That would be a huge distraction for Apple. The phone is where the action is and they need to compete as hard as they can.
Shifting any engineering talent towards laptop/desktop chips is fighting last decade's war, for a shrinking (relatively) pie.
The only reason to do that would be if it gave them some research edge that they could use elsewhere. Like cloud servers or something.
As the article says, already more iphones than PC's. Why bother.
The way to use this power is to keep pushing where they have the advantage (mobile) and hammer PC's until you have all the power you need in your phone/tablet. Software will grow up around the ecosystem until it eats PC's, and Apple can ignore the PC market entirely. Forget about PC's, focus on beating Android and they'll beat PC's as a bonus.
One of the reasons to do this would precisely be to free up engineering talent that would otherwise have to be dedicated to developing specialized laptop logic board designs (and other specialized parts). If you can build a relatively flexible internal hardware platform that can handle laptop, tablet or phone use cases, then your production teams can focus on core problems for a specific product while borrowing whatever solutions they need from other teams. Apple already has similar chips and boards running iPads and iPhones, so why not consolidate laptops too?
For some evidence that this could be happening, look at ifixit's macbook teardown[0]. The logic board seems to be approaching the size of an iphone logic board [1]. Someone at Apple has to be asking a what-if question here when looking at this thing and thinking about how they could go all the way.
I've been following Apple and the rest of the industry since the mid 90's. Based on what I've seen I suspect that they already have built a handful of ARM Mac prototypes and maintain an ARM version of macOS, probably in a fairly serious way since the A7 shipped. macOS shares quite a bit of code with iOS and has been multi-architecture from the beginning. First bring-up was probably on actual iPhone or iPad dev hardware but they have the in-house fabrication capability to make machines that are visually the same as an existing model (probably 13" Air) with more or less the guts of an iPad. My guess is there are a small number of these floating around campus.
With those assumptions I still would bet against us seeing ARM Macs in the next couple years. There are a lot of advantages to keeping Macs single-architecture. Recent Intel CPUs are efficient enough that saving power there won't make a huge difference in user-available battery life (screen, DRAM, network, etc all use some too). Intel can price their CPUs aggressively enough to make cost not a major issue.
I do think it will eventually happen mainly because it will give them more control over their product line. Indicators of an upcoming shift will be if Intel screws up another generation of mobile CPU (it happened with Skylake, could happen again), Apple adds features like PCI-e for discrete GPU support (ok, they did in A9), and if Ax series performance starts to exceed corresponding Intel mobile parts by >50% (enough to compensate for the overhead of binary translation for legacy x86 apps).
> if Ax series performance starts to exceed corresponding Intel mobile parts by >50% (enough to compensate for the overhead of binary translation for legacy x86 apps).
Another milestone for an Ax laptop would be a good power efficiency story. If Apple can show an Ax based laptop with exceptional battery life under real working conditions that would be quite attractive to many people, even if translated programs experience a 50% drop in performance.
In addition to the Appstore bitcode stuff other comments mention, the Universal Binaries tools that were used for the mass powerpc->intel migration will also help to diminish the need for binary translation in many cases. UBs mean that you don't solely need to rely on Bitcode translation via the iOS/Appstore, either, so these fanciful Ax laptops could run all normal non-appstore apps quite easily.
>Apple adds features like PCI-e for discrete GPU support (ok, they did in A9), and if Ax series performance starts to exceed corresponding Intel mobile parts by >50% (enough to compensate for the overhead of binary translation for legacy x86 apps).
When it happens, it might not need binary translation.
For real -- having most of the *nix niceties in a visually appealing, hassle free environment is so nice. We are on Red Hat at work, and the hoops one has to jump through to get reasonable fonts up and running and the huge inconsistencies between desktop apps makes me pine for the Mac.
It's true for enterprise distributions, sure, but something like Arch Linux+GNOME 3.20 is REEALY nice. I use a Mac at work and my Arch with GNOME setup at home is easily matching the Mac and even outperforming it in terms of package management etc. Mac still handles high DPI a lot better, (but GNOME is getting there too) and if I was doing media content production, Linux probably wouldn't cut it for me, but I'm programmer and Linux is just the sweet spot of power + choice + usability for me.
Arch + Gnome 3.20 is what I use on my desktop too. The UX is excellent. Too bad that linux battery life and gpu-switching is still not up to scratch for laptops.
Very short sighted of parent post. One of the main reasons I don't consider (very hard) switching away from iPhone is the integration with my Mac (which I do still think is worth using, at least unless they ruin this platform).
As sidenote I should mention after reading your comment (which I agree completely) I felt sad for Microsoft, they are really in bad position, they lost mobile war and their most important income (PC sector) shrinking.They are late, and they are losing already.(I know MSFT is strong in enterprise sector)
Yeah I think you're losing perspective that outside of startup hipster tech bubbles, Windows is still the most popular desktop computing platform.
Mobile/tablet devices have struggled heavily at overcoming the one thing that's preventing their full adoption: making them great for content creation, not just content consumption.
A lot of big tech companies aren't building their products on .NET, for a variety of reasons that mostly boil down to it not being the right tool for the job. Recently the .NET web dev situation has improved. (Whether it has improved relative to the alternatives is arguable.) But when those companies were getting started the original iteration of ASP.NET was the only game in town, and it was very much not something you'd want to use for what any of those companies do. Nor was it intended to be.
.NET's always been more of a player on the in-house enterprise dev front, and probably always will be. (That might change with .NET Core, but it also might not.) Which means that you're not going to see much about .NET on TechCrunch. Inventory & financial management for Fortune 500 companies might be the silent majority for software development work, but it's rarely newsworthy.
Their consumer product escapades are very visibly, famously star-crossed, but their revenues have always shown consistent growth, and over the past few years they've had some realizations and made some re-alignments that put them in a good strategic position for supporting all their other operations.
That is a Silicon Valley view of Starbucks full with Macs and iPhones.
Here in Europe Microsoft is doing just fine across the continent companies, not only enterprise. Including selling Windows licenses for computers used in Android development.
Anectdote, but all Android developers (and WP devs forcwhat it's worth) are uding Macbooks Pro as their work machines. That's in Europe, and not even in Western.
I wonder how many of those devs "pivoted" from being web devs.
Best i can tell, Mac got a inroad in the web development world thanks to:
A. Being unix based and thus similar to the then prevalent LAMP stack.
B. Apple having a entrenched position in media production, and web development, via newspapers and broadcasters going online, becoming a part of that economy.
A) was important: Windows was going through some painful transitions and still required paid compiler licenses, which meant porting open source apps was painful, and while Linux was many things, a comfortable desktop OS was not one of them, especially if battery life or not crashing mattered to you (usually video driver or power management bugs rather than the kernel but people with jobs to do don't make that distinction).
