You've already drunk the Kool-Aid and this isn't even a thing yet.
Last edited:
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 Plans to Ditch Intel and Use Custom Mac Chips Starting in 2020
- Thread starter MacRumors
- Start date
- Sort by reaction score
You've already drunk the Kool-Aid and this isn't even a thing yet.
Many people here seem to think Apple is capable of immediately taking the desktop CPU performance crown from Intel. There's a good chance that they wouldn't even manage to compete with AMD.
Many also seem to think that CPU performance has stagnated. That's not really true. Single-threaded performance has steadily improved year on year by 5-10%. Energy efficiency has also steadily improved. For the 'mainstream' platforms - i.e. iMacs and MacBooks, it's true that Intel hasn't delivered game-changing products (until 8th gen with quad core CPUs suitable for the ultrabook form factor). Look at AMD - even factoring in Ryzen, their year-on-year gains are no better than Intel's - and you can bet that both these companies are working very hard to uncover additional performance - especially single-threaded.
The real truth is that the low hanging fruit has disappeared, and getting these single-threaded gains for *high performance* CPUs is very difficult. You can't simply add more cores - you have to have an interconnect, memory bus, cache design etc that allows you to scale the core count. You can't simply add more frequency, because that increases the thermal envelope, power usage etc.
Getting CPU gains for the low energy ARM CPUs has been a different story - there was plenty of low-hanging fruit to tick off, and as a result we've seen multiple companies make huge strides (Apple, Qualcomm, Samsung among others). The signs are that this rate of improvement is also starting to slow down.
Now - instead of looking at mainstream consumer products, look instead at the high-end desktop and server segments. Total compute power per socket, along with bus bandwidth, has improved *dramatically* in the last 10 years. Where once we had 4 cores per CPU we now have 26! This is one of the areas where Intel (and AMD) have substantial design experience and patent warchests that make it difficult for other designers to compete.
CISC vs RISC isn't really an argument here either. X86 has some baggage, but modern X86 CPUs get rid of that pretty damn well, and it's very difficult to argue that modern ARM is really a RISC architecture and doesn't have its own oddities and baggage. Certainly on a laptop or desktop the 'X86 tax' isn't going to be that significant.
As for any actual transition to ARM, a non-x86 Mac would almost certainly kill my interest in the Mac. Apple have done a pretty good job of pushing me into a corner with their decisions of the past 5 years, and even without an ARM transition it's far from certain whether I will ever buy another Mac as my primary machine (after always owning a personal Mac for the last 15 years).
PPC to Intel had real benefits, and even then it had a significant amount of pain with older software and Rosetta. Intel's chips were hugely more performant than IBM/Motorola's, which reduced the pain a lot, and Apple was able to immediately take advantage of better interoperability. Boot camp. Virtualization allowing VMWare Fusion and Parallels. This was *huge*.
What is the benefit to moving to ARM? Better battery life? Unlikely to be much. Better performance? Unlikely. Virtualization of ARM? Well, iOS is the closed platform - developers already compile iOS apps to x86 to run them in the simulator - it's easier to make iOS work on the X86 Mac than make the rest of the world work with an ARM Mac.
If this really is Apple's plan, Cook is going to return the Mac to it's NuBus days. Yuck.
Many also seem to think that CPU performance has stagnated. That's not really true. Single-threaded performance has steadily improved year on year by 5-10%. Energy efficiency has also steadily improved. For the 'mainstream' platforms - i.e. iMacs and MacBooks, it's true that Intel hasn't delivered game-changing products (until 8th gen with quad core CPUs suitable for the ultrabook form factor). Look at AMD - even factoring in Ryzen, their year-on-year gains are no better than Intel's - and you can bet that both these companies are working very hard to uncover additional performance - especially single-threaded.
The real truth is that the low hanging fruit has disappeared, and getting these single-threaded gains for *high performance* CPUs is very difficult. You can't simply add more cores - you have to have an interconnect, memory bus, cache design etc that allows you to scale the core count. You can't simply add more frequency, because that increases the thermal envelope, power usage etc.
Getting CPU gains for the low energy ARM CPUs has been a different story - there was plenty of low-hanging fruit to tick off, and as a result we've seen multiple companies make huge strides (Apple, Qualcomm, Samsung among others). The signs are that this rate of improvement is also starting to slow down.
Now - instead of looking at mainstream consumer products, look instead at the high-end desktop and server segments. Total compute power per socket, along with bus bandwidth, has improved *dramatically* in the last 10 years. Where once we had 4 cores per CPU we now have 26! This is one of the areas where Intel (and AMD) have substantial design experience and patent warchests that make it difficult for other designers to compete.
