With the iPhone being over 80% of annual revenue, it's not the suicidal scenario it was in ~2007.
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 Interested in Developing ARM-Based Mac Processors and iPhone Modems in House
- Thread starter MacRumors
- Start date
- Sort by reaction score
What makes you think this would be suicide? Computers still need software, but there are plenty of ways to remedy that throughout this thread.
It’s not that they are. Geekbench isn’t a particularly good indicator of how the A11 would perform if dropped into a MacBook Pro in place of th current CPU. Geekbench is influenced by too many other platform factors, even across more similar devices.
Take a look at those benchmarks again. Does anyone seriously think that the A11 in the iPhone 8 (and 8 Plus) is faster, both in single core and multi core performance, than the A11 in the iPhone X? Because that’s what GeekBench would lead you to believe.
Now consider all of the other differences between an iPhone and a MacBook Pro. Place that in light of the comments posted above about Apple using complementary ARM chips.
Apple could move to ARM in Macs. Sure. The hardware is probably “fast enough” and, for basic consumer use, the software library is probably adequate. And maybe Apple has other strategic business reasons that might motivate such a change, probably not in the next couple years, but maybe five or more years out. But right now, certainly, and I suspect going forward, performance of its
ARM architecture, and nothing more, certainly isn’t a reason to go that route. It isn’t there the way some in this thread seem to think it is.
Now, unloading a bunch of CPU hungry secondary and background tasks to a co-processor specially designed for the purpose (welcome back to the ‘80s!) is another story. That could make all kinds of sense, allowing Apple to lard up its Macs with a bunch of other hardware and software features that would otherwise bog down the main CPU. A CPU that will likely remain designed and fanned by Intel for some time.
This reminds me of a rumor that was floating around a few months back, where supposedly Apple was working on putting a chip in the Mac to do just this. It sounded like it would be used for background tasks while the laptop sleeps (I guess Power Nap 2.0?) like fetching email and such, but it seems like a good start.
Why would Apple bother making their own CPU for MacOS systems? That is not where the money is.
Exactly. Apple is playing a long game, and a lot of the worries expressed in this thread are premature.
ARM's specialty is low power consumption. It's design is perfect for mobile and small devices. Apple proved with the Air that they want a mobile computer and moving to ARM will allow them to do that, removing the current heat issues that Air's have with using Intel chips. Using an ARM chip will reduce heat, reduce power usage and extend battery usage. You put an ARM chip in one of the latest macbook Air's, the battery wont last for a few hours, it will last a couple of days and that will be it's selling point.
I remember back in the days of powerpc, intel and AMD. The huge majority of people did not pick their pc based on the hardware spec's, they based it on what operating system they liked, price and after care support. The same will happen if Apple moved everything to ARM, you would then have ARM, Intel or AMD with buyers making the same decisions as they did back when it was powerpc days.
I remember back in the days of powerpc, intel and AMD. The huge majority of people did not pick their pc based on the hardware spec's, they based it on what operating system they liked, price and after care support. The same will happen if Apple moved everything to ARM, you would then have ARM, Intel or AMD with buyers making the same decisions as they did back when it was powerpc days.
Inevitable!! The iOS platform is so popular and the home-made chips so powerful and purposeful that Apple could design and run better computers by making their own architectures tailored to specific computer usage. Outside companies are too inefficient for Apple’s needs.
PIA-side: all software needs to be revamped.
I’d suspect 2020-onward begins a serious change in Apple computing.
PIA-side: all software needs to be revamped.
I’d suspect 2020-onward begins a serious change in Apple computing.
Your timeline has a couple problems. First, nobody ever said Macs couldn't make a transition. Nobody relevant anyway.
Second, the transition *to* x86 brought a specific benefit other transitions didn't, and the transition *away* from it removes that benefit.
Also legally might be a problem for Apple as well. because Arm will want licensing fee's of course.
They are still producing hardware for anything with an A10 CPU and older but Apple are producing A11 and onwards in house. They are no longer designing and manufacturing new graphics architecture as of earlier this year.
Apple has an ARM Architecture license, the least expensive kind there is, I think. That allows them to run off and design their own ARM ISA compatible processors, which is exactly what they've been doing for years. Qualcomm also has an architecture license. There's also nothing to stop Apple from extending their silicon architectures in any direction they please, so the ARM relationship is nothing at all like the Intel relationship.
I believe either Microsoft or Intel bought the last few companies with that patent portfolio. I know Microsoft bought the former PPC companies that were formerly selling x86 on PPC macs just after the Intel switch. They used that in the back and forth architecture changes in XBox land. I know Intel bought some other ones just to keep people from leaving their hardware.