A Mac was the one OS where it was easy to run PHP and Photoshop, and the small size + long battery life era starting with the Titanium PowerBook didn't hurt at all. I knew it was going to be interesting when all of the developers and Unix admins I knew started buying Mac laptops for daily use. If only I'd bought more AAPL at $17…
Well, I am yet to work in any client that uses Mac for any kind of development besides iOS, and even when they do, they timeshare an iMac across teams.
That is one thing to ponder. For MS the home/consumer desktop has been pretty much a means to an end.
This in that they can point to the consumer sales and say that those are people already familiar with MS products, and thus require less training when hired by a company.
That is, if said company base their computing environment on MS products.
MS going cloud is basically saying to said companies that MS can handle all the day to day support issues, for a fee. Thus the companies can reduce the size of their IT department.
How is it going to kill Intel? Apple's percentage of the desktop market is still a rounding error. If you think they're interested in selling CPUs to others, you need only look back to history to see it's VERY unlikely that will ever happen. If they aren't selling to others, they aren't killing Intel.
Not to mention, even if Intel managed to be "defeated" in the processor space (which I consider just about impossible at this point) - they've got better fab production than anyone else in existence. They've also got lots and lots and lots of products that aren't x86 CPUs.
7.4% of the worldwide computer market is hardly rounding error. Lenovo/HP/Dell put together account for a hair over 50% of the market, although it wouldn't surprise me if the total profits from Apple's 7.4% were greater than the total profits from the top three's 50% due to significantly higher margins.
By comparison, that's roughly equivalent to Honda's market share in the automotive industry, and I don't think anyone would dismiss Honda's sales as being a rounding error.
Other than that, I agree, they are clearly not the sort of company that would be interested in selling chips to other vendors, at least based on their history up to this point. If there were to start selling CPUs to others, it would not be playing to their strengths and would be a huge departure for them as a company.
Exactly. A lot of people seems to think Apple with only a 7% marketshare is small. But Apple are using some of the highest margin chips on their Laptop. I dont think it is anywhere close to 50% of Intel's PC market profits, but definitely in the range of 20-30%. Not to mention the marketing value of being inside Apple.
The next interesting question is at what price point would Intel give and Apple accept to continue using Intel x86 CPU inside Mac.
What you don't understand is that they're changing the game like the iPhone did to the blackberry. The iPad Pro is a computer that can work for most general use cases, which don't require heavy computing power or large storage onsite. With more computing power, it allows for more business applications to be ported over as well. Apple isn't aiming to replace intel in computer chips, it's aiming to replace the need for a laptop or computer.
Well you've got to consider what's going on at the low end of the market too. If Chromebooks are eating into Windows laptop sales and mostly running ARM, and Apple switches to ARM for some or many of its high end computers, which make up more than 50% of that market segment, that's definitely some bad news for Intel. It won't kill them though, just push them out of the consumer desktop/laptop market.
Apple sales are less than 1% of Intel's revenue. It will mean literally nothing.
You can paint me a GIANT skeptic of chromebooks in any setting outside of education. My step-mom got one as a present (as in for FREE) from her daughters. She's BUYING a Windows laptop because of how frustrated she is that it works with exactly nothing. Can't use her fitbit. Can't update her GPS. No way to sync her iPhone. etc. etc. etc.
It's a niche product for nerds. Any standard consumer will purchase one and never buy a second once they figure out all the limitations of their "cheap" Chromebook. That's ignoring the fact that I've seen more Intel based Chromebooks in the wild than ARM. I completely disagree with your "mostly running ARM" assessment and the only facts I can find on it do as well:
https://en.wikipedia.org/wiki/Chromebook
Respectfully I disagree. It's actually a great product for the majority of non-geeks who surf the web, take a few photos and watch Netflix.
Most newer devices sync over wifi to the cloud. I am pretty sure your Mom's iPhone can do that. The number of things you need a full blown PC for are diminishing.
If she has an iPhone why would she need to hook her Fitbit up to a computer? You can do everything you need from the phone via Bluetooth. And what is she syncing her iPhone with? I've not needed to sync my iPhone with a computer in years.
Every non-tech person that I know that has a chrome book loves it.
If you have an iPhone -- iTunes for music. Since I only use Linux at home now (yes, yes -- I should look at BSD variants, another discussion), I reluctantly got an Android this year. :-(
Any song you've downloaded from iTunes will be available on your phone through the iTunes app. If you've ripped a bunch of music then I guess you are right, but I find I don't use either, and instead use Spotify and Amazon Prime Music for all my music needs.
As for ARM inside the chromebooks : currently Intel sells chromebooks chips at low prices , and for a relatively smaller segment of the market - so they're OK with that.
What happens when ARM chips improve, and chromebooks hold a much larger share?
It is my understanding that Apple purchases the highest-quality chips of the highest-quality runs and pays a premium for it. What fraction of Intel's profits do Apple represent?
A relatively small part. Apple computer sales are not that impressive. They may be a large individual conntributor, but they alone do not sell enough to make a big difference.
Yes, Macs are a pretty small figure. The new iPhone modem deal Intel struck with Apple for around $1.5 billion is a bigger deal, which would be in the ballpark of 2% of Intel's ~$55 billion in revenues.
But Intel likes Apple Macs as a show pony. Intel has for years been desperately trying to convince hardware manufacturers to make both more consumer appealing PCs and new consumer devices around their chips. PC makers kept making the same old big beige towers, with the only differentiating factors being price (race to the bottom) and speed.
Apple doesn't like to compete in that space and likes to do other things and pushes on different features of chips that allow them to make smaller machines, quiet machines, fully integrated machines, battery friendly machines, and focus on industrial design.
Intel knows this and likes to use Apple to push other manufacturers to think in similar ways. The results have been pretty good as there are now PC competitors that make things like the All-in-one-flatscreen iMac, and the Macbook Air. If you remember way back before the iPhone, the original Apple TV was Intel based and showed Intel chips in mainstream consumer non-PC use cases. (This is different than concept demos because this was a real shipping product that actually managed to make enough money to sustain itself.) Also, Apple's obsession with reduction, like eliminating ports has been good for Intel because Apple helps encourage adoption of new standards they want to put out. For example, Apple embraced Thunderbolt at the beginning and was willing to drop legacy ports while most manufactures would normally hedge by keeping legacy ports. Apple pushing hard on this helps makes the 3rd party adoption faster which emboldens manufacturers to migrate faster themselves.
