Regarding ARM-based Macs

Regarding ARM-based Macs

There's a story going around about Apple working on ARM-based Macs. In other words, Macs that, instead of Intel chipsets, use the same kind of Apple A7-style chipsets found in the iPhone and iPad. The story is nothing new. ARM-based Macs, touchscreen Macs, iOS Macs, Retina Macs... Apple prototypes pretty much anything and everything any reasonable person would expect them to. A thousand no's for every yes requires a very high prototype to product ratio, after all...

It's fun to speculate — would an ARM-based Mac run iOS or OS X? If the former, could it emulate the latter so it could run OS X apps? If the latter, would it be powerful enough to emulate Intel to run apps not ARM-ready at launch? How would Windows Bootcamp be maintained, if indeed its still a desirable feature to offer?

There's even more to consider than there was when Apple prototyped "Marklar", the Mac on Intel project that ultimately resulted in OS X 10.4.4 Tiger.

Not everything prototyped makes it into production, of course. For example, Apple has been vocal about not liking the ergonomics of touchscreen Macs. It should surprise no one that Apple is stating an informed opinion, one of experience and insight gained from first-hand testing and experimentation. That's how Apple's works. That's how they make great products.

Apple hasn't said anything publicly about the pros or cons of ARM-based Macs, but it's likewise not hard to remember the conditions that led to the Intel switch and imagine there'd need to be similarly strong reasons to warrant an ARM switch. Unless and until that happens, unless and until Apple shows an ARM-based Mac on stage, the what is far less interesting than the how or the when.

Have something to say about this story? Leave a comment! Need help with something else? Ask in our forums!

Rene Ritchie

EiC of iMore, EP of Mobile Nations, Apple analyst, co-host of Debug, Iterate, Vector, Review, and MacBreak Weekly podcasts. Cook, grappler, photon wrangler. Follow him on Twitter and Google+.

More Posts



← Previously

Apple adds two new iPad verses: Orchestrating sound, exploring without limits

Next up →

iMore store Memorial Day Sale! Save 20% on all iPhone and iPad accessories!

Reader comments

Regarding ARM-based Macs


I see no reason to do this. Current Intel chips are generations ahead of ARM in terms of raw computing power, and if this is a power consumption issue, the haswell and upcoming broadwell chipsets work exceptionally well in laptops.

Plus this would also cripple current Mac OS X app performance for years (remember the Photoshop debacle), and permanently kill bootcamp/parallels.

This simply makes no sense on any level.

If you're talking about ARM v7 chips like the A6 and older, you're right. ARM v8 is another story. Its designed specifically to be able to scale up to the desktop. The A7 is almost neck and neck with Intel's Bay Trail ATOM as far as performance goes. Apple is a good year ahead of everyone else with ARM v8 design and they wouldn't have to rely on Intel's schedules if they were able to design their own desktop-class CPUs.

That being said, there would be downsides to switching over now. Despite Apple's promotion of it, Thunderbolt is very much an Intel product. An Apple desktop OS running ARM would also be at least partly reliant on Window RT gaining traction. Sharing a processor code base with Windows helped the Mac expand and thrive. We wouldn't really want a situation we had with PowerPCs were it wasn't in the best interests of programmers to fully optimize for the hardware. Though that might be a tad bit overblown with all the mobile development for ARM.

Here's what Anand also doesn't realize (going by his "x86 myth busted"): Apple (and other ARM chip vendors) are "neck in neck" in Intel's Atom, while Intel is half a node ahead of ARM foundries (22nm vs 28nm) and basically an extra full node ahead in terms of improvements with Trigate.

So it's 1 and a half nodes ahead of ARM, yet it's still barely "neck in neck". Yet, Anand told us that "x86 myth is busted". With such a huge process advantage Atom should be wiping the floor with high-end ARM chips, but it doesn't, because the x86 bloat is still very much real.

Also Atom is only neck in neck on the CPU side with its vast process advantage. On the GPU side, despite this advantage, it's a full mobile GPU generation behind. Its "desktop class" IVB/7th gen GPU usually testing at HALF the performance of most ARM GPU's. I don't think it even has support for OpenGL ES 3.0 yet.

Not to mention Atom chips still cost 50-100 percent more than ARM chips with the same performance, and Intel has to lose $1 billion every QUARTER to subsidize Atom, so it makes it at least not look extremely unacceptable to OEMs.

A Haswell Intel Core i5 chip costs well over $200, basically an order of magnitude more than a high-end ARM chip, yet it's within 2.5-3x the performance of such an ARM chip. If Apple doubles up in CPU performance again (much easier to do it for ARM these days, while Intel's chips are growing by 5 percent in overall performance over the past few generations), it can come a lot closer to Intel's performance soon, and what most people will think is good enough, while still costing 10x less, which gives Apple the opportunity to price its laptops not only much more competitively, but also maintain a high profit margin, that's currently eaten by Intel. Many are already saying Nvidia's Denver will have roughly the performance of a mainstream SNB Core i5, which is within 30 percent of the performance of a mainstream Haswell Core i5.