CISC vs RISC isn't really an argument here either. X86 has some baggage, but modern X86 CPUs get rid of that pretty damn well, and it's very difficult to argue that modern ARM is really a RISC architecture and doesn't have its own oddities and baggage. Certainly on a laptop or desktop the 'X86 tax' isn't going to be that significant.
As for any actual transition to ARM, a non-x86 Mac would almost certainly kill my interest in the Mac. Apple have done a pretty good job of pushing me into a corner with their decisions of the past 5 years, and even without an ARM transition it's far from certain whether I will ever buy another Mac as my primary machine (after always owning a personal Mac for the last 15 years).
PPC to Intel had real benefits, and even then it had a significant amount of pain with older software and Rosetta. Intel's chips were hugely more performant than IBM/Motorola's, which reduced the pain a lot, and Apple was able to immediately take advantage of better interoperability. Boot camp. Virtualization allowing VMWare Fusion and Parallels. This was *huge*.
What is the benefit to moving to ARM? Better battery life? Unlikely to be much. Better performance? Unlikely. Virtualization of ARM? Well, iOS is the closed platform - developers already compile iOS apps to x86 to run them in the simulator - it's easier to make iOS work on the X86 Mac than make the rest of the world work with an ARM Mac.
If this really is Apple's plan, Cook is going to return the Mac to it's NuBus days. Yuck.
Here is the thing: every IAx86-64 CPU has a chunk of chip real estate set aside for handling 32-bit code. They are essentially trapped by their backward-compatibility. No one wants an Intel chip unless it has 32-bit capability – well, maybe two or three people.
Apple burned 32-bit apps to the ground with iOS 11. Lost some old stuff, some of it pretty good but neglected. So now, as their own chip designer, they can literally shed ARM 32-bit compatibility, making the processor a little lighter (not as hot). Already, the A10 is close to passing Xeon. A few more tweaks and Intel will be fighting to catch up.
And you can run a Windows 10 on Raspberry Pi, not that it is going to give you access to all the great Windows apps. But MS may get stuck on Intel where performance will not be improving as much. The closed Intel ecosystem could well fold under the weight of Intel trying to keep it going on their own, competing with dozens of ARM licensees.
Or, perhaps, they have some sort of hybridization scheme that would make the transition all but seamless.
Apple burned 32-bit apps to the ground with iOS 11. Lost some old stuff, some of it pretty good but neglected. So now, as their own chip designer, they can literally shed ARM 32-bit compatibility, making the processor a little lighter (not as hot). Already, the A10 is close to passing Xeon. A few more tweaks and Intel will be fighting to catch up.
And you can run a Windows 10 on Raspberry Pi, not that it is going to give you access to all the great Windows apps. But MS may get stuck on Intel where performance will not be improving as much. The closed Intel ecosystem could well fold under the weight of Intel trying to keep it going on their own, competing with dozens of ARM licensees.
Or, perhaps, they have some sort of hybridization scheme that would make the transition all but seamless.
And also tomorrow, since ARM CPUs will never gain the same prevalence in the areas were especially BSD and Linux are important.
It’s not as if Apple is going to suddenly create chips as powerful as Intel’s without running into similar realities. I don’t really think this will solve more problems than it creates.
What is gained? Apple making their own chips and having total platform control.
What is lost? Virtualization. Binary compatibility.
Is Apple so arrogant as to believe their chips can outpace Intel’s? I’m continuously confused by this rumor. And I doubt anyone is dying to deal with more fat binaries and more Rosetta Stone and more waiting for major software vendors to update everything (and be forced to upgrade software that already works).
If the Mac was still Apple’s flagship product, I’d see a reason. It’s not though.
What is gained? Apple making their own chips and having total platform control.
What is lost? Virtualization. Binary compatibility.
Is Apple so arrogant as to believe their chips can outpace Intel’s? I’m continuously confused by this rumor. And I doubt anyone is dying to deal with more fat binaries and more Rosetta Stone and more waiting for major software vendors to update everything (and be forced to upgrade software that already works).
If the Mac was still Apple’s flagship product, I’d see a reason. It’s not though.
While discussing we need to consider that Apple has four different customer segments:
- Consumers
- Education
- Professionals
- Enterprise
Most consumers don't care about what's under the hood or technical specs. To them cost, ease of use and "just works" is most important. Apple's profits are high because of economies of scale though the use of the same hardware inside the iPad, iPod touch and iPhone. Adding Macs to that this will drastically boost Apple's margin.
