than sufficient for a standard user's traditional "PC" needs then Apple moving to ARM would be the start of a large of issue for Intel (and AMD).
We've gone over this already.
AMD has an ARM license and makes ARM chips. Just swap X86 shipments for ARM shipments and AMD is fine.
I have little doubt there are ARM-powered Macs running OS X in a lab somewhere in Cupertino, possibly for years.
While we'll never know, it's possible they started doing this in 2012.
ARM unveiled their new 64-bit architecture in October 2011. Most likely Apple taped out one or two 64-bit chips that never saw service in a shipping product, but remained in the lab for study. Finally, in 2013, Apple's 64-bit design was ready for primetime, in the form of the Apple A7 processor. By then, there have surely been ARM-powered Macs in a lab.
It needs to be said again. iOS is OS X.
Here's a speculative way for Apple to Phase Intel out if they were planning on it. When they start planning it, they will put both an Intel chip and an ARM chip in the same Mac, and initially, the ARM part will only run ARM software, and all binaries for the OS X version that supports this will support running either the Intel or the ARM binary. Then when users flip on power-management features, the Mac would prefer the ARM binary and turn off the Intel chip except when there is no ARM binary for something that is running. Users will be warned when software without an ARM binary is launched that it will impact their power management. Eventually all software on the App store will be required to have Intel and ARM binaries or they will be delisted. A simple recompile will solve most things except for those who used depreciated features from older frameworks. Aren't you glad you picked SWIFT now?
Then when there's enough established ARM base, Apple will start making the Intel part optional, first in the laptops, then the desktops. This may eventually leave only the Mac Pro and the high end iMac as the only standard Intel CPU offering.
I say this speculatively, since an Intel CPU and an ARM CPU sharing the same hardware resources is likely impossible, though it may be possible for AMD to releash such an animal http://www.enterprisetech.com/2014/08/11/amd-details-seattle-arm-server-chip/ "These next-generation ARM chips, expected in 2015, are being developed under Project SkyBridge, which will create identical pinouts for ARM and X86 variants of the Opteron chips so companies can use them interchangeably on system board."
We endeavor to innovate so they'll continue to look to us as a supplier," Intel's Ultrabook chief Greg Welch said when asked about Apple's potential plans in 2011.
Where "innovate" means "flog the same old x86 architecture for all it's worth."
........ it would reduce the cost of "PCs" thereby increasing unit sales without reducing the cost for Windows licenses to the OEMs.
There are rumours that with Windows 10 MS wants to 'monetise Windows in a different way", so quite what this would mean for OEMs isn't clear. Whether it means a simple alteration in license costs or a move to a subscription model for all users, consumer or business remains unknown. I imagine though that it would have an impact both on consumer perceptions and OEM profit margins.
AMD has an ARM license and makes ARM chips. Just swap X86 shipments for ARM shipments and AMD is fine.
In no way does having an ARM license make a company "fine". There would be a lot more players and AMD suffer at both ends, but they will surely suffer with x86 if Apple sets an ARM trend on traditional "PCs."
It's inevitable. AX performance is nearly good enough for consumer use of OS X right now.
All Apple needs to do is add one more CPU to the 3-core A8X. Or maybe 3 cores is enough already.
The A7 crossed the 64-bit barrier way back in 2013 so that's no longer a gating factor.
And I think these periodic rumors about AX-powered MacBook Air are being planted by Apple.
To help prepare Intel and the rest of the world for Intel-free consumer Macs.
Great quote from a great article…
"At the launch of the iPhone 5s, Apple referred to the A7 as being "desktop class" - it turns out that wasn't an exaggeration."
Here's a speculative way for Apple to Phase Intel out if they were planning on it. When they start planning it, they will put both an Intel chip and an ARM chip in the same Mac, and initially, the ARM part will only run ARM software, and all binaries for the OS X version that supports this will support running either the Intel or the ARM binary. Then when users flip on power-management features, the Mac would prefer the ARM binary and turn off the Intel chip except when there is no ARM binary for something that is running. Users will be warned when software without an ARM binary is launched that it will impact their power management."
That seems like a fairly complex solution, especially from a users perspective. I can't see Apple releasing a product with Windows style warning messages about power management popping up whenever you open certain programs.
