Will Apple Switch Macs from Intel to ARM?

Since Apple introduced the Macintosh in 1984, Macs have used three different CPU families. First was the Motorola 680x0 family, and it was followed by the PowerPC, which Apple helped develop in conjunction with Motorola and IBM, in 1994. Thanks to 680x0 emulation built into the Mac OS, PowerPC Macs were able to run most 680x0 apps.

In June 2005, Steve Jobs announced that Apple would switch from PowerPC to Intel "within the next year", and the first Intel Macs arrived in January 2006. Once again, the operating system had software, known as Rosetta, that allowed new Macs to run apps designed for old Macs. Apple continued to include Rosetta as part of OS X until OS X 10.7 Lion was introduced this past July, at which point PPC support vanished and several important legacy applications (most notably Quicken) were left behind.

Lately there have been rumors that due to the success of its iDevices, which are now built around an ARM processor designed by Apple, we're going to see Macs move from Intel to ARM, probably starting with the MacBook Air within a year. This week Low End Mac's staff discusses the pros and cons of such a switch.

Austin Leeds (Apple Everywhere): Whether I believe the ARM rumors or not is moot. I do, however, believe that Apple has been planning to switch to ARM for some time. The architecture is not quite advanced enough to power the latest Macs yet, but its progress has been stunning over the past five years. With their diminutive size, power-consumption, and heat production, Apple's A series CPUs could easily be grouped in sets of 8, 12, 16, 24 - whatever it takes to match the Intel CPUs.

The pros of such a switch would be independence from Intel's release timetables, and the rendering of pointless the argument that "Macs are just overpriced PCs with a different OS". If Macs could show a genuine increase in performance with the A series CPUs, like they did in the PPC days, then Apple would have that much more in the way of bragging rights.

The cons? Compatibility and reliability. I haven't experienced any problems with my A4, but that's not to say that a group of them wouldn't have some serious issues working together. Compatibility could be overcome with another emulation system such as Rosetta, but since IBM now owns the technology that drives Rosetta (and arguably forced Apple to kill Rosetta), it remains to be seen how Apple will pull another emulator out of thin air.

Dan Knight (Mac Musings): I am very skeptical of Apple moving Macs to ARM in the short term, primarily because ARM processors simply don't offer sufficient power at present to replace even the slowest 11" MacBook Air with its 1.6 GHz Intel Core i5 CPU, although a dual-core ARM processor may be able compete with the Intel Atom, a processor used in many netbooks. (Base on benchmark results published on Android Authority, the A4 was about 30% more powerful than the fastest competing dual-core ARM chips.)

Apple's A6 is rumored to be a quad-core processor, and when it ships we may finally see an Apple A series processor pass the 1 GHz mark, but the fastest ARM chips today seem to be 1.2 GHz, which is just a 20% boost. Couple that with four cores, and you could end up with enough power for a notebook computer - but still not as much as the slowest MacBook Air.

Further, ARM is a 32-bit architecture, which is limited to 4 GB of memory space. Every Mac made today has a 64-bit CPU and runs 64-bit OS X Lion, so moving to ARM would be a significant step backwards in an age where 6-8 GB of system memory is becoming more common. I just don't see that happening.

I don't see any reason that Apple couldn't or shouldn't produce an iOS notebook, especially with the quad-core A6, but I don't see any way for Apple to move Macs to ARM until there's a 64-bit version. At that point, with sufficient memory space and cores and clock speed, ARM could emulate Intel sufficiently well to make such a transition possible so Mac users wouldn't have to leave all their software behind, which I see as a very significant factor.

At present, the added bonus of Intel Macs is the ability to run Windows and apps for Windows. Although Windows 8 will be available for ARM, I don't see that becoming a significant option because existing Windows software used on the desktop is universally written for Intel's x86 processors.

Tim Nash (Taking Back the Market): One of the biggest reasons for staying with Intel for Macs is Windows compatibility. An increasing number of MacBook Airs and Pros are sold to people who want a well built laptop that is properly supported by the manufacturer, runs Windows well, and doesn't come laden with crapware. Even if Windows 8 runs well on ARM, there will be plenty of people and organisations that want to run legacy Windows apps on good hardware that is Intel based, because they rightly feel there will be no significant gain from trying to get old software running on ARM.

From Apple's point of view, Macs and MacBooks are still very profitable and are increasing their unit sales, revenue, and market share Year on Year with updates of components and OS X, but with little need to radically change the form factor. Macs as they are have become a cash cow that Apple should continue to milk for as long as there is a PC market.

Simon Royal (Mac Spectrum): There hves been a lot of rumours of Apple moving towards the portable device market with the iPhone/iPod/iPad, and with the whole move with Lion looking more like iOS, it is easy to see why.

I don't see Apple ditching their Macs just yet. I do see it happening possibly in a few years time when the portable market has caught up in terms of power.

The whole of the computing world seems to be moving towards portable, what with smartphones being more advanced, and netbooks and tablets becoming faster and more powerful. This is true not only of Apple, but Dell, Microsoft, Blackberry, and many others too.

I don't see the benefit for Apple to make yet another platform jump. Okay, most smartphones already run on ARM processors (Android, Symbian, and iPhone), and Apple's portable range already runs on ARM technology, but like others have already pointed out, it is too underpowered to replace the current range of Macs.