I agree with you, Intel is here to stay. But let's remember that Intel only has better fabs because they've had almost unlimited resources to reinvest in them as a result of the incredibly high margins on their processors which is the result of almost no competition. If the margins on those processors fall, their investment in fabs is sure to follow.
Can't find a good source for how much of intel's sales are Apple x86, but AWS does indeed use Xeons exclusively (with some Nvidia Tesla stuff) for EC2 according to https://aws.amazon.com/ec2/instance-types/
They did it before in the PowerPC to Intel change. Migration was helped by:
- Fat binaries (Mach-O binaries can support multiple platforms).
- A PowerPC emulator for applications that are not ported.
I think they are even better positioned for an architecture change than during PowerPC -> Intel. They now have the app store where they can impose certain requirements (like supporting two architectures). And they now have their own compiler backend, which opens the possibility of supporting architecture-independent IR (along the lines of bitcode).
There was also the 68K transition to PowerPC a long time ago. And more recently, the 32-bit to 64-bit transition.
And for a lot of the older Mac developers, iOS was kind of like another migration. And iOS has gone from armv6 to armv7 to arm64, plus the x86 and x86-64 simulator targets. Not to mention that there is now an LLVM Bitcode requirement for Apple TV and watchOS.
Apple and its developer community has a lot of experience with architecture migration. Each transition built on the experience of the previous and got smoother each time. Apple has been very good insulating their frameworks and tools from the architecture, and the Apple developer community has gotten very good at following Apple's guidelines to minimize disruption since there have been so many of these transitions.
Forgot to add that I think the real thing keeping Mac on Intel is the bullet point that Intel Macs can install Windows.
In the PowerPC days, Apple had trouble convincing people to switch to Mac because Windows being so dominant, everybody was afraid they might need Windows for something and that would make a PowerPC Mac an expensive mistake. The Intel switch alleviated many fears because in the worst case and Mac OS didn't work out, they could just install Windows. Bootcamp and virtualization provided additional options.
Apple has always seen Mac and iOS as different markets. Mac is still a small market compared to Windows. They are still trying to grow which means still trying to convince Windows users to switch and the safety net of Intel is still useful. (Windows RT isn't a realistic option.) I personally haven't seen iPhone users wishing for a big laptop or desktop that runs their same apps.
Windows was already on ARM with Windows RT, for which I believe Microsoft did the work of porting UEFI and standardised bridge (PCI) technologies to ARM.
While speculating: Apple could conspire with Microsoft to standardise these technologies together to lock up their ecosystems against the Linux / Android monster :-)
The hard part for Microsoft is not the porting of Windows to ARM. The hard part has been and continues to be convincing existing Windows developers to port their apps to UWP.
The reason Windows on Mac is useful to people is because people have may programs that do not have acceptable equivalents on Mac OS. These apps are usually legacy programs that probably fill special niches, and the cost of "modernizing" is usually not justified. This effort would be required to port these apps to Windows ARM, which last time around also required porting to UWP which Microsoft is still pushing, but not easy to actually do for a lot of code bases.
A new Microsoft irony kicks in here is that if they are successful in convincing the huge Windows ecosystem to migrate to UWP away from Win32, et. al, they may actually kill their lock-in advantage. A serious rewrite at this stage may also invite Mac, Linux, iOS, Android ports. In this case, Apple no longer needs to care about Intel Windows and this bullet point becomes less compelling and maybe they could reconsider.
You can use the new Centennial technology to wrap an old Win32 application as an UWP Store app. If the app was made with .NET it might even run on ARM.
I think this market is getting smaller and smaller year over year (I no data to back this up unfortunately). They could honestly build a MacBook with ARM in the near future while keeping the pros around with x86 for a while longer. They could reduce the price by hundreds of dollars switching off of Intel processors which would undercut a huge amount of competition.
The sole reason I can use a Mac for my paid work is, that I can run Windows and x86-Linux VMs, as I need to run Intel based software on those operation systems.
Well yes and no. I ran Windows in VirtualPC (I think) on a PowerPC MBP for a while. Pretty slow but good enough for Outlook, Word etc that I needed the compatibility for.
Say, does anyone know who owns the IP for FX!32 nowadays?
Putting aside the fact that Windows RT was a commercial failure and discontinued which makes it weird to push yet another effort that this would be anything more than some niche demo thing like Raspberry Pi support, Windows in of itself is not what concerned potential Mac switchers.
People are concerned that they have some mission-critical, unreplaceable app that only runs on Windows, that has no acceptable Mac analog (doesn't exist, doesn't have feature parity, data formats are incompatible).
Windows on ARM doesn't solve anything because almost no Windows apps that people care about were ever ported to Windows RT. (It was probably more likely there was a Mac port.)
Not to mention it's been a long time since Intel really stopped the show with a CPU release. That was not the case way back when Apple switched to Intel. Granted I remember the RISC vs CISC wars with nostalgia.
And Apple could transition even easier now thanks to it controlling most Mac apps through the Mac Store. All Apple needs to do is announce they're going to enforce ARM support in the Mac Store a year or two before doing it, and that's it.
Also, isn't Apple already encouraging the use of some intermediary bitcode for iOS (and macOS?) apps? Wouldn't that already made apps architecture-agnostic?
You're not being serious are you? MAS holds such an insignificant role in Mac App development that any requirements it sets around the transition would weaken its cause.
Also mo, bitcode would not be able to be used to facilitate an Intel to ARM transition - bitcode is still fairly processor-specific. Bitcode is more for being able to adapt to minor changes in instructions in the same arch
>You're not being serious are you? MAS holds such an insignificant role in Mac App development that any requirements it sets around the transition would weaken its cause.
Negligible? If you exclude Adobe and MS, most apps people use are available through the Mac App store. And more can be made available if Apple pushes more for it.
With the restrictions it currently imposes (sandboxing), I think we would see a significant withdrawal from the Mac platform. Many developers would stop making apps (especially those that are for power users, which when turn more devs away from macs). I doubt Adobe Creative Suite (Photoshop) would be technically possible with sandboxing, so Adobe would withdraw from Mac. MS Office would probably be gone as well. Parallels and other virtualisation software would find it extreme difficult to run.
Valve would have to pull Steam because you can be as sure as hell they aren't going to give Apple a 30% cut of Steam sales. There goes Skype as well.
Without special deals with lots of developers, I doubt we would see many more apps on the App Store and instead we'll see the Mac be an unsuitable platform for many/most people. It would turn into a glorified Chromebook.
In this hypothetical situation Microsoft doesnt make the stupid move of requiring all apps go through their store which imposes its own limitations (like sandboxing).
