The first device containing Apple's own chip was the iPad in 2010. At that time, the A4 processor contained a single core and its performance cannot be compared to today's generation at all. For five years, there have also been rumors about the integration of these chips into Mac computers. As mobile chips rapidly increase their performance every year, their deployment on desktops is a very interesting topic.
The previous year's 64-bit A7 processor was already labeled as "desktop-class", meaning that it is more like large processors than mobile ones. The latest and most powerful processor - A8X - was put into the iPad Air 2. It has three cores, contains three billion transistors and its performance is equivalent to the Intel Core i5-4250U from the MacBook Air Mid-2013. Yes, synthetic benchmarks do not say anything about the real speed of the device, but at least they can mislead many that today's mobile devices are just polished ink with a touch screen.
Apple really knows its own ARM chips, so why not equip your computers with them too? According to analyst Ming-Chi Kuo from KGI Securities, we could see the first Macs running on ARM processors as early as 2016. The first capable processor could be the 16nm A9X, followed by the 10nm A10X a year later. The question arises, why should Apple decide to take this step when processors from Intel are steaming to the top?
Why ARM processors make sense
The first reason will be Intel itself. Not that there is anything wrong with it, but Apple has always followed the motto: "A company that develops software should also make its hardware." Such a state has its advantages - you can always optimize both software and hardware to the highest level. In recent years, Apple has demonstrated this directly.
It's no secret that Apple likes to be in control. Shutting down Intel would mean simplifying and streamlining the entire production process. At the same time, it would reduce the cost of manufacturing chips. Although the current relationship between the two companies is more than positive - you would rather not rely on each other when you know that you can produce the same thing at a lower cost. What's more, you would manage all future development entirely yourself, without the need to rely on a third party.
Maybe I made it too short, but it's true. In addition, it would not be the first time that a change of processor manufacturer would occur. In 1994 it was the transition from Motorola 68000 to IBM PowerPC, then to Intel x2006 in 86. Apple is definitely not afraid of change. 2016 marks 10 years since the switch to Intel. A decade in IT is a long time, anything can change.
Today's computers have enough power and could be compared to cars. Any modern car will take you from point A to point B without any problems. For regular riding, buy the one with the best price/performance ratio and it will serve you well at an affordable cost. If you drive often and further, buy a car in a higher class and perhaps with an automatic transmission. However, maintenance costs will be slightly higher. Off the road, you can certainly buy something with a 4x4 drive or a straight off-road car, but it will get used regularly and the costs of its operation will be high.
The point is that a small car or a car of the lower middle class is fully sufficient for most. Analogously, for most users, an "ordinary" laptop is enough to watch videos from YouTube, share photos on Facebook, check e-mail, play music, write a document in Word, print a PDF. Apple's MacBook Air and Mac mini are designed for this kind of use, although they can of course be used for more performance-demanding activities.
More demanding users prefer to reach for a MacBook Pro or an iMac, which after all have more performance. Such users can already edit videos or work with graphics. The most demanding of the demanding reach for uncompromising performance at an appropriate price, i.e. the Mac Pro. There will be an order of magnitude fewer of them than all the other mentioned models, just as off-road cars are driven far less than Fabia, Octavia and other popular cars.
So, if in the near future Apple will be able to produce an ARM processor such that it would be able to satisfy the needs of its (at first apparently less demanding) users, why not use it to run OS X? Such a computer would have a long battery life and could apparently also be passively cooled, as it is less energy-intensive and does not "heat" as much.
Why ARM processors don't make sense
Macs with ARM chips may not be powerful enough to run a Rosetta-like layer to run x86 applications. In that case, Apple would have to start from scratch, and developers would have to rewrite their apps with considerable effort. One can hardly argue whether developers of mainly popular and professional applications would be willing to take this step. But who knows, maybe Apple has found a way to make x86 apps run smoothly on "ARM OS X".
The symbiosis with Intel works perfectly, there is no reason to invent anything new. The processors from this silicon giant belong to the top, and with each generation their performance increases with lower energy consumption. Apple uses a Core i5 for the lowest Mac models, a Core i7 for more expensive models or a custom configuration, and the Mac Pro is equipped with very powerful Xeons. So you will always get enough power, an ideal situation. Apple could find itself in a situation where no one wants its computers when it breaks up with Intel.
So how will it be?
Of course, no one outside knows that. If I were to look at the whole situation from Apple's point of view, I would certainly like it once similar chips were integrated into all my devices. And if I am able to design them for mobile devices, I would like to practice the same for computers as well. However, they are doing great at the moment even with the current processors, which are stably supplied to me by a strong partner, although the release of the upcoming new 12-inch MacBook Air may have been delayed precisely because of Intel's delays with the introduction of the new generation of processors.