If Apple were to jump to another architecture again, this would alienate those who liked dual-booting or virtualising Windows - something that has been a big bonus to Intel owners. It would also mean yet another Intel to ARM Universal Binary situation, and this architecture hopping could put off software developers, as some of them have already had a hard time switching from PPC to Intel.

Personally, I think Apple might be planning it. They might be looking to move from big powerful Macs to small portables running on future high powered ARM devices running the iOS across tablets, laptops, iPods, and phones.

With the exception of the Mac Pro,all of Apple's desktop Macs have gotten smaller. Apple has just ditched the MacBook. Now we have even smaller Macs minis and ultra thin iMacs, and the MacBook Air is about as thin as you get.

Could this be Apple downsizing its devices? Could they be needing to pull away from Intel quickly? Could we see a shift in Apples direction, especially without Steve Jobs at the helm?

Charles Moore (several columns): I would rate the likelihood of ARM-powered Macs in the future as very strong to inevitable. There have been seemingly plausible reports that Apple has been testing potential use of its in-house designed A series of ARM processors in future MacBooks, especially once a quad-core A6 ARM chip, expected to be ready sometime next year, is ready.

A China Economic News Service report by Steve Chuang recently reported that Taiwan Semiconductor Manufacturing Co., Ltd. (TSMC), reportedly the worlds largest semiconductor foundry by market share, has started trial production of the ARM-based A6 processor in cooperation with Apple, with the production design to be finalized in the first quarter of 2012 and to be publicly unveiled in Q2/12 at the earliest. It is speculated that Apple's mid-term plan will be to eventually power at least its smaller, lighter Macs, presumably beginning with the MacBook Air, with A6 silicon.

This prospect is being taken seriously by Intel. In an interview with CNET's Brooke Crothers, Intel's Ultrabook group director Greg Welch affirmed that it would be imprudent for Intel to to be dismissive of the ARM in Macs rumors, but in the meantime they'll continue to innovate so Apple will continue to consider them as a CPU supplier - but Intel will also be hedging its bets with continued development of the Ultrabook platform.

Advantages for Apple in migrating Macs to ARM, aside from greater in-house control over product engineering and lower costs, would be smaller-sized components that would facilitate making their computers even thinner and with improved battery life. Downsides would include the probable necessity of a major OS X rewrite, and likely backward software compatibility issues that would make the troubles caused by OS X 10.7 Lion's dropping Rosetta PowerPC emulation look trivial. Even more problematical would be the probable loss of Windows compatibility, which has been such a key factor in Apple's market resurgence since the migration to Intel silicon in 2006. A workaround for that might be a hybrid CPU with tandem x86 and ARM processor cores, with Intel having reportedly affirmed willingness to consider working with Apple on projects like that.

Personally, I think it's more a matter of when than if, and the first wave of OS X/iOS convergence represented by Lion is an indicator of what's coming.

Adam Rosen (Adam's Apple): The decision to switch to ARM will likely be made on two primary factors: whether the design offers a roadmap for continued performance improvement with lower heat and power consumption, and whether the chips can be manufactured in high enough volume to supply Apple's needs. Both of these issues are what drove the change from PowerPC to Intel.

If Apple does make the change, we're faced (again) with legacy software compatibility issues: existing Mac software as x86 binaries, and Windows compatibility for the same reason. A fantastic side benefit of Apple switching to Intel chips has been Windows compatibility, through Boot Camp and virtualization software. The latter is very popular, with many of my business clients running Parallels or VMWare on their Macs. This feature alone has sold many Macs to new users, it would be foolish to remove this capability.

If everything "just works" for end users, either CPU architecture is adequate. But if Apple switches to ARM and then pulls another Rosetta-killing stunt, the headaches for Mac users might cause defections - either to Windows, or to an iPad with iOS.

Dan Bashur (Apple, Tech, and Gaming): ARM CPUs are powering devices running iOS with the greatest of ease. In addition, without the heavy demands of Flash, the web-driven content we typically consume on more powerful platforms is delivered to these lower powered devices in virtually the same quality through HTML5. The days of Flash being relevant became numbered as soon as Apple pulled the plug on it for iOS devices. Going forward, I would not be surprised if newer Macs begin to follow suit and eliminate Flash altogether, as well with future versions of OS X. [Flash is not included with OS X 10.7 Lion.]

Meanwhile, as ARM CPUs become more and more efficient, and with Apple's dependence on Intel to deliver CPUs that fit into their product designs, it becomes more and more obvious that the proverbial writing is also on the wall for Intel. Just look at the heat being emitted from the latest Core i7 MacBook Pro! When a device is running hot enough internally to boil water (the i7 quad core mobile CPU is designed to run up to 105 degrees C within spec), the design is clearly flawed and not built properly to be in a portable device.

Intel is currently scrambling to come up with more efficient ultra-low power solutions (in addition to the ultra-low power Core 2 Duo and i3), but ARM has a huge lead in miniaturizing this technology, while improving on it at a much faster pace as they continue to meet (and possibly exceed) Apple's needs in iOS devices. I envision a future with clusters of ARMs running in tandem to power Macs similar to how the PlayStation 3 was built with IBM's Cell processor technology. Octo-core ARMs are right around the corner, and with Nvidia's Project Denver, it's clear the the scales are tipping in favor of ARM in higher powered and more graphically demanding applications. When ARM is able to overcome the 64-bit hurdle and embed that kind of technology into their processors, the time will come when Apple and Intel part ways.