I agree that Apple would need to continue with Intel chips for its high end Macs at this stage though. So right now I can't see Apple transitioning to ARM for the Mac. I think they could down the road though, once they have a custom ARM chip that is capable of serving the Mac Pro as well or better than the x86.
The market debacle of IBM leaving the desktop market to a crappy Motorola doesn't change the fact that RISC > CISC. ARM v. x86 proves that more in today's market.
IBM basically decided (and rightfully so at the time) that the Apple market was not where they wanted to invest resources - they wanted to fight Intel harder in the big iron market, and it has been successful for them.
Intel is a very strong and competitive company. They won't back down from a challenge. Intel will keep innovating on x64 to keep Apple.
I seriously doubt that any ARM design will be as powerful as an Intel chip without running into the same problems as Intel has had to deal with. After all, it is the laws of physics that are limiting factors.
I tend to believe Apple is going to switch, but if that's their intent, I wonder if there's anything Intel can do to keep them a customer. I can think of a few strategies for Intel:
1. Cut prices.
2. Offer to fab Apple's ARM chips (if you can't beat em', join em')
3. Offer Apple early access to next gen Core chips (that is, keep Apple first in line for new chips, giving them a 3-6 month lead over other OEMs)
I can't believe they'd do #1. It would just facilitate Apple gaining marketshare at the expense of OEMs that pay a higher price for the same CPUs. Intel just can't do this.
The second one seems more possible. The devil would definitely be in the details, though. Apple would have to make sure they really get Intel's best effort. They'd probably have to get Intel to build a fab specifically for Apple, there would have to be penalties if new processes didn't come to Apple first, etc. I doubt Intel would accept the terms Apple would demand, but I think this is more likely that a price cut.
The third one seems most likely. Intel has given Apple preferential treatment in this way before -- they could do it again to keep Apple on board. But they would probably only agree to do it for 2 or 3 years, at most.
My hunch is that none of these will work out. Ultimately their interests are too much at odds. Intel doesn't want any single OEM to become too big and powerful. It wants them all to fight for scraps while Intel sucks in most of the profits. So any favoritism Intel shows to a customer is really just an attempt to weaken another customer that is becoming too much of a threat. Lately those subsidies have been indirectly aimed at weakening Apple (contra revenue for Atom-based tablets; support for ultrabooks). Obviously Apple really doesn't like that.
Intel is a very strong and competitive company. They won't back down from a challenge. Intel will keep innovating on x64 to keep Apple.
I seriously doubt that any ARM design will be as powerful as an Intel chip without running into the same problems as Intel has had to deal with. After all, it is the laws of physics that are limiting factors.
research the difference between CISC & RISC microprocessors.
This is exactly the reason ARM can do more with less compared to Intel
Coincidentally, it is the same thing that had, mhz for mhz, PowerPC out performing x86.
It is not about the laws of physics, rather about the decades of baggage that is attached to the instruction set Intel has chosen to maintain.
Windows RT shows that a full desktop system can be ported over to ARM ok and this example shows why adding display touch support is not a good idea because even for occasional use, users would be frustrated by it:
[VIDEO]
The benefits that an ARM design gives Apple is that it would be cheaper, maybe longer lasting on battery, certainly fanless. If they use a mobile GPU, it will use OpenGL ES so the CPU and GPU will cause software incompatibility. It would work as a basic Mac though.
They couldn't do a transition like they did with PPC to Intel and just stop using Intel. Having an ARM model on the low-end shouldn't be much of a threat to Intel at all, it just helps Apple grow its Mac userbase. They could double their Mac unit sales and still sell the same amount of Intel models they already do.
I actually have a lot of nostalgia for PPC. The promise of RISC processors was a good one, but IBM just wasn't up to the task of looking toward efficiency and Intel caught up. I still, at least in concept, prefer the RISC model.
Comments
I have little doubt there are ARM-powered Macs running OS X in a lab somewhere in Cupertino, possibly for years.
While we'll never know, it's possible they started doing this in 2012.
ARM unveiled their new 64-bit architecture in October 2011. Most likely Apple taped out one or two 64-bit chips that never saw service in a shipping product, but remained in the lab for study. Finally, in 2013, Apple's 64-bit design was ready for primetime, in the form of the Apple A7 processor. By then, there have surely been ARM-powered Macs in a lab.