Mind you, it would be a completely insane move for Apple (or Microsoft) and I can't see them doing this any time soon. It would be absolute suicide.
How is Adobe going to do it when Windows follows the same footsteps than Mac OS X?
The only way will be the store, or why do you think Microsoft is making it easy to port Win32/WPF (legacy) applications to the store model?
When the applications that matter like Adobe are on the store, and Apple has proven the "my way or the highway" works, they will slowly disable the "legacy" model.
I have been through enough computing changes to believe this will indeed happen.
> thanks to it controlling most Mac apps through the Mac Store
Serious question: Do you have any evidence to support that claim?
I could well be in the minority, but I don't publish through the Mac app store and I don't know anyone else who publishes exclusively through the Mac app store (except Apple).
Well you'll have to if Apple decides that ARM MacBooks will only be able to install apps from the Mac app store - they already have a precedent in iOS and precedent of pushing developers to do things their way.
iOS had a very different history with developers. The desktop isn't the mobile space.
Desktop developers have their own solutions in place and have for a long time. You'd be forcing them to give up a ton, in exchange for nothing of value. I'd expect many would just stop developing applications at all instead of comply with store requirements as they stand right now.
That's not to say those problems couldn't be resolved in the future.
All nice and good, and where would those developers go, considering that both Google and Microsoft are following similar sandbox approaches?
Those developers can choose between stay in business or go broke, because I doubt GNU/Linux or *BSD users would pay for desktop software as their current Mac OS X customers.
I don't think Adobe and Microsoft, the two biggest third party Mac developers, care what Apple imposes on the Mac App Store.
And they certainly won't be too happy about having to rewrite, again, their applications (I think Office for Mac is not yet fully 64bits, but I could be wrong, haven't checked in a while.)
> I don't think Adobe and Microsoft, the two biggest third party Mac developers, care what Apple imposes on the Mac App Store.
Microsoft must care somewhat because they have spent a lot of time fully sandboxing Office 2016. Since OneNote is already available in the MAS [1], I reckon Microsoft will eventually add the remaining Office apps at some point in the future.
> I think Office for Mac is not yet fully 64-bits
Microsoft released the first 64-bit version of Office 2016 for Mac (15.25.0) last month [2] [3]
As the article says, bear in mind that a lot of those applications are now on iOS. They seem to think a shift to ARM might also mean a shift to iOS, with the iPad Pro acting as the bridge device to encourage software makers to provide compatibility (which seems to be working). One possibility might be Apple segmenting its laptop range, with the MacBook being shipped with iOS, and the MBP with OSX.
They say in the article that the entire PC market shipped fewer units than the iPhone alone, one quarter last year. There's certainly a risk that the consumer end of the laptop/desktop market gets dragged in behind as an 'additional device' for the mobile operating systems, bringing with it its walled gardens and concept of 'devices as utilities' rather than generic computing platforms.
I suspect as strange as it might seem to us a lot of people would jump on a laptop with the managed qualities of a utility device, especially if it meant they could get a Macbook for $150 less.
The one think keeping me so far from buying an iPad Pro and probably will cause me to buy a macbook next is the software limitations of iOS which prevent it from being a generic computer. So if Apple moves to ARM, it probably would be rather some more macOS style operation system. But that switch would be not really difficult as iOS and macOS largely share the same codebase.
The history of Mac OS X on Intel (the Marklar project) is very interesting. Most of the work was done by a single dev (ex-director who demoted himself to engineer) who wanted to spend more time with his family. So he worked from home.
Changing to Intel are kinda easy, because applications (think Photoshop, etc) already has optimized code for x86 (think SSE/AVX and such, which are hand-coded). While undoubtedly many also has NEON nowadays, I still think it is on different scale.
One other factor is the degree to which hand-optimized code in general has been going away: Apple heavily encouraged people to use system frameworks like CoreImage, CoreVideo, etc. both to ease the transition and to take advantage of their significant optimizations, and that appears to have been followed by many developers. Adobe presumably still has a whopping case of Not-Invented-Here but I'd be surprised if a significant number of apps couldn't be ported with very modest changes.
My prediction is that we will see a new MacBook (Air) next year (probably October-November), with an ARM CPU running MacOS 11. Apple with bump MacOS (nee OSX) a major version number and keep it in parity with iOS, which will also be version 11 next year. Essentially iOS and MacOS becoming front ends to the same OS (which they almost are).
They will start with the ARM CPU on a single MacBook (an entry level one) for a few years before pushing it our across the whole range once app compatibility has been achieved.
The shift to Intel was a move of necessity, a huge risk, and ultimately a major pain for Apple's customers not talk about third party developers. Apple took a huge hit to their goodwill. That was predictable, but they went ahead and did it anyway. Why? Because they had to. PowerPC was falling further and further behind. It was alright for Macs to be more expensive than PCs, but more expensive and slower? Furthermore, PowerPC was holding Apple back from taking the Mac where they wanted it to go (ultralight laptops).
None of the issues that caused the Intel transition apply to Apple now. Intel is not holding Apple back, because Apple already uses its own chips in what it considers the future of computing: iOS devices. At the same time Macs use exactly the same chips as PCs, which means there is no risk that PCs will suddenly leapfrog the Macs. There is in other word no major need to make a shift. But the risks in doing so would be the same. Apple would once again be pissing of its customers and 3rd party developers. And for what? A marginal improvement in battery life? To save a fragment on bill of materials on a product line which already enjoys healthy margins, but which also is a tiny fragment of their overall business?
Considering how light and fast the 9.7" and 12.7 iPad Pro is having a 2lb 13" Air with similar speeds becomes fairly distinctive. It further separates the Air vs Pro lines and could have ~18 hours of battery life.
Right now the basic MacBook is 1200$ has a larger screen and lower weight than the smallest Air. The 11" Air on the other hand is much cheaper.
> It was alright for Macs to be more expensive than PCs, but more expensive and slower?
Isn't that becoming the case with iPads and (low end)Macbooks? If the rumored performance of new A10X chip is any indication, that is the direction we are going.
The PPC era did not just end with "more expensive and slower", it ended with "more expensive and slower than the competition". If buyers refuse the Macbook in favor of an iPad, Apple won't shed much of a tear.
Last time I checked, PC desktop were dominated by Windows by about 90% market share and that market, while shifting to mobile, isn't tiny either.
If they could come up with a desktop that is as small as Apple TV while providing enough performance of a large box desktop or shaves off 300g off of laptop without compromise or doubles battery power, that would be shocking even to average consumers.
As I have been losing interest on Apple products for the past 5 years not seeing much of innovations, I think it's time they take the risk while they can.
