Intel CEO Responds to Rumors of ARM-Based Macs, Says Relationship With Apple Is 'Strong'

[junkstory]

Apple already has an OS X port for ARM, remember, they based ios off of it. Similarly, Apple already had an OS X port for Intel, when they were on PPC.

For GPUs, Apple can still partner with Nvidia or AMD. Not that many native games are available for OS X, anyway, most just use a wrapper based off of wine. But gaming, was never Apple's target market, in any case. Apple has a lot more power with devs now thanks to the Mac App Store, and Adobe is moving into the cloud anyway with CC, so they can still develop some type of interface for Mac to work with it, from a front-end perspective.

What would be Apple's business case for making an ARM-based Mac? People aren't going to flock into Apple stores wanting to buy one just because the chip architecture is different. If anything, there would be less demand as less apps/games would be supported initially.

The only reason to go ARM would be for cost and battery life. Cost isn't even an issue as an average Mac is $1500+.. the cost of the chip is insignificant. Battery life is already superb. The current Haswell MacBook Air can get 12 hours of battery life, I'd imagine the Broadwell can get 14 hours, and Skylake could probably stretch to 15+ hours. Beyond Skylake, the bottleneck of battery life would be from other components like the screen, SSD, and WiFi.

 

[LordVic]

What would be Apple's business case for making an ARM-based Mac? People aren't going to flock into Apple stores wanting to buy one just because the chip architecture is different. If anything, there would be less demand as less apps/games would be supported initially.

The only reason to go ARM would be for cost and battery life. Cost isn't even an issue as an average Mac is $1500+.. the cost of the chip is insignificant. Battery life is already superb. The current Haswell MacBook Air can get 12 hours of battery life, I'd imagine the Broadwell can get 14 hours, and Skylake could probably stretch to 15+ hours.

The business case is Gross margin and complete control over their own pipeline products.

removing intel would have both those benefits to Apple

Right now Apple is at the mercy of Intel for the parts. This means they can only produce computers with the CPU featureset intel provides at the costs of intel chips.

moving it internal design and Arm they could control the exact design and elements in the chip, featureset, at their own whims and not intels.

they could also reduce the costs of their CPU's to improve their overall profit margins. Likelihood from a consumer end user side, we would never see those cost savings.

 

[b34g]

What would be Apple's business case for making an ARM-based Mac? People aren't going to flock into Apple stores wanting to buy one just because the chip architecture is different. If anything, there would be less demand as less apps/games would be supported initially.

The only reason to go ARM would be for cost and battery life. Cost isn't even an issue as an average Mac is $1500+.. the cost of the chip is insignificant. Battery life is already superb. The current Haswell MacBook Air can get 12 hours of battery life, I'd imagine the Broadwell can get 14 hours, and Skylake could probably stretch to 15+ hours. Beyond Skylake, the bottleneck of battery life would be from other components like the screen, SSD, and WiFi.

Apple's business case for switching to intel from powerpc was performance-per-watt, it would be a similar case here. The 11 inch air is still officially 9 hours, which could be better.

 

[alent1234]

Apple's business case for switching to intel from powerpc was performance-per-watt, it would be a similar case here. The 11 inch air is still officially 9 hours, which could be better.

except intel was faster than powerPC. ARM CPU's are slower than Intel CPU's.

the case would be to capture the low end market and make a profit. the MBA has a crappier screen than the MBP so you can build a Mac around cheaper parts and sell it at a profit in the $500 range with keyboard, mouse and monitor.

 

[Michael Goff]

except intel was faster than powerPC. ARM CPU's are slower than Intel CPU's.

the case would be to capture the low end market and make a profit. the MBA has a crappier screen than the MBP so you can build a Mac around cheaper parts and sell it at a profit in the $500 range with keyboard, mouse and monitor.

Not just CPU, but GPU as well.

 

[b34g]

except intel was faster than powerPC. ARM CPU's are slower than Intel CPU's.

the case would be to capture the low end market and make a profit. the MBA has a crappier screen than the MBP so you can build a Mac around cheaper parts and sell it at a profit in the $500 range with keyboard, mouse and monitor.

That is comparing current ARM mobile processing power to intel desktops, it is easier to scale up ARM than it is for intel to scale down its processors. It won't be faster at the high-end for a long time indeed (but then, apple doesn't compete at the very high-end in terms of pure specs), however, it is fast approaching, that ARM may became decent enough for the middle-end.

 

[junkstory]

the case would be to capture the low end market and make a profit. the MBA has a crappier screen than the MBP so you can build a Mac around cheaper parts and sell it at a profit in the $500 range with keyboard, mouse and monitor.