This is true if you credit Intel for creating the Athlon, Athlon 64 (first 64-bit desktop CPU), and Athlon X2 (first dual-core desktop CPU), and forget the Itanium FIASCO. Hopefully, Intel learned the lesson once and for all.
Intel (and everyone else) knows that Apple is the trend setter. If ARM has reached a point that the power per watt is more than sufficient for a standard user's traditional "PC" needs then Apple moving to ARM would be the start of a large of issue for Intel (and AMD).
We've gone over this already.
AMD has an ARM license and makes ARM chips. Just swap X86 shipments for ARM shipments and AMD is fine.
It needs to be said again. iOS is OS X.
Here's a speculative way for Apple to Phase Intel out if they were planning on it. When they start planning it, they will put both an Intel chip and an ARM chip in the same Mac, and initially, the ARM part will only run ARM software, and all binaries for the OS X version that supports this will support running either the Intel or the ARM binary. Then when users flip on power-management features, the Mac would prefer the ARM binary and turn off the Intel chip except when there is no ARM binary for something that is running. Users will be warned when software without an ARM binary is launched that it will impact their power management. Eventually all software on the App store will be required to have Intel and ARM binaries or they will be delisted. A simple recompile will solve most things except for those who used depreciated features from older frameworks. Aren't you glad you picked SWIFT now?
Then when there's enough established ARM base, Apple will start making the Intel part optional, first in the laptops, then the desktops. This may eventually leave only the Mac Pro and the high end iMac as the only standard Intel CPU offering.
I say this speculatively, since an Intel CPU and an ARM CPU sharing the same hardware resources is likely impossible, though it may be possible for AMD to releash such an animal http://www.enterprisetech.com/2014/08/11/amd-details-seattle-arm-server-chip/ "These next-generation ARM chips, expected in 2015, are being developed under Project SkyBridge, which will create identical pinouts for ARM and X86 variants of the Opteron chips so companies can use them interchangeably on system board."
We endeavor to innovate so they'll continue to look to us as a supplier," Intel's Ultrabook chief Greg Welch said when asked about Apple's potential plans in 2011.
Where "innovate" means "flog the same old x86 architecture for all it's worth."
........ it would reduce the cost of "PCs" thereby increasing unit sales without reducing the cost for Windows licenses to the OEMs.
There are rumours that with Windows 10 MS wants to 'monetise Windows in a different way", so quite what this would mean for OEMs isn't clear. Whether it means a simple alteration in license costs or a move to a subscription model for all users, consumer or business remains unknown. I imagine though that it would have an impact both on consumer perceptions and OEM profit margins.
I can see it 5 years from now...Intel will only be in Pro level Macs. Mac Air and Mini may run on custom Apple chips.
It's inevitable. AX performance is nearly good enough for consumer use of OS X right now.
All Apple needs to do is add one more CPU to the 3-core A8X. Or maybe 3 cores is enough already.
The A7 crossed the 64-bit barrier way back in 2013 so that's no longer a gating factor.
And I think these periodic rumors about AX-powered MacBook Air are being planted by Apple.
To help prepare Intel and the rest of the world for Intel-free consumer Macs.
In no way does having an ARM license make a company "fine". There would be a lot more players and AMD suffer at both ends, but they will surely suffer with x86 if Apple sets an ARM trend on traditional "PCs."
Great quote from a great article…
"At the launch of the iPhone 5s, Apple referred to the A7 as being "desktop class" - it turns out that wasn't an exaggeration."
Of course, now that Anand works for Apple some will probably say anything positive he's ever said about Apple is now just propoganda.
...
"At the launch of the iPhone 5s, Apple referred to the A7 as being "desktop class" - it turns out that wasn't an exaggeration."
Of course, now that Anand works for Apple some will probably say anything positive he's ever said about Apple is now just propoganda.
I can picture Anand's screening interview:
Interviewer: "So can I see some of your prior work?"
Anand: "Oh yes you can! Slap your eyes on THIS!"
That is probably what he is thinking ...
That seems like a fairly complex solution, especially from a users perspective. I can't see Apple releasing a product with Windows style warning messages about power management popping up whenever you open certain programs.