It's a shocking news if x86 drags Intel (and to some extent Windows) to secondary position, if Apple does it right.
I think you vastly overestimate how much average consumers and businesses know or care about any of those issues. Apple could release a chip that was 500% faster than the equivalent Intel chip while only consuming 75% of the power and it wouldn't move the needle on the distribution of Mac vs. Windows purchases. The vast majority of people and businesses who buy Windows buy in for the ecosystem and nothing else.
PowerPC to Intel had some built in advantages: Intel was really pulling away from PowerPC and it gave Apple immediate hardware parity. Before Intel you could love their software but there were real questions about performance.
The advantages with arm seem less obvious, why would you do anything but let the market decide? Build up bigger and better iPad pros until it erodes you MacBook market.
Arm on desktop (win or Mac OS) has to happen now that the chips are this quick. I don't really need a phone with desktop perf, but I'd love a laptop with phone class battery life.
No doubt full Windows is already running on Arm at Microsoft - the problem is likely that legacy Win32 x86/x64 apps need to run on Arm which requires some sort of effort similar to when Apple switched to Intel. Only having.
To add to this, third party Windows programs have very little in common with each other, as developers on that platform use hundreds of different UI toolkits, libraries, etc making porting to ARM infeasible due to the number of dependencies that need to be ported along with it.
Contrast this with the situation on Apple platforms, where nearly all developers use Cocoa and the vast majority of third party libraries already support ARM in addition to x86. Most of today's Mac apps could be running on a hypothetical new ARM based Mac with developers doing little more than ticking an additional architecture box in Xcode and recompiling.
Apple has everything that kept MS from succeeding with their ARM desktop OS transition and could even make the leap relatively smoothly.
"Full" emulation has to be provided - that is, an emulator that lets ALL Windows programs written years ago to be run on ARM Windows without recompilation.
They have started to move to platform independence with UWP, but legacy software must also work otherwise it's not Windows.
Microsoft have tried to run Windows on multiple platforms before, and IIRC, failed. NT was available for the Alpha and MIPS, but I believe almost everything had to be emulated.
Yes, and phones. No ARM windows desktops or laptops, and no ARM Windows tablets for the "full" Windows however.
They need an ARM version of Windows that runs Win32 Intel programs. They already have an emulation layer for compatibility with 16-bit win32 Intel programs in Windows today. An ARM compatibility layer would be slower, but still doable (c.f. Apple when switching to intel)
A shift to ARM is a big deal for developers, so they'd probably want to give them a year like they did the last time this happened. It's also the kind of thing that Apple would want to announce at WWDC. We just had one of those a couple months ago, so that makes the next opportunity for a switchover more like summer 2018.
Shouldn't the iPad simply become the MacBook Air? There will be a convergence with the low-end laptop. It's simply a matter of getting desktop software running on iOS. As with the new Airpods, you instantly pair your keyboard, if it's not attached, and mouse. Apple's strength is iOS so they should grow it into the laptop.
I have been predicting this for a while. I keep saying "hmm maybe in a year or two" but it never seems to happen :)
I still believe the future of macOS will be ARM though. Apple have invested so damn much in making their own SoC that I can't imagine them not going that way. Soon (within a year or two? heh) they will have an SoC that they can pop in the MacBook and performance will be so close to the m-series Intel chip that only lab benchmarks will show a difference (and even that will be tiny).
Once that happens Apple will own the own compute process. While very cool it also sucks as I have no doubt the Mac will become as locked down a iOS. No sideloading and certainly no dualbooting. The Mac will essentially be yet another appliance albeit it one with some kind of development environment. It won't be something you can hack on though (well not without having to actually hack it first).
I'm pretty skeptical of the notion that iOS could replace macOS any time soon. Here's an experiment if you don't believe me: grab a mac, run the iOS emulator on it full-screen and try to do all your normal tasks in the emulator, without jumping out to the desktop. Let me know how that feels after a few hours.
You could argue that with some engineering time they can add some affordances to the iOS interface to make it more pleasant on desktop. Sure. But every app is also going to need to adapt for this -- they need to drag along the entire ecosystem. Honestly it would be way easier to make macOS run on ARM than it would be to make iOS run on desktop. They already did this whole processor shift thing once and it worked out just fine. Just package things in universal binaries and emulate the old CPU for legacy software. Not to mention with something like LLVM it's not crazy to imagine distributing applications in a near-native bytecode that gets JIT'd to the target machine.
I know microsoft has gotten pretty close to accomplishing the entire one-os-for-everything model, but look how much time and money it's taken them, and the amount of good will they burned through getting there.
A lot of people forget, but iOS shares the same underpinnings as MacOS X. In fact, when Steve Jobs announced the original iPhone, he didn't say it ran iPhone OS, he said it ran MacOS.
I wouldn't be surprised if there's macOS already ready to go for ARM, much like there was an Intel MacOS for years before the transition was announced.
As for whether iOS could replace macOS, if you go back to my first point, think about iOS as a version of macOS that was stripped down and has been built back up over the intervening decade. I don't think they could make the switch tomorrow, but in a couple of years, it's not unreasonable to think there could be a reunification of the iOS and macOS codebases, creating, essentially a fully-featured iOS for desktop hardware with the features we've come to expect from modern desktop OSes.
My wife was commenting the other day, when she noticed a photo of the iPad Pro 12 with the Smart Keyboard, that Apple should make a touchscreen laptop that would let people install either Mac OS or iOS. I don't personally know if such a device would be accepted but there are some people interested in such a device.
Personally, I prefer they keep Mac OS and iOS separate.
What does this have to do with anything? Mac OS is totally portable (at one point. OpenStep provided transparent support for four CPU architectures), and server-side software even more so. Similarly, Apple could presumably port whatever subsets of Mac OS are needed as the mood takes them.
It wasn't that long ago that a universal binary on OS X meant four architectures: 32 and 64-bit for both PowerPC and Intel processors. That part of the Mach binary format worked quite well as most people never even noticed.
I never said macOS wasn't portable, so I'm not sure what you're getting at? My point is that iOS would suck on desktop without a lot of work, and a lot of that work would fall on app developers, not just apple.
The article is saying that intel should be worried because Ax chips are competitive with their chips. Whether iOS is a good replacement for Mac OS is irrelevant. Mac OS can run on Ax chips with no significant effort from Apple and not much from third party developers.
Both the iPhone 7 and the next next iPad Pro will be significantly faster than the current Macbook (and Macbook Air) and that's probably also the case for the next iteration, if it sticks with Intel. That alone is reason to switch to ARM. Compiling to ARM will be relatively easy, but still emulation will be needed for many applications, and that might be harder than emulating PowerPC.