Can I bring powerful enough processors that will at least be at the level of those in the Macbook Air? If so, will I later be able to deploy (or be able to develop) ARM also in professional computers? I don't want to have two kinds of computers. At the same time, I need to have the technology to run x86 applications on an ARM Mac, because users will want to use their favorite applications. If I have it and I'm sure it will work, I'll release an ARM-based Mac. Otherwise, I'll stick with Intel for now.
And maybe it will be completely different in the end. As for me, I don't really care about the type of processor in my Mac as long as it's powerful enough for my work. So if a fictional Mac contained an ARM processor with performance equivalent to a Core i5, I wouldn't have a single problem not buying it. What about you, do you think Apple is capable of launching a Mac with its processor in the next few years?
ARM means throwing away 99.9% of usable software and using the computer only for reading email, Facebook and Angry birds (yes, I'm willingly exaggerating), and I hope no one who isn't a total dick will allow that. Switching to Intel simplified things a lot. After all, quite a few people in the scientific sphere go to Macy. That's my opinion.
Moving to ARM means a major loss of compatibility and an even greater squeeze into a tightly guarded ecosystem of pre-selected applications. Add to that the plans to remove the USB port from the Macbook Air and other ideas, and the result is a computer that will only be compatible with itself, and maybe even then. For me it would mean the end of MAC as I like it:-( But who knows, change is life...
Why have the weakest iMac and mini in Intel when their performance is just as weak as the last Generation?
For me, to test at least these models and go in a similar direction as with Retina, a gradual flooding of the product line.
You're kind of forgetting about the upcoming generation of intel Core M processors…
history repeats?
I believe it will come very soon. I wouldn't be surprised if it was the expected air. It's a device for the masses who don't do much work, but rather have fun and a coffee. I can also imagine some kind of emulator for x86 and since apple is in, most developers rush to be the first to remake their software for the new hardware. Flusbrok will go there so why not? :-)
I think the only drawback would be incompatibility with the old software.
An advantage is, for example, a possible drop in price. Even for the cheapest Macs, an Intel processor with a chipset costs from USD 250. Apple can make A8, A9, ... processors under 50 USD.
Another benefit would be to use the Metal graphics API on Macs. The performance of the quad-core A8 with current graphics would be somewhere on the level of the latest game consoles.
Basic versions of macs could have ARM processors, and the most powerful macs could still stay on Intel processors. Apple has its own app store and its own developer environment with its own translator. Programs that currently only use the standard API provided by Apple would be easy to translate for ARM processors.
Later, the most powerful versions of Macs could have ARM and Intel processors. OS X and iLife and iWorks programs would run on ARM processors and Photoshop, Premiere, Final Cut and other programs would still run on Intel processors.
Unfortunately, the disadvantage would also be incompatibility with the new software.
Why should there be incompatibility with new software? I write a program for OS X, compile it for Intel and ARM and send it to the App Store. If I only use what the API offers from Apple, then there can be no problem with that. If I use something special for an Intel processor, of course it won't work. Apart from professional software that requires high performance, there is no reason for this.
The user launches the App Store, downloads the version of the program for his processor and uses it.
I once read an article that when they tried OS X on an Intel processor, they bought the most expensive Sony Vaio they could find in the store, and it worked in two days. In two weeks, even the basic programs from Apple came.
The main question is to what extent Apple would be able to ensure the running of x86 programs on ARM. Especially for company use, at least the occasional use of something from Windows is essential. So if x86 Windows virtualization could work on ARM without any noticeable performance loss, apparently users wouldn't really care what processor is physically there as a result. But I'm afraid it wouldn't do without the need for emulation and the associated significant performance drop. But who knows. For example, Cupertino managed to find an ideal way of emulating the processor architecture in such a way as to preserve the high performance and energy efficiency of such a solution.
Apple doesn't have to secure it. He already has his own app store. It is enough to ensure that developers have time to adapt their programs to ARM, which should not be a problem. And at some point I will definitely be selling Macs with Intel processors.
There should be a new ARM core next year and the performance should be 3.5 times higher than the current one, so emulation could be overcome.
I can imagine that next year at WWDC, Apple will introduce OS X on ARM processors, offer developers a mac mini with an ARM processor to test their programs, and sometime half a year before Christmas, they will start selling basic versions of the mac mini, macbook air and iMac with ARM processors. And a year later, Intel will leave processors only in the 15″ macbook pro, 27″ iMac and Mac Pro. And sometimes in 2-3 years Intel will leave the processor only in the Mac Pro.
Why the incompatibility? Windows 10 will now also run on ARM. Trď is a new version of the RasPi 2 with an ARM SoC from BroadCom, and Windows is supposed to run there. So I see no reason why not ARM on the Mac. IBM PowerPC applications were also rewritten and compiled to x86 instructions when Apple started with Intel. For Apple, the ARM architecture would be more advantageous in the direction of further miniaturization and lighter, more subtle machines, and with bigLITTLE technology, even with more cores, the battery would last longer. It would use its same chips across iPhones, iPads, Macs, Apple TVs, and everything else it has on offer now.
Especially so that they can then optimize so well when running on different HW.