Intel has the low-end of the market covered pretty well. An Intel Atom Z3735F could be had for $17/chip, and it can run Windows 8.1 decently. A similarly spec'd ARM chip may be just a few dollars less. Is it worth re-engineering an operating system to save a few dollars per machine?

 

[Nightarchaon]

I agree with you that i can see them moving to an App Store only system, same as iDevices. Personally, i've been slowly growing to love that idea. Not having to keep track of where stuff is installed, organizing files, all that mess... much easier on an iDevice. But when i think this, i remember a guy i knew decades ago telling me how much he hated Windows, that he preferred DOS... he thought that having a command line interface gave him so much more control. Change is hard for most people.

Sadly , even now on OSx and Windows 8.1 that old command prompt / terminal windows is where 70% of my job has me living, whilst we have iDevices in the company, i dont see us every being locked down to App Store only devices, purely because we need to run such a diverse series of applications (our oldest application is still DOS based and still interfaces with an old Sun Solaris maniframe, it deals with about 1.4 million clients details and there is no sign of it getting updated , because "it works" and is bullet proof)

an ARM/App Store based apple mac would mean all the macs in our organisation would have to go, no Wine integration for applications means no sale for us. (not to mention the few machines that need parallels to allow full windows / mac OSx side by side integration)

If apple went Mac only, our current Mac offering would have to be abandoned as we would not push it out via the appstore. it fails on too many of apples draconian requirments

 

[joema2]

What would be Apple's business case for making an ARM-based Mac?...If anything, there would be less demand as less apps/games would be supported initially...

There is no good immediate business case that warrants the tremendous risk of a hobbled machine with poor application compatibility. If it had 50% better battery life users wouldn't care if x86 was good enough (and getting better with each version). If it cost users software compatibility of their favorite apps, they just wouldn't buy the ARM Mac.

It would also splinter the Mac product line between x86 and ARM.

Furthermore it would cost Windows compatibility. No more boot camp, no more running Windows apps (efficiently) via Parallels, etc.

This was discussed last year at some length, and the same reasons still apply: http://www.extremetech.com/computin...its-own-arm-chips-in-macbooks-and-the-mac-pro

Now, if the plan is to make *dual* CPU Macs with both x86 and ARM, and which can natively run both iOS and x86 apps, that would be compelling. However it is very challenging technically.

If Apple thinks ARM can do better long term than Intel as a desktop/laptop CPU, that would be a reason. However that is a big guess. Intel spends $11 *billion* per year on new fabs and products.

Even though the above Extremetech article says there's no major difference in ARM vs 86 power efficiency at the same performance level, there is informed academic CPU research which holds the *right* non-x86 architecture can significantly out-perform Intel at the same feature size. Proof of this are VLIW-based DSPs. In theory a general-purpose CPU could be designed using similar architectural methods.

Intel desktop CPU performance is limited by three fundamental parameters: (1) Clock speed/heat, with no significant improvement in sight (2) They are roughly at the superscalar execution limit (8 way) with no further major improvement possible (3) Multiple cores are limited by software, Amdahl's Law, and increasingly by heat.

Is Apple looking that far ahead? iOS integration may be a closer issue, but Apple has long stated iOS and OS X will be forever separate.

Another possibility is it's just a feint to to leverage negotiation with Intel.

 

[DudeDad]

Deja vu:

"What, Apple is going to Intel and leaving the PPC behind. It's the end for Apple"

Yeah, deja vu.

 

[gnasher729]

Games tend to be the ultimate test though. If all software can be "simply recompiled" why isn't every game available on Windows available on Macs?

It's a simple recompile if you stay on the same operating system. Apple could build an iPad with iOS using an Intel processor, and all iOS software would run on it with a simple recompile. Apple could build a Mac with an ARM processor, running MacOS X, and it would be a simple recompile. Making iOS ARM software run on that Mac would _not_ be a simple recompile, and the same is true with Windows software.

 

[Badagri]

Deja vu:

"What, Apple is going to Intel and leaving the PPC behind. It's the end for Apple"

Yeah, deja vu.

 

[gnasher729]

How similar or dissimilar the CPUs are is not important. The compiler guarantees the program generated will behave exactly the same regardless.

Intel processors, both 32 and 64 bit, are very similar to ARM processors (32 and 64 bit) in the way that they behave, even though the instruction set is totally different. PowerPC processors are big-endian while Intel/ARM are little-endian; that requires a competent programmer to run with a simple recompile; Intel and ARM are so similar that you need a programmer who isn't badly incompetent.

 

[Count Blah]

What would be Apple's business case for making an ARM-based Mac? People aren't going to flock into Apple stores wanting to buy one just because the chip architecture is different. If anything, there would be less demand as less apps/games would be supported initially.