I wonder where the performance ceiling is going to be for the Apple processor line. Looking at Intel, I'd have expected them to have hit it already, but given the huge increases in single threaded performance so far year on year, it's hard to believe that it'll peter down to <15% per year from now on.
If performance is going to keep improving at the current rate, then the A11X will be significantly faster than Intel's 2017 Cannonlake mobile lineup. Switching away from Intel would give Macs another big leg-up relative to Windows laptops.
Macs will never switch to iOS though. Processes on macOS have a freedom that's incomparable to that on iOS, and once the genie is out of the bottle, you can't put it back in again. But Apple will keep working to make iOS sufficient for increasingly more people, and Macs will keep getting more niche. Apple could even introduce a laptop-like form factor like the Surface Book with iOS on it (the iBook Pro?).
If you look up Geekbench scores for these devices at least, the iPhone 7 gets a higher score than the 12" MacBook, the latest MacBook Air, is similar in performance to a 2013 MacBook Pro, and even has better single core performance than the top of the line 12 core Mac Pro.
Still apples to oranges, and it's just one benchmark. How long does e.g. Npm install take with the same project, compiling the same dependencies? ARM might come to desktop eventually, and it's damn impressive that for a single benchmark from a single test, someone was able to get the ip7 to beat a >1000 days-old Mac Pro in one metric, but they've learned from the PowerPC switch and a new arch won't happen overnight. I very highly doubt that the MacBooks hopefully coming later in 2016 will be anything other than Intel, hopefully Kaby Lake. I'd predict that the earliest we'd see an ARM MacBook would be 2018, and my guess is that they'd do it as a separate mode (so, have the A12X CPU in there with an Intel Coffee Lake or Cannon Lake main x86 CPU) and let it do something like dual-boot iOS and macOS, or have way tighter integration (run iOS apps as just another window in macOS, for example...dual-booting seems un-Apple-y, even with BootCamp).
I doubt we will ever see a laptop running iOS. iOS was designed solely for a touch interface. At the same time, apple has been outspoken about not wanting touch screens on laptops because of ergonomic issues and other things.
They will however happily sell you a $170 keyboard for your ipad...
But snark aside, apple is always outspoken against things... until they release their own version of it. That's not some apple specific trait of course.
- Most 3rd party Apple development is already happening on ARM. This would inject new life into the desktop/laptop product line. Think of the thousands of apps that could now run on those platforms.
- Apple gets rid of a massive chunk of BOM cost improving margins, i.e. Intel CPU.
- No need to emulate X86 if Apple can do binary translation instead. There are already OSS projects like MC-Semantics (https://github.com/trailofbits/mcsema) that can take an X86 binary and emit LLVM IR. In theory you could give that to the ARM LLVM backend and relink to ARM frameworks. This would be an interesting problem to solve vis-a-vis code signing. One way would be for Apple to sign a code signing identity for each user, have the OS translate the binary, prompt the user for a password to get the private key from the keychain, sign the new executable, save it, and then run it.
- I actually wonder how many desktop apps are commonly used on the macOS platform, and how big of a problem in 2016 an arch switch would be. Muggles are already predominantly using mobile devices. I'm guessing 99% of users use one of 100 (or less) desktop apps. I imagine it to be quite the long-tail.
How plausible is it to stick an ARM chip and a (lower-end) Intel CPU on the same board and make them work together? The transition to an ArmBook would be much smoother and I'm betting it won't affect cost of the laptop that much. After all, an Apple TV, which is essentially an iPhone 6 board without the battery and screen, costs just $149.
That actually was done toward the tail end of the 1980s PC wars. The Commodore Amiga and several other machines could sport x86 add on boards that would let them run DOS apps natively.
It would not be good for a laptop though, and that is almost all of Mac sales. Too much space and power consumption.
Pretty much every iteration of the Gameboy does this (at least up through the Nintendo DS) in order to achieve backward compatibility, so it has at least some plausibility.
The iPhone 7/7 Plus single-core score is 3,450 and the multi-core score is 5,630.
The current top intel chip (i7-6700K) single-core score is 5326 and that same chip multi-core score is 17003 and the i7-6950X multi-core score is 29877. http://browser.primatelabs.com/processor-benchmarks
That's still a pretty good lead. Now perhaps with the added room of a laptop they could make an arm chip that was closer in performance but would it be worth it?
You're comparing a chip with two[1] cores to a 4 or 10 core chip so the comparison is complicated. It'd be straightforward for them to add more cores – and we'll see that with the next iPad Pro, I'm sure – but at some point they're going to hit scaling issues for things like memory bandwidth.
The real difference, though, is power: that 6700-K uses 91W and the A10 should be around 2W. There's no way you can put a 6700-K in a normal laptop and the mobile processors appear only to be delivering single-core scores which are already within iPhone 7 range.
The other question is how often single-core or even multi-core performance matters for the average user vs. GPU performance since that's often the only place where people are limited by something other than network or storage speeds. I think a growing number of people are within the range where their needs would be satisfied with perhaps an extra core or two but a lot more GPU capacity.
The i7-6700K is a 91 watt chip. Remove that and they are fairly close, worse the A10 line has been improving much faster than Intel's chips over the last five years.
Also, don't forget it's really the mid range chips that matter and the A10 is very much a mid range chip without the benefit of excessive binning.
What's the power consumption of the last AX CPUs btw ? I just laughed thinking about the iPhone 7 main "board" that fits under the space key. An A10f laptop would be a the board + 4 batteries and a IO hub... how long would that last ?
The AX line is really a system on a chip, but the CPU is something like 0.1 to 0.2 watts. Remember these get ~1/3 of a ~2,0000 mAh battery and can last for several hours.
iPhone 6 has a 1810 mAh battery at 3.82volts = just under 7 watt hours. If the phone including display, cellphone chips, wifi, Bluetooth, Ram, and COU etc add up to 4 watts you would not get 2 hours of use from the phone.
To the extent you're comparing A10 power usage to Intel's published numbers, keep in mind that Intel's TDP is what the processor uses to operate at the rated clock speed at full load indefinitely. In typical usage (where the CPU operates in short bursts and sleeps the rest of the time), the average power consumption will be a lot lower.
7 / 2.5 is still under 3w for the total prone draw.
All of which suggests in a laptop with active cooling and higher energy budget say 20W you could sustain 2+ ghz x6 cores or 20,000 to 24,000 with minor adjustments. Which is in the ballpark of Intel's high end desktop performance and faster than their mobile chips.
The real issue is Intel is dealing with a well funded competitor who does not care about x86. And the computer history has a long line of companies who where eaten from below as people where happy to pay a lot less for fewer features.