I agree that Apple would need to continue with Intel chips for its high end Macs at this stage though. So right now I can't see Apple transitioning to ARM for the Mac. I think they could down the road though, once they have a custom ARM chip that is capable of serving the Mac Pro as well or better than the x86.
IBM Power processors still kick a$$.
The market debacle of IBM leaving the desktop market to a crappy Motorola doesn't change the fact that RISC > CISC. ARM v. x86 proves that more in today's market.
IBM basically decided (and rightfully so at the time) that the Apple market was not where they wanted to invest resources - they wanted to fight Intel harder in the big iron market, and it has been successful for them.
I seriously doubt that any ARM design will be as powerful as an Intel chip without running into the same problems as Intel has had to deal with. After all, it is the laws of physics that are limiting factors.
I really hate to say this but this guy's face does not say "innovation" to me... It says more like "Dead Man Walking."
Yeah, he looks like an IBM executive from 1955.
I tend to believe Apple is going to switch, but if that's their intent, I wonder if there's anything Intel can do to keep them a customer. I can think of a few strategies for Intel:
1. Cut prices.
2. Offer to fab Apple's ARM chips (if you can't beat em', join em')
3. Offer Apple early access to next gen Core chips (that is, keep Apple first in line for new chips, giving them a 3-6 month lead over other OEMs)
I can't believe they'd do #1. It would just facilitate Apple gaining marketshare at the expense of OEMs that pay a higher price for the same CPUs. Intel just can't do this.
The second one seems more possible. The devil would definitely be in the details, though. Apple would have to make sure they really get Intel's best effort. They'd probably have to get Intel to build a fab specifically for Apple, there would have to be penalties if new processes didn't come to Apple first, etc. I doubt Intel would accept the terms Apple would demand, but I think this is more likely that a price cut.
The third one seems most likely. Intel has given Apple preferential treatment in this way before -- they could do it again to keep Apple on board. But they would probably only agree to do it for 2 or 3 years, at most.
My hunch is that none of these will work out. Ultimately their interests are too much at odds. Intel doesn't want any single OEM to become too big and powerful. It wants them all to fight for scraps while Intel sucks in most of the profits. So any favoritism Intel shows to a customer is really just an attempt to weaken another customer that is becoming too much of a threat. Lately those subsidies have been indirectly aimed at weakening Apple (contra revenue for Atom-based tablets; support for ultrabooks). Obviously Apple really doesn't like that.
Intel is a very strong and competitive company. They won't back down from a challenge. Intel will keep innovating on x64 to keep Apple.
I seriously doubt that any ARM design will be as powerful as an Intel chip without running into the same problems as Intel has had to deal with. After all, it is the laws of physics that are limiting factors.
research the difference between CISC & RISC microprocessors.
This is exactly the reason ARM can do more with less compared to Intel
Coincidentally, it is the same thing that had, mhz for mhz, PowerPC out performing x86.
It is not about the laws of physics, rather about the decades of baggage that is attached to the instruction set Intel has chosen to maintain.
iOS is the mass market. Windows apps that are Windows only are the niche now.
Apple is a mass market company.
Most people probably use an iOS device now and most of them are probably Windows users, not that that makes a difference.
iOS has the leverage.
Although we have Mac, Windows and Chromebook, our entire family uses iOS primarily (iPhone & iPad).
In fact, an iPhone 6+ could be your overall primary computer. Add a keyboard and you have your mobile computer. Or iPad.
Anyway, the question is: why does Apple need Intel? Really. They are in the cat bird seat.
In certain regards. For Apple's need, which is a heavy focus on low-power, they aren't even in the running.
[VIDEO]
The benefits that an ARM design gives Apple is that it would be cheaper, maybe longer lasting on battery, certainly fanless. If they use a mobile GPU, it will use OpenGL ES so the CPU and GPU will cause software incompatibility. It would work as a basic Mac though.
They couldn't do a transition like they did with PPC to Intel and just stop using Intel. Having an ARM model on the low-end shouldn't be much of a threat to Intel at all, it just helps Apple grow its Mac userbase. They could double their Mac unit sales and still sell the same amount of Intel models they already do.
I actually have a lot of nostalgia for PPC. The promise of RISC processors was a good one, but IBM just wasn't up to the task of looking toward efficiency and Intel caught up. I still, at least in concept, prefer the RISC model.