The only reason to go ARM would be for cost and battery life. Cost isn't even an issue as an average Mac is $1500+.. the cost of the chip is insignificant. Battery life is already superb. The current Haswell MacBook Air can get 12 hours of battery life, I'd imagine the Broadwell can get 14 hours, and Skylake could probably stretch to 15+ hours. Beyond Skylake, the bottleneck of battery life would be from other components like the screen, SSD, and WiFi.

They are trying to maximize profits by fleecing their bags-O-cash customers for RAM an storage, now that it's all glued in. So what are the other components they can use to bolster the bottom line? They don't make monitors, motherboards, nor power supplies, so CPU it is. Keeping more of the $$$ in house is a HUGE incentive for Apple.

If they can make the move palatable to enough people to make $1 more, they'll do it. I can only hope that those who laughed and ridiculed me for being so angry with Apple's "all-glued-in/anti-consumer" approach, are those hardest hit by this. Sucks when Apple does something you REALLY don't like, doesn't it?

 

[firewood]

Intel has the low-end of the market covered pretty well. An Intel Atom Z3735F could be had for $17/chip, and it can run Windows 8.1 decently. A similarly spec'd ARM chip may be just a few dollars less. Is it worth re-engineering an operating system to save a few dollars per machine?

Apple doesn't buy Atoms. They buy dual core i5's for their low end MBA11's, which cost them significantly more than $17.

An ARM Mac could allow them to price consumer and educational market MacBook Airs a bit more aggressively, with equal or better profit margins, while offering customers even better battery life. Little OS re-engineering will be required, as they have already ported most of OS X to ARM as part of their internal iOS development process.

The Mac Pros (etc.) will keep the Intel room-heating CPUs and fans.

 

[firewood]

Another possibility is it's just a feint to to leverage negotiation with Intel.

That's a very strong possibility. Apple could spend a good fraction of a Billion $$$ taping out ARM processors for MacBooks, and end-up saving even more money by forcing Intel to offer them better deals to keep Apple's business.

Or if Intel reacts too slowly, edu Mac buyers and consumers suddenly get a new product with even better battery life.

 

[djgamble]

I still don't see why there would be an ARM-based Mac... it would perform like a 10 year old Intel Mac and have serious issues trying to run any Intel-based apps even if the software to do so existed.

I've been proven wrong many times before... but I think there's other chips that would be considered for cheap, low-power but full-featured devices. That is, unless ARM chips are moving at a faster pace than Intel chips, and are expected to overtake them one day. Well then... maybe it's one of those 'in the future' explorations.

To be fair, that's what Intel was... IBM made faster chips, but Apple kept developing an Intel-based OS X. One day we saw the latest PPC technology in the Xbox 360, followed by an announcement that Apple was dumping PPC chips a few months later because Intel had a better roadmap. Strange things happen!

 

[Michael Goff]

Apple doesn't buy Atoms. They buy dual core i5's for their low end MBA11's, which cost them significantly more than $17.

An ARM Mac could allow them to price consumer and educational market MacBook Airs a bit more aggressively, with equal or better profit margins, while offering customers even better battery life. Little OS re-engineering will be required, as they have already ported most of OS X to ARM as part of their internal iOS development process.

The Mac Pros (etc.) will keep the Intel room-heating CPUs and fans.

Room heating? Have you used a modern, within the last three years, processor?

 

[firewood]

I still don't see why there would be an ARM-based Mac... it would perform like a 10 year old Intel Mac...

A single core of last year's A8x beats a single core of my 2010 MacBook Pro on several benchmarks, even though designed for a tiny heatsink. A newer arm64 chip designed for a bigger heatsink and faster memory would greatly narrow that 4 year performance gap.

 

[Tech198]

Here we go...

- Better power,
- Ultra thin Macbook 12-inch approaching (maybe)

You mac will be like a Mac Pro, all sorts of devices hung of it since APple will migrate to one "easy" port.... (with an array of Y connectors)... ARM based also means better CPU, but Apple now controls what OS u can put on...... (more limiting on the horizon.)


I knew it wouldn't be long before they did this on desktop.

Messy stuff.

How can Apple turn something nice into something ugly ?


Here i come "other world"

 

[mattoligy]

Apple could have done this years ago on Intel chips. They could make it available now on Intel Macs. There's an emulator, no simulator that already runs iOS software on OS X for the programmers coding iOS apps. That could be repurposed, polished up for general use, and there it is. No ARM chip required.

OR, all iOS programmers could "just recompile" their iOS apps to run on OS X. Just ask Nutjob. Apparently so many of them don't want the added revenue or are too lazy to "just recompile" their iOS apps for OS X.