How would ARM present a legal problem for Apple to support x86. x86 is Intel IP. The legal impediments to supporting x86 ISA would come from Intel. Emulation would not be a problem, as Apple already did it on PowerPC. Emulation also isn't necessarily the performance hindrance many claim. Microsoft is emulating x86 on ARM in their upcoming Win10 port to Qualcomm chips.
The thing that makes it inevitable that Apple will use their own chips in all their devices is the fact that they need to have more control over the development and supply chain. The issues that constrain them today on the laptop and desktop side, with Intel seemingly delaying each major release of their chips, makes it very difficult for Apple, and every other PC manufacturer, to roll out product on time.
Of course Apple also wants to be able to bring the benefits of what they've paid for, in designing mobile chips, to the rest of their line which reduces their costs even if initially there will be a lot of up-front costs to make the initial transition.
Those people worrying, or worse complaining, that their needs won't be met because the current A11 chip isn't powerful enough, are clearly not recognizing that Apple isn't just going to plop a mobile A11 chip into a MacBook Pro and call it done. Who knows what Apple will do, but it could include running 4, 6, 8 or more A chips in a single rMBP or iMac, creating a supercomputer of sorts. Or they could just make a single chip that is far more powerful, with more bandwidth to deal with typical desktop or laptop computing needs.
I'm excited to see what the future brings, especially with Apple having far more resources to be able to execute on their vision - something that Steve Jobs was only every able to partially do because they didn't have the skills or resources they do now.
Of course Apple also wants to be able to bring the benefits of what they've paid for, in designing mobile chips, to the rest of their line which reduces their costs even if initially there will be a lot of up-front costs to make the initial transition.
Those people worrying, or worse complaining, that their needs won't be met because the current A11 chip isn't powerful enough, are clearly not recognizing that Apple isn't just going to plop a mobile A11 chip into a MacBook Pro and call it done. Who knows what Apple will do, but it could include running 4, 6, 8 or more A chips in a single rMBP or iMac, creating a supercomputer of sorts. Or they could just make a single chip that is far more powerful, with more bandwidth to deal with typical desktop or laptop computing needs.
I'm excited to see what the future brings, especially with Apple having far more resources to be able to execute on their vision - something that Steve Jobs was only every able to partially do because they didn't have the skills or resources they do now.
Funny you should mention that.
I was in a meeting where the President of a multi-billion dollar real estate development company said he did 100% of his work from an iPad Air. But I guess running a multi-billion dollar real estate development company isn't "real work".
Why would Apple continue to support two underlying CPU architectures for their iOS and macOS product lines? They're already making their own ARM processors, so there is NO bother.
It it’s “in the ether” that Apple is “talking about” it.. then it’s for sure going on. Apple was quietly porting Next/Darwin kernel to x86 for several years before there was any talk of Mac on Intel. MacOS is decended from iOS now... there was a big flip a few years back when they redid all the iWork apps... behind the scenes they refactored all their code for a common platform and re-added features for a few years to cover the change.. but now they’re all iOS apps first, macOS apps second. The only main app not redone is Xcode. And with Swift Playground, and the push to Swift code over old Objective-C and aggressive depreciation of lots of APIs they clearly have it working, just not ready yet.
Correct. Apple has a first class chip design team in-house. Why on Earth wouldn't they leverage that to the hilt? Every dollar Apple sends to Intel benefits Apple competitors in some way. Ever dollar Apple spends on its in-house team benefits Apple only.
I disagree, I think they can emulate x86 on ARM with a performance penalty and still get away with it. It would cause most SW devs to start cross compiling for ARM natively and x86 dependence won't be necessary. The bigger issue is ARM performance itself, the A11 bionic chip may benchmark well, but for real workloads, x86 processors are still significantly ahead. Most people don't realize that ISA and microarch are completely 2 different things. Both x86 and Apples implementation of ARM use specialized VLIW microarch, are heavily pipeline, and use various HW accelerators. With this regard, x86 is a much more mature platform.
Huh. There are far more ARM focused (iOS) developers today than there are x68 developers.....
the last time a company dictated what chips apple could use , look what happened
motorola anyone ? the g4 screwup...
sj would want tc to go 100% apple
it's all about control..
[doublepost=1506703462][/doublepost]apple don't need intel
intel need apple
motorola anyone ? the g4 screwup...
sj would want tc to go 100% apple
it's all about control..
[doublepost=1506703462][/doublepost]apple don't need intel
intel need apple
I'd be happy with just a couple of bionic arms.....
On topic, such a move by Apple will initially be laborious and costly, but I'm sure they've come to the conclusion that controlling their own destiny going forward, and specifically within the exciting new AI arena, far outweighs the not inconsiderable negatives of in-house designing and fabbing.