Now, whether Apple will actually go this way remains to be seen. Perhaps they decide to go the way of turning iOS into a "notebook OS" instead, by allowing you to put multiple apps on the screen, and other such tweaks, that transition people from PC's to "iOS books" or whatever they will call them. If the market accepts that, they don't even have to bother to transition their Mac apps from x86 to ARM. But the decision has a lot less to do with the performance of ARM in the future, and more to do with strategy for how they want this to happen.

Personally, I also think turning iOS into a notebook OS as it "matures" as a disruptive technology that it is, is a more viable strategy. But it's important to realize that whichever way they go (and I do think they will do one of these two things) - Intel loses.

I think Apple is still on a Wait and See Approach. Yes it is not hard to double or even triple the performance of Apple A7 SoC, It has to do so at power / performance ratio better then Intel. While ARM manage to win at the mW and <5W level. Intel has been winning everything above that. No ARM chip is even as powerful as Intel Haswell at 10 - 15W. And this isn't going to get any better when 14nm Broadwell comes out.

The Gap is definitely shrinking. Since over 30% of BOM cost on Mac belongs to Intel. Intel is likely to lower the price to keep Mac / Apple on its platform for as long as possible.

And in the long term I think what Apple really want is Intel to manufacture SoC for them.

For Apple, I don't think it would be an either-or thing. Unlike Microsoft with Windows RT, Apple has a long history of managing compatibility over two processors: in 1995 or so, they switched from Motorola 68k to PowerPC; Classic MacOS for PowerPC had a compatibility layer (emulator? I don't think that's the right word...) for 68k binaries. Likewise, when OS X went to Intel from PowerPC, they had Rosetta, which ran PowerPC applications pretty much seamlessly. The transition from Intel 32bit to Intel 64bit was also basically invisible to most users (again unlike Microsoft's transition). There is always going to be some performance hit if the Apps are processor intensive, but Apple has made these transitions very smoothly in the past.

In other words, they could have consumer laptops and desktops running on ARM and the higher end Pro models running on Intel 64bit. Apple's own applications would probably be available in Universal binaries (they have the resources to do this no problem). Third party developers could create Universal binaries to avoid the performance hit, or let them run in the compatibility mode. From the user's point of view, it would be no more painful than any of the previous transitions. Probably less noticeable, with faster CPUs and more experience.

Previous processor transitions were born from necessity. This one, if it happens, would not be. There's nothing wrong with intel, if anything chipsets such as broadwell are the best things out there by a comfortable margin. There's no logical reason to switch, it would simply create a lot of work for devs, break windows compatibility, and lower performance considerably.

Which is why I don't believe it will happen. Heck just look at the debacle that is windows RT for a fantastic example of the perils of breaking ALL existing apps for no gain whatsoever.

You assume however that Intel put it's best CPU engineering resources into the version of Atom tested against the A7, which isn't true. But trying to compare a hypothetical A9/A10 vs. the best of Intel (Haswell Core i7), I don't see there being much of a contest.

Besides, the real problem isn't chip design, it's running a fab company. Unless Apple wants to go down that path in an attempt to "own it all", they will always be beholden to whatever process TSMC and Samsung are at, which is historically 1-1.5 nodes behind what Intel is at with its best chips. It would be a ton of work for not much savings.

First of all, ios on a notebook is not a good thing unless you only use it for web browsing. The ios isn't support a mouse pointer. do you thing a laptop like device which inclued a multi-touch display and no trackpad or mouse support is better and an iPad with bluetooth keyboard? Second, an arm processor isn't powerful enough for desktop operating system at a moment. But the application support problem is the most important problem if apple really getting the arm based mac on the market.

Well Apple also said they didn't like phones bigger than 3.5 inches. That that was the perfect size anything above that didn't make sense. They also said the same thing about smaller tablets.

Steve Jobs said those things, but he was willing to change his mind and had a company that could deliver when he did.

Instead of switching the entire Mac line from Intel all at once, Apple could create a new category (12" ARM-based MBA) to create demand + excitement to begin the transition. The appeal would be a 24-hr bat life laptop with retina display with an easy transition for devs to port apps from iOS.

With Apple investing $1+ billion over the past year to build a chip development team, the switch from Intel is a matter of when, not if.

That's a good point about battery life, too. A retina Macbook Air will not achieve the current 12h record the Air achieves right now. It will get 6-8h. But using the same large battery, they could get 2-3x the battery life with an ARM chip (so at least 15h, while pushing 4x more pixels).

Added a link to the story for context, also removed the video since it seemed to be confusing some people. I'm not saying no to the rumor, I'm saying it's not new and the video explains way.

After Surface Pro 3 my expectations couldn't get any higher for the Apple's vision of desktop/mobile systems and laptop/tablet future. I would have be blind to buy a MacBook Air now because the world has moved on.