And I just can't get how this vision of ARM-based Airs coexisting with Intel-based Macs (otherwise) is not an explicit shot of the dreaded "fragmentation". There would almost certainly be OS X software that would not run on the ARM-based machines. It would be "For OS X" but "on Intel-based Macs Only". Maybe vice-versa too? Can we actually picture Apple doing that?

And, as someone astutely asked several pages back, how do we end up with Apple pricing a 128GB Air at $500 vs. a 128GB iPad priced at $700?

All I see in this concept is corporate benefits of greater profitability, better product development timeline control and a little more "big reveal" surprise. I am not seeing the consumer benefits. If it's about cheaper Macs, I don't know how cheaper is compatible with fattening margins. As others have offered, Intel makes cheaper chips than those Apple choses to use now. If it's about running iOS software on Airs, maybe there's a little something there? But why not repurpose the emulator/simulator? Or "just recompile" those apps? Or buy a keyboard case for an iPad? After that is just falls off for me.

Losing the ready availability to use Windows when needed- or sometimes necessary- is a massive negative. Even if "we" can all get behind the switch, as soon as a Mac can't run Windows, those that must run Windows software or connect to Windows networks have little choice but ALSO buying a Windows machine. If you have 2 laptops, are you really going to take BOTH with you when you'll have to interface with Windows in some way as part of the trip? Probably not. Other than corporate benefits for Apple, I just don't see it.

Very good post, made a good read. However you did miss one point... Desktop class applications like photoshop finding there way to (hopefully a larger) iPad down the line due to big software companies having created ARM ports of their apps for a new mobile MacBook! If developers are pushed to port their powerful desktop class apps to ARM for a mobile MacBook line, why not release on iPad (pro) also for added revenue!? Even if they were not touch ready for input all we would need is Bluetooth mouse functionality added to a pro iPad line and with the already present Bluetooth keyboard support an iPad pro would be a worthy laptop replacement no? Especially as these A chips get more powerful down the line... The A8X was impressive, who knows what the A9X will show us...

 

[HobeSoundDarryl]

It's a simple recompile if you stay on the same operating system. Apple could build a Mac with an ARM processor, running MacOS X, and it would be a simple recompile.

I'd like to see this proven. While I can imagine Apple putting in the time behind the scenes so that OS X on ARM would run (including OS X's many stock programs), I'd like to see non-Apple OS X software run from that first day or first week... stuff like the Adobe Suite, MS Office, Pixelmator, etc down to smaller, popular apps like Skype, Handbrake and so on. Take the "Top 100 Third Party apps" for OS X and you & others make it sound like all 100 should run with a "simple recompile"... like expectations for all 100 should be met within a week or so. I just don't think so.

Jump back to the last time Apple hopped from one CPU type to another and it was a long wait for Adobe Suite, MS Office, etc. We got by on the PowerPC versions while we waited because of Rosetta; but it was a long time for Intel versions of either of those to show up. Both were running on OSX before and after the switch to Intel. Why didn't they just "simply recompile" those programs for Intel Macs?

I can imagine versions of Calendar, Mail, Calculator, etc being quickly converted. I just don't see third party stuff being so easy. Nevertheless, if Apple goes there, I'll hope you are right... because the alternative is giving up on Apple and going back to Windows rather than waiting & hoping for months or years.

 

[Tech198]

Well, based on the fact u can do the same thing with programming code, and not all of it turns out ok, it would be a challenge...

Not saying it would work. but i'm not saying it wouldn't either. It just may not be 100%

 

[katewes]

I think Apple needs to get this clear, that if Mac goes to ARM, a big bunch of us are out of here. Apple trait is to trash its users, e.g. discarding matte screens needed by professionals, dumbing down pro-features in its software. Many business users need Intel compatibility to run business software.

I want to debunk the usual reasons for going to ARM:

-- Intel's slow development cycle: I don't care. I buy Macs on average once every 2-3 years, so I am in no rush to get the latest processor. All processors are fast enough for what I want to do. So it's meaningless to me for Apple to give away Intel compatibility, just for their self-interest to control the processor. We live in a wider community where users have to interact in a PC-driven world, and that requires compatibility. Apple needs to get that, in no uncertain terms.

-- ARM's lower power-consumption. True, but that, in no way, trumps the above need for Intel-compatibility. Nowadays the battery can last for an entire work-day, so ARM's power-saving features are not a game-changer that, in itself, warrants discarding Intel. I repeat, ARM's power-capacity is nice, but not sufficient to warrant a change from Intel to ARM.

 

[coolfactor]

I think the best approach would be a dual-CPU strategy, both Intel AND ARM in the same machine. Leverage each for their strengths. Detach the screen and the Intel-powered Mac becomes an ARM-powered iOS device. Re-attach and it switches back. iOS apps runnings alongside OS X apps without virtualization.

Apple would be able to achieve such a device and competitors wouldn't able to match it.