Further more, designing own chips allows greater control of power management, optimise, graphics, hardware acceleration etc. Think of the advantages gaming consoles have over gaming PCs: Better optimisation for intended usage, smaller foot print and lower cost.
Fear not though. Intel based MacBook Pros and workstations will be the last to be phased out, if not at all. Power users and enterprise customers have nothing to worry about in the medium term.
More than ten years ago, most phone and PDA manufacturers were running Windows CE. Apple didn't go down the route of putting a desktop style OS on mobile device. The barriers to succeeding with own-chip powered computers is lower.
I suspect there are a lot of people out there in the same boat. I keep a Windows partition on my Mac so I switch over and play any of the hundreds of Windows-only games out there, but I know others have reasons that are far more important or critical to their work.
Hopefully Apple will think to include some way to continue running Windows on the Mac. I wouldn't bet on it, but here's hoping.
Windows compatible Macs won't end production overnight. Apple compatible Macs will still remain available for atleast another 7-10 years.
This will be a lengthy transition and there will be choice.
I am not seeing any ARM references in the article other than from a previous rumor. In my opinion, it is just as likely that they have struck a deal with AMD to use their chips perhaps fabbed by someone else or with a bunch of tweaks like in the Xbox One / PS4.
If they really are taking on the challenge of moving to another complete architecture then they are probably the only company in the world that could make that work as they control the OS as well as the hardware. If they do manage to create a CPU / App combo which is significantly faster than an Intel then that'd be great news for the consumer as for the first time in a long time you'd have 3 competitors in the chip market, if nothing else it should motivate Intel.
Until we know what really is happening though I am most certainly not going to put off buying another Mac if needed, just waiting for the 8th gen Intel quads to end up in the 13 MBP then we probably have 18 months to 2 years before something launches with a non Intel chip (if it happens) then several years more support for the architecture after that.
If they really are taking on the challenge of moving to another complete architecture then they are probably the only company in the world that could make that work as they control the OS as well as the hardware. If they do manage to create a CPU / App combo which is significantly faster than an Intel then that'd be great news for the consumer as for the first time in a long time you'd have 3 competitors in the chip market, if nothing else it should motivate Intel.
Until we know what really is happening though I am most certainly not going to put off buying another Mac if needed, just waiting for the 8th gen Intel quads to end up in the 13 MBP then we probably have 18 months to 2 years before something launches with a non Intel chip (if it happens) then several years more support for the architecture after that.
I'll guess this will make your iMac into a giant iPad. Good luck with running parallels on a custom cpu.
Wonder if we'll see macOS adopt iOS features like not being able to downgrade the OS.
Doesn't anyone think it strange that, this being a site with "Mac" in it's name, there are 400-odd comments about how people are going to miss Microsoft Windows?
Some of you seem to remember the PPC days. Some were too young.
They were good days. And back then, Apple didn’t have the money they have now.
Apple have tremendous cash reserves needing to be spent to avoid having large chunks given away to the government.
Can they strike a deal with AMD? Yep.
Can they buy AMD or some other Intel competitor? Yep.
Can they hire the top talent away from Intel? Also yes.
Sure, this is a dangerous move. But stagnation is death. And frankly, for the first time my 4-year-old computer is still basically top of the line. I don’t need a new Mac for years. I used to buy one every 2-3 years.
This makes sense for Apple for these reasons:
1. “Spend it or lose it” cash.
2. Ability to compete with anyone on Earth (this was not true when the G3/G4 chips were out - though I will say, the G4 was a GREAT chip.)
3. Ability to drive greater shareholder value with higher profit margins and/or increased sales.
4. Greater flexibility with software and hardware.
5. A much larger user base that forces everyone to port their apps.
6. You never know what they will come up with in terms of highly customized backwards compatibility and emulation. Remember: billions and billions and billions of dollars...
There are probably a ton of reasons they know about but we do not based on the roadmap.
So I’ll close with this: those of us who did our work on PPC Macs do not have the same memory of those days that some here seem to project. Sure, there were problems. But I loved my PowerBook G3, my iBook clamshell, my eMac G4, my PowerMac G3, and my Titanium PowerBook with a (then) ultra-fast G4.
I’m not denying the reality that there are drawbacks and dangers ahead. But there’s a glorious uniqueness to Apple hardware when Apple has more control. It’s hard to explain it. But it’s very tangible...