I doubt it would run OSX. That OS wasn't designed for systems with low resources (there is a reason why Mavericks has memory compression) and it's tied to the dead PC era. Chip performance is also an issue; I couldn't even imagine the horror of opening iTunes on an ARM machine. iOS is also an awkward choice as Apple has gone out of their way to not have a hardware keyboard tied to iOS products. I don't know, maybe it will be a brand new OS (one built with hardware keyboards in mind)?

Basically this rumor seems forced to me as it's a new product that nobody seems to be asking for.

I think this is more targeted towards Chromebooks than full-blown notebooks. This will be a cloud computing device that depends on web applications to compensate for the lack of native applications for the first year and then, if all goes well, a new App Store would emerge. I wouldn't be surprised if this thing has a touchscreen.

What about a MacBook Air running ios with retina display, 4 gigs of RAM with a A8 AMD processor, could this be the 12" retina MacBook Air that's been rumored?

Sent from the iMore App

I don't know why you would downgrade from os 10 to ios. Ios to me doesn't feel like a complete operating system to me yet. It needs a file system. I need to be able to move data around on it like a regular computer for me to replace my computer with it. I need to have software that is as useful as Photoshop or aperture. I need to be able to plug in hard drives. Or connect to other drives with other files on them. Like if they are on an airport time capsule or whatever. I look forward to the day where people don't even want a regular computer anymore and all they need is an ipad. Apple has nearly perfectly redesigned the personal computer with the ipad. They just need to extend the usefulness of their creation. Make a touch version of aperture. Start making some real pro level apps. And get more hardware interconnections going on. A start would be an adaptor that allows you to plug in a special apple hub that you can connect hard drives, CD burners, flash drives. Any ipad should be able to see and make use of files on a time capsule. I should be able to look at and edit raw camera files. Import camera files export them to a time machine. Or save them online to the cloud.

I hope they don't do this until and unless they can have the same performance *and* gain significantly in battery life.

A few other points:

1) A laptop that runs iOS makes no sense. iOS is predominately a touch based OS and laptops are keyboard and mouse/pad devices. Mixing the two doesn't bring significant user benefits and for the people who want to do this, there are existing ways to do so.

2) A significantly less powerful Macbook running OS X doesn't make much sense. It would have to be priced far below the entry Airs, but those are comparatively cheap for what you get. Apple doesn't make cheap crap, so they're not going to shove something like this in cheap plastic case and price it at $500.

3) Battery life is getting close to being a non-issue. My 11" 2013 Air gets 7-9 hours of life off a charge. The 13" gets about 12-13 hours. Yes, doubling that would be nice in t he edge case (stuck at an airport with not outlets etc) but the marginal utility is getting less and less.

I agree with most all of your points. The only way I can make sense of it is to think that they are actually making iOS *desktops* and not laptops or tablets.

Could ARM chip talk be about an iPad Pro with a physical keyboard, and very close to a MackBook Pro in ability, but just lighter, and very much portable. I would love to see this, and use Auria which is a DAW designed for a touch screen.

"An iPad ... with a physical keyboard, ..." is not an iPad though is it? It's a hybrid laptop.

This is the essential problem the article is trying to work out. The only devices that make sense of the rumour, are devices that typically and currently run OS X.

Arm based Macs make sense to me.

Much like the Intel transition it's not "today's" Arm processor that the the Golden Egg it's where Arm is going to be in 2015/16.

You have to look at Pro and Con here.

Apple most likely has an Arm Architectural License. This gives them freedom that they will never have to produce their own designs. Pro

No Thunderbolt - Con

Arm cores are power efficient and multiple processors can be joined via CoreLink - Pro

Arm makes virtual more difficult - Con

If I'm a betting man I'm going to say that Apple starts with the low end (Mac mini and entry level iMac). These computers really only need to be fast enough to handle your basic tasks (Web, light document creation, multimedia). The rest of the lineup will stay Intel for the foreseeable future. Depending on the success of the Arm platform we could see hybrid Intel/Arm systems or further proliferation of Arm throughout the lineup.

You don't want to look at just architecture. That's a mistake. Intel macs make gaming a possibility. If that went away I would go back to windows. It would be the biggest mistake they could make on the mac platform. Intel chips are already getting much more power efficient. Sure look where arm is going, but don't be blind to where intel is going. Intel chips will be right up there with arm. The only reason arm gets more battery life is because it is a less powerful chip. Intel makes less powerful chips as well. And intel is willing to work with apple on system on a chip designs. I think apple is going to continue working with intel. If anything maybe apple will make some really low power systems based on intel atom chips. I think they would do that before making an arm mac.

There are two ways to look at this.

Apple is literally intending to switch to ARM. or..
Apple claims they intend to switch to ARM as a way of putting pressure on Intel to
produce better, less greedy processors.

Either way, its a win for Apple.

I firmly predict this transition to ARM will never happen for Macs. Therefore, based on my past track record, you should expect ARM based Macs at the next major announcement.