When the next iPhone has a single-core score with a wide margin over the single-core score of the fastest current normal i7, and an equivalent multi-core score against the same i7, then we can say Apple has caught up.
Why wide margin of single? Because artificial benchmarks are not a good measure, and due to unrelated factors, and also due to how most programs are strongly single threaded in performance, program execution performance in the real world no longer linearly scales with raw IPS increases, nor does it scale linearly with IPS + cache/memory performance increases.
Also, the whole "but the iPhone chip is only dual core, and modern i7s are quads"... too bad. If you're going to make the argument of desktop performance, then you have to apples-to-apples the comparison. You can say, however, the new iPhone CPU has good performance, because it legitimately does... for a phone, not for a desktop.
Pretty bad article. Why recommend the move to iOS on laptops instead of simply recompiling MacOS to the ARM architecture. Not an issue in the Linux world, dont see why Apple would have any issue with that if they really wanted to.
As some commenters pointed out, Apple already has a big part of the OS ported to ARM. iOS shares stuff with macOS - XNU Kernel and Foundation libraries. To have a complete port they would need to port the drivers, AppKit and their applications. And it seems it wouldn't be very hard for them because they already did that once (from PowerPC) so they have some experience with that.
Much easier this time around given that Carbon is no longer in the picture. There will be holdouts like Adobe and Microsoft, but those companies are not the powerhouses they used to be on the Mac.
The switch to ARM is largely a matter of 'when' not 'if' at this point.
Given that Apple control their own CPU designs, how hard would it be for them to add a hardware x86 ISA emulation layer over their ARM CPU for use with MacOS? That way the OS and newer software could be compiled for ARM, and legacy software could be run in x86 mode with reasonable performance without the need for rosetta-style emulation.
Parallels is performant because it isn't emulating the entire processor. Most instructions are passed through to the underlying hardware without translation.
If I remember correctly, Transmeta used a VLIW CPU where there is a low level firmware doing translation from x86 (or what do you need) to the native ISA.
Yes – that “Code Morphing System” had a complex layered process where code started running in an interpreter and was progressively optimized as it was repeatedly executed. The main problem they hit is that even with a simplified CPU, the economies of scale are brutally unforgiving to a niche product and they could never demonstrate a compelling advantage to prospective buyers: there were a few microbenchmarks where it looked competitive but real world code never showed much benefit and the larger engineering teams Intel/AMD could afford to support continued to race ahead.
In the case of Apple, the situation could be different because they're guaranteed to ship in significant volume, their potential budget for R&D is massive, and most importantly they control the entire OS and have significant sway with the application developers, not to mention how far dynamic binary translation has improved. It'd still be a huge project but it's not unrealistic to image something like running a slow, offline binary translator over the entire contents of the App Store, which simply wasn't the kind of thing anyone could do in the 90s.
This is one reason I quit the whole CPU commentating thing. People kept rewriting the same article (with the same comments thread attached!), and I kept rewriting the same responses.
I think there is something satisfying about the topic, that we enjoy rehashing. And to opine on what's new since last time, hoping for exciting new developments.
I think the future is very powerful mobile devices that have a very smooth docking experience. In business environments this is good because knowledge workers don't have to lug around laptops. Instead, walk into a conference room, sit where there is a monitor/mouse/keyboard and your work environment is right there on you phone. Sort of like Google's multitude of mini-conference rooms with audio/visual setups, but augmented with shared peripherals.
I'm still hopeful this will be the case -- that all this multi-device cloud syncing nonsense today is just a fad and we'll all be carrying around phones with terrabytes of storage and desktop grade performance. Not holding my breath, though.
While technically it would be fascinating to have an ARM based MacBook - just imagine what Apple could do with more power/thermal budget - the big deal breaker will be the compatibility with x86 baseded code. Currently Macs are very popular because they can run (partially via VMs) basically any program written in the last decade. This is not easy to give up, especially on the MacBook Pros. There could of course be an entry level ARM MacBook. Unless of course, they make some processor which can run both ARM and Intel code, but that sounds a bit far fetched.
IIRC, Longsoon is a modified MIPS chip with extensions designed to allow running x86 code via qemu very efficiently (relatively speaking, at least). Since Apple designs their own cores, they could attempt doing something similar with ARM.
Apple controls both the hardware and the operating system, so they could build such an emulation layer into the OS. I have no idea how crazy this idea is in reality, but in my imagination it could be done.
They could start requiring bitcode for Mac App Store submissions, and compile things to ARM and x86 as long as they have to support both platforms. Or they could just bring iOS to their desktop class computers. Also, why not emulate x86 apps on ARM? Sure, that would be terrible for games, but might work well for most productivity apps.
The operation system is not the problem. They for sure have a full macOS for ARM working. Its the applications. Most desktop apps are not sold via the App Store, and there are the VMs (Windows, Linux) to think about. For my work for example, I need to be able to run x86 binaries at full speed. So an ARM based Mac would not work for me.
Unfortunately, bitcode is too low level to use for cross-compiling. It's an intermediate representation that's already "on its way" to being code for one architecture.
It is more useful for supporting a new CPU feature or maybe working around a compiler bug without fully recompiling.
That link is referencing Geekbench 3, not Geekbench 4.
I did some searching and found the following quote concerning the Geekbench 4 benchmark:
“These updated workloads include several well-known codebases that are used every day on mobile devices, such as LLVM, SQLite, and PDFium. These updated workloads model real-world tasks and applications, and provide an objective measure of the performance of the CPU in your phone or laptop.”
That sounds like it is testing a lot more than hardware accelerated encryption.
I am not sure how good of a benchmark geekbench is, but ever since AMD stopped pushing Intel, they have been a little complacent. When AMD introduced Athlon 64 it really forced Intel to step its game up. The core 2 duo was pretty impressive. AMD hasn't had a competitive product in years and Intel has just been providing some incremental improvements here and there.
With shrinking transistor sizes they can shorten their pipeline maybe add a few more front end decoders, add even more cache. If you have a shorter pipeline the branch misses aren't as bad. There is a lot of stuff they could do which they probably aren't since they haven't had a real incentive to do so.
I think sometime next year when the A10X is released the question of MacBook Pro or iPad Pro will come down to the following:
“What device of, which both have the same performance and storage, do I choose? The one with the USB ports, the inflexible keyboard, bloated software and costs 600$ more, or the other one with a touchscreen, optional keyboard, and much much simpler software?”
Also I think that Apple won’t do any ARM MacBook at all. If we consider a RICE scheme for gauging priorization for such a feature like ARM CPUs in MacBooks:
Reach: Modest, there aren’t that many Mac users compared to iOS
Impact: Positive impact on Mac users is limited. Performance won’t be that better, mostly equal. Software availability could be limited for a while during a transition.
Confidence: Apple has made a couple architecture transitions on the Mac before, so they know how it works. But they don’t know how app developers will react. Will Adobe, Microsoft etc. get onboard, or jump ship?
Effort: Rather high. It’s one of the things where 10% of the implementation will need 90% of the effort. Mostly it’s just a recompile to run. But optimizing software for x86 is different than for ARM. The performance critical parts of many system frameworks which have no existing iOS port have to be rewritten to be efficient on the new architecture.
- Comparing the brand new A10 to a 2013-era Intel.
- Asserting iOS for workstations.
While many people could probably get by on iOS, it is particularly crippled for people that need to do technical work.
However, I found the idea intriguing, of having a workstation with multiple ARM or Apple chips in it with a proper OS. I know they exist already, but having Apple behind them would make a big difference.
I sometimes wonder if iOS and OSX will end up merging and Apple notebook, iPhone and iPad hardware becoming more of the same thing, all running on Apple chips and the same appstore.
After I saw the benchmarks of Apple’s A10 Fusion chip and it was clear they continued focusing on desktop class performance in phones, my thoughts were:
1. Apple is working hard to make the switch from Intel to their own ARM chips and in a few short years we will see MacBooks and MacBook Pros with ARM architecture.
2. The iPhone becomes the new ultra portable desktop computing device. You plug it into a dock (maybe backside of an Apple 27" monitor) and boot up an ARM version of macOS from your iPhone.
> 2. The iPhone becomes the new ultra portable desktop computing device. You plug it into a dock (maybe backside of an Apple 27" monitor) and boot up an ARM version of macOS from your iPhone.
Why would they do that - Apple is very profit oriented and if they can sell you Apple monitor, Apple iMac, Apple iPhone and all the dongles to connect them together, that maximizes their profit.
Plugin your iPhone into a monitor is not a good way to extract maximum profit - well maybe if it'll only work with Apple monitor and nothing else.
Apple does not work in a vacuum: PCs still exist, Android still exists, and things like ChromeOS are competing in certain spaces. Apple is also different from most competitors because they also make money after the device purchase from purchases of app, music, video, ebook, etc.
If the question was just “iPhone or iPhone + MacBook?” there's no doubt they'd prefer the latter but if the question is instead “iPhone or iPhone + Windows PC” I have no doubt they'd throw the Mac under the proverbial bus if it would keep your money from going to a competitor.
I guess that Apple would be already able to produce laptop SoCs running at 5-15W TDP, with e.g. 4 to 6 CPU cores at 2-3GHz, outperforming newest Core i7 chips for laptops.
That could allow to maintain, or even increase, profit margings, with margin for lowering product price. Running that path, in my opinion, Apple could take 50% of laptop market. E.g. high quality laptops with BOM (bill of materials) of 150 USD selling for 400-600 USD. Or even less.
Somebody mentioned on here a few days ago that it's possible Apple would kill their laptop line entirely. I scoffed at the time, but this article has made me wonder. If the iPad offered the same screen, had better battery life, was just as fast if not faster and had a more stable OS... why wouldn't they kill their laptops entirely? Their laptop sales are a small and shrinking portion of their business.
Even if Apple doesn't pursue this for it's MacBook (I'm assuming they wouldn't for the Pro), I think it's a good indicator for where ARM is headed in general. As these things get faster than low power Intel chips and Google continues to add functionality to Chrome OS, like Android app support, it's likely to have a significant impact on Intels share of the PC market.
Nah, the only things that will hurt Intel are "good enough" SoCs out of Asia and possibly ARM in bulk in the server racks.
Apple's "special sauce" only remains special as long as it remains an Apple exclusive, and in no way will that be a Intel threat no matter how much the SV goes round the infinite loop...
As crazy as it might sound, perhaps the easiest way to gain a larger market would be some Xserve style server hardware with ARM chips. The typical (web-) serving software gets rebuilt quite often, so legacy code is a minor importance, and having lots of efficient processors is really a huge thing.
Surely the iPhone's new chip should worry professional and creative Mac users even more. The Mac Pro is now so outdated that the iPhone is catching up to it in single-threaded performance, and one wonders if macOS is much longer for this world.
I've said it before here: Intel allowing ARM chip makers to use its 10nm process is not an "automatic win" for the company, as many seem to believe it is.
If Apple is already able to match some of Intel's mainstream CPUs with an inferior process, what do you think is going to happen when Apple can use the same 10nm process Intel uses?
a) Apple embarrasses Intel with a much cheaper chip (I'm talking ~5x here) that can surpass some of its most powerful notebook chips, while both using the same process.
b) Apple replaces Intel in most of its Mac lineup with the exception of Macbook Pro (and even those chips could be replaced with cheaper AMD Zen chips soon).
c) Intel is ultimately chased to the more profitable "up-market" (typical innovator's dilemma), and loses control of the consumer market.
d) All of the above.
I hope Apple does switch to its own ARM chips for the Mac line. It's what Steve Jobs would've done and would've wanted, but it remains to be seen if Tim Cook is willing to follow in his footsteps to that degree.
For those still worrying that "Mac performance would suffer" - newsflash - Macbook Air has already been using the lower (than Core i3) performance "Core M5" for at least a generation. So if you have the latest Macbook Air, you're already benefiting from less performance than you did 2-3 years ago with Core i5.
It wouldn't surprise me one bit if Apple announces an A11-powered Macbook Air next year that is 30%+ faster than the previous "Intel Core" Macbook Air. It would also be a great and very smart way to launch the ARM-powered Macbooks because people would have little to complain about (other than perhaps a few initial compatibility issues, but I think they will get over those quickly as developers update their apps).
I see the iPad Pro gain more functionality and be the "computer" for most people. Its possible that slowing the development of Macs further may actually help future iPad Pro sales.
Off topic but I'm starting to think people find any excuse to use the word "portmanteau". Maybe that should be a new law like betteridges law. (Wikipedia is littered with it)
Shifting any engineering talent towards laptop/desktop chips is fighting last decade's war, for a shrinking (relatively) pie.
The only reason to do that would be if it gave them some research edge that they could use elsewhere. Like cloud servers or something.
As the article says, already more iphones than PC's. Why bother.
The way to use this power is to keep pushing where they have the advantage (mobile) and hammer PC's until you have all the power you need in your phone/tablet. Software will grow up around the ecosystem until it eats PC's, and Apple can ignore the PC market entirely. Forget about PC's, focus on beating Android and they'll beat PC's as a bonus.