They were good days. And back then, Apple didn’t have the money they have now.
Apple have tremendous cash reserves needing to be spent to avoid having large chunks given away to the government.
Can they strike a deal with AMD? Yep.
Can they buy AMD or some other Intel competitor? Yep.
Can they hire the top talent away from Intel? Also yes.
Sure, this is a dangerous move. But stagnation is death. And frankly, for the first time my 4-year-old computer is still basically top of the line. I don’t need a new Mac for years. I used to buy one every 2-3 years.
This makes sense for Apple for these reasons:
1. “Spend it or lose it” cash.
2. Ability to compete with anyone on Earth (this was not true when the G3/G4 chips were out - though I will say, the G4 was a GREAT chip.)
3. Ability to drive greater shareholder value with higher profit margins and/or increased sales.
4. Greater flexibility with software and hardware.
5. A much larger user base that forces everyone to port their apps.
6. You never know what they will come up with in terms of highly customized backwards compatibility and emulation. Remember: billions and billions and billions of dollars...
There are probably a ton of reasons they know about but we do not based on the roadmap.
So I’ll close with this: those of us who did our work on PPC Macs do not have the same memory of those days that some here seem to project. Sure, there were problems. But I loved my PowerBook G3, my iBook clamshell, my eMac G4, my PowerMac G3, and my Titanium PowerBook with a (then) ultra-fast G4.
I’m not denying the reality that there are drawbacks and dangers ahead. But there’s a glorious uniqueness to Apple hardware when Apple has more control. It’s hard to explain it. But it’s very tangible...
Not really.
I spend 90% of my time in OSX, but I still need to run software which only works in Windows.
One of the reasons I got my Mac was because I liked that for when I needed it, I could still run the other software.
If I had a choice between running OSX exclusively or Windows exclusively, it would have to be Windows.
I remember PPC chips that were supposedly faster but were not, that ran so hot they required a water-cooled system that leaked and burned out the computer.
This is a bad, bad move if it is true. I don't want more integration with my phone. I want a good computer.
This is a bad, bad move if it is true. I don't want more integration with my phone. I want a good computer.
Well that speaks volumes about the state of Mac software in general doesn't it? That you would give up the 90% of your time in OSX for the 10% you spend in Windows as you can't do/work without it but you could ditch OSX/macOS.
Yes I saw it. Intel and MS know it’s over for the Windows era, but some consumers don’t seem to get it.
There was also a rumor that Apple would use an Intel and an Apple-chip together, the extra A10 chip would add the cost with $25 translating to $100 retail. Not that bad considering it would almost double the performance and allow for a low-power mode next to a full power mode.
Adobe wont port their iOS apps to desktop. Adobe already deliver a bare minimum to Mac users. Their apps clash with macOS. They release bug ridden dross despite months of macOS beta availability. What I’m saying is: Adobe can’t get any worse.
Also intel haven’t released a MacBook Pro chip worthwhile since 2012 retina MacBook came out. That is why this is good! Now I’ll have an 8 year old computer by the time I upgrade.
I’d rather a great Mac than an average Mac that can run windows.
I can't say I really understand the problem here. As someone who doesn't know a great deal about computers, it seems to me that the pros are:
- Better integration with iPhone (Apple's best selling product), iPad etc.
- Much improved battery life
- Better control of their products
- More reliable upgrade of chips
The cons seem to be:
- Third parties will need to re-write their apps, but will be in the same language as iOS
- Macs won't be able to run Windows anymore, but that must be an issue for a very small number of users who could just get a Windows machine in addition to the Mac.
I don't see this as a surprise at all. Apple are all about going it alone and have no interest in integrating their products with anyone else's (just look at the removal of the h/phone jack, the fact iMessage doesn't work on non-Apple products, the lighning connector etc.).
- Better integration with iPhone (Apple's best selling product), iPad etc.
- Much improved battery life
- Better control of their products
- More reliable upgrade of chips
The cons seem to be:
- Third parties will need to re-write their apps, but will be in the same language as iOS
- Macs won't be able to run Windows anymore, but that must be an issue for a very small number of users who could just get a Windows machine in addition to the Mac.
I don't see this as a surprise at all. Apple are all about going it alone and have no interest in integrating their products with anyone else's (just look at the removal of the h/phone jack, the fact iMessage doesn't work on non-Apple products, the lighning connector etc.).
Whereas for it is means the end of using Macs - no longer able to support the software (not "apps") that I need to run my business...
Last edited: