iOS 10, OS X 10.12, Apple A10 chip, and where this is all headed

[HatMine]

…which does absolutely nothing for thousands of legacy applications, of which the vast majority is not written in Swift.

Stating the obvious. If there's something Apple is good at, then that would be to swiftly change route. First with the CD/DVD-drive, then with flash (on iOS), and now with the headphone jack. How many headphones won't be incompatible with the new iPhone 7? By taking a step forward in a new direction, and focusing on it wholeheartedly, they sooner or later have most people by their side.

Swift is already being used for apps in the Mac & iPhone AppStores, and Apple will undoubtedly make sure that all standard-apps in macOS run on the ARM-mac. What's more is that they will of course work with big developers to port their apps to Swift if they haven't done so already. Lastly, I didn't say that they will do this NOW. The only date I have mentioned is Q3 2017, and that's at the very best.

 

[leman]

Objective-C was being compiled to ARM ages before Swift made an appearance and while Swift might make cross-platform development somehow simpler, I am fairly sure that most Objective-C (or C++/C/etc.) code will compile to ARM without much of a problem. People don't do architacture-dependent code that much these days.

 

[HatMine]

Objective-C was being compiled to ARM ages before Swift made an appearance and while Swift might make cross-platform development somehow simpler, I am fairly sure that most Objective-C (or C++/C/etc.) code will compile to ARM without much of a problem. People don't do architacture-dependent code that much these days.

That's a fair point, although I'm not sure that C++ code can be compiled to run on ARM (citation needed). Objective-C however, should work with ease.

 

[leman]

That's a fair point, although I'm not sure that C++ code can be compiled to run on ARM (citation needed). Objective-C however, should work with ease.

Both gcc and clang can compile to a number of targets, of which modern ARM and x84 are only a very very small subset. The question is whether your code makes certain assumptions about the architecture. And in case of modern iOS, even that doesn't matter much as the endianness, as well as sizes and alignment of basic types are the same between ARM64 and x86. Basically, what this means is that if your C/C++/etc. code is 64-bit clean (works with 64bit OS X) and doesn't use CPU-specific intrinsics/assembly, then it should compile to ARM64 without any hiccups.

 

[SlCKB0Y]

If there's something Apple is good at, then that would be to swiftly change route. First with the CD/DVD-drive, then with flash (on iOS), and now with the headphone jack. How many headphones won't be incompatible with the new iPhone 7?

Yes but they always provide a bridge to support people. They produced the external optical drive, a 3.5mm headphone adaptor and when they switched to Intel, Rosetta. This will not be possible with a switch to ARM in the timeframes talked about.

 

[SimonTheSoundMa]

I can't see Apple moving their Intel based machines to ARM anytime soon. ARM is roughly a decade behind x86 in terms of performance, ARM was faster than x86 in the 1990's on Acorn manchines but that was another time.

What I'd really like to see is Apple start to introduce their own DSPs like they do with their ARM chips to the x86 platform or utilise some that are already there and not used. Audio, graphics manipulation etc. For example of an unused DSP, Thunderbolt 2 controllers can do h264 video conversion, there is 512MBs of memory on rMBPs logic boards attached to it and its only real use is in Airplay and TCP/IP data over Thunderbolt. iPhone's are speedy at video conversion, why not use this hardware for application level tasks such as in iMovie.

 

[jpn]

They could literally just take an Ipad Pro, attach a keyboard to it, and call it a Mac. The one thing holding them back is the fact that they would need to make macOS run on ARM, and all which that means.

But thanks to swift, which can compile apps to run on ARM AND x86, the app compatibility isn't as big of a problem as it could've been. Like I've said before, I wouldn't be surprised if we saw an ARM MacBook in 2017 or 2018.

However, what's more likely to happen is that they start to make iOS even more like macOS (feature wise), and continue to place their hopes on the Ipad Pro.

i think this is the going to be the actual evolution.

and macbook air could be the 1st machine that this is reserved for.

 

[Toutou]

For the price of one Four core Xeon, Apple could put ten A9Xs in to a desktop PC. That would be twenty core beast with ~50W TDP.

lets say that there is going to be a four, six and eight core versions of next A10 chip. That will definetelly rival Intel.

This might surprise you, but shoveling more cores into a computer doesn't necessarily make it faster. Only for very specific tasks that are easy to parallelize. Otherwise it's just wasting space and power.

 

[grahamperrin]

… APFS will be used for both iDevices and Macs (+ Apple TV and Apple Watch, I believe) …

https://forums.macrumors.com/posts/23234541 notes Apple's plan for APFS to be the default file system for "all" Apple products, and my uncertainty about Time Capsule.

 

[dogslobber]

You can guarantee that OS X runs on ARM. There is no doubt whatsoever. In fact, it would be reckless for them to NOT have ARM running OS X right now.

 

[lysingur]

Firstly, Apple didn't switch because of Pentium 4, they switched because of the future roadmap of Intel vs. that of PPC along with the massively more energy efficient processors.

The speed claims I mentioned were those cited by Apple when they released their Core Duo iMac and Macbook Pro at Macworld 2006.



Estimates put total number of Windows computers and servers at about 1.5 billion and iOS + OS X at just over 1 billion.



I'm not sure what you mean by saying the software is converging since they are both based on the same underlying OS. I would argue that the two platforms are undergoing integration not convergence, and this allows them to work much more harmoniously together.



It's not even remotely possible. The highest end of ARM currently performs in line with the lowest end of Intel. It won't come close to matching the high end of Xeons within your 3 year period, nevermind the multiples it will need in speed in order to emulate Intel (unless you think Apple would switch without compatibility).

Even after you consider all this, and even if ARM can massively scale up in terms of speed, there is no guarantee they can do this and keep all the power efficiences.

https://www.macrumors.com/2016/09/15/iphone-7-faster-than-macbook-air/

Let's just wait for another year or two.

 

[lysingur]

It's not even remotely possible. The highest end of ARM currently performs in line with the lowest end of Intel. It won't come close to matching the high end of Xeons within your 3 year period, nevermind the multiples it will need in speed in order to emulate Intel (unless you think Apple would switch without compatibility).

Even after you consider all this, and even if ARM can massively scale up in terms of speed, there is no guarantee they can do this and keep all the power efficiences.

It seems like it won't even take longer than a year before we see a Mac with a fully functional ARM chip in it. So I'd say "remotely possible" is putting it too strongly, isn't it?

Apple has to be really confident about their own ARM's R&D to be putting it in the Pro lineup — this takes care of your worry that ARM needs to scale up in terms of speed.

The switch, like I said in my earlier posts, is already well underway and it'll happen within the 3-year timeframe that I talked about. It's also very obvious that the switch from NVIDIA to AMD in MBP 2016 has something to do with Apple's plan to transition away from Intel. AMD is probably working on a graphics card specifically for Apple's ARM as we speak.

 

[Janichsan]

It seems like it won't even take longer than a year before we see a Mac with a fully functional ARM chip in it. So I'd say "remotely possible" is putting it too strongly, isn't it?

SlCKB0Y wasn't referring to the possibility of an ARM CPU in a Mac in general, but about the likelihood of having an ARM CPU that competes with high-end Intel CPUs in regard to performance anytime soon.

And that is in fact not even remotely possible.

Apple has to be really confident about their own ARM's R&D to be putting it in the Pro lineup — this takes care of your worry that ARM needs to scale up in terms of speed.

Honestly, have to taken a closer look at what Apple calls "Pro" devices recently? It's more than enough reason to be very, very worried…

The switch, like I said in my earlier posts, is already well underway and it'll happen within the 3-year timeframe that I talked about. It's also very obvious that the switch from NVIDIA to AMD in MBP 2016 has something to do with Apple's plan to transition away from Intel. AMD is probably working on a graphics card specifically for Apple's ARM as we speak.

Unfortunately, I believe you are correct. The day Apple will cease to be a computer manufacturer that could be taken seriously comes closer…

 

[lysingur]

SlCKB0Y wasn't referring to the possibility of an ARM CPU in a Mac in general, but about the likelihood of having an ARM CPU that competes with high-end Intel CPUs in regard to performance anytime soon.

He was actually replying to my prediction that there'll be an ARM-based Mac in three years (#22). At least that's the premise of the discussion. If you've read the entirety of the thread, you'll see that we weren't talking about high-end Intel CPUs, at all.

 

[Janichsan]

He was actually replying to my prediction that there'll be an ARM-based Mac in three years (#22). At least that's the premise of the discussion. If you've read the entirety of the thread, you'll see that we weren't talking about high-end Intel CPUs, at all.

Apparently in contrast to you, I have read the whole thread. That's why I see the fault in your line of argument starting from your claim that the "only thing left" for a transition (implicitly of the whole Mac product line, including desktops) would be an ARM CPU "powerful" enough also for actually professional applications, to SICKB0Y's assertion that the performance of ARM CPUs might compete with low-end Intel CPUs, but not with higher-end variants (like the Xeons he explicitly mentions), to you waving triumphantly an article about an ARM based co-processor for extreme low power (and low performance) tasks that is working alongside an still existing Intel CPU.

 

[wrldwzrd89]

This is a very interesting direction this topic has taken... I must admit. On a related note, Apple's got too big of a marketing opportunity with iOS 11 and tvOS 11 coming... not to turn macOS (and heck, watchOS too) to 11 as well. Thinking of some ideas for marketing slogans: "Quad-Core. To the eleventh power." or perhaps "We turned everything up to 11." This second idea sounds more like something Apple might do, at least to me.

 

[Pit bull puppy]

I cannot see Apple switching to an ARM chip for the laptop line up, wouldn't it mean a complete redesign of the main boards and maybe even the case design. i also cannot see Apple switching to AMD processors since they seem to like to run very hot and burn up.

 

[wrldwzrd89]

I cannot see Apple switching to an ARM chip for the laptop line up, wouldn't it mean a complete redesign of the main boards and maybe even the case design. i also cannot see Apple switching to AMD processors since they seem to like to run very hot and burn up.

Regarding AMD, their upcoming Ryzen line looks to have solved the power consumption problems, and could even beat Intel in certain situations. We shall see, once these are released. Like you, I'm not confident yet.

 

[cynics]

It seems like it won't even take longer than a year before we see a Mac with a fully functional ARM chip in it. So I'd say "remotely possible" is putting it too strongly, isn't it?

Apple has to be really confident about their own ARM's R&D to be putting it in the Pro lineup — this takes care of your worry that ARM needs to scale up in terms of speed.

The switch, like I said in my earlier posts, is already well underway and it'll happen within the 3-year timeframe that I talked about. It's also very obvious that the switch from NVIDIA to AMD in MBP 2016 has something to do with Apple's plan to transition away from Intel. AMD is probably working on a graphics card specifically for Apple's ARM as we speak.

I didn't see any mention of x86 in that article, maybe I missed it. Seems mostly speculation.....

Until Apple can get around that we won't see an ARM only Mac. No one is going to buy a computer if it only supports 1st party software.

 

[lysingur]

Apparently in contrast to you, I have read the whole thread. That's why I see the fault in your line of argument starting from your claim that the "only thing left" for a transition (implicitly of the whole Mac product line, including desktops) would be an ARM CPU "powerful" enough also for actually professional applications, to SICKB0Y's assertion that the performance of ARM CPUs might compete with low-end Intel CPUs, but not with higher-end variants (like the Xeons he explicitly mentions), to you waving triumphantly an article about an ARM based co-processor for extreme low power (and low performance) tasks that is working alongside an still existing Intel CPU.

Where you're wrong is I never said anything about "professional applications". An ARM chip simply needs to be powerful enough to satisfy most users, i.e. the non-Pros, for the switch to be feasible. This has been my argument since the very beginning. How you missed it after having "read the whole thread" is anybody's guess.

The article is simply to point out that that's the direction Apple is taking. SICKBOY's position is that the switch won't happen, at least in the three-year timeframe that I talked about. But, again, as I said earlier, it's already happening and the article is just more proof to that.

You only see it as my waving an article "triumphantly" because you got your feelings hurt.

I didn't see any mention of x86 in that article, maybe I missed it. Seems mostly speculation.....

Until Apple can get around that we won't see an ARM only Mac. No one is going to buy a computer if it only supports 1st party software.

Nobody is going to buy it now. But three years from now? It seems like a real possibility. With the kind of leap cloud services and mobile network technology is making, most things can be done over the cloud by that time anyway. Software development is already less platform-dependent than ever before. This trend will continue.

Also, not to nitpick but ARM-based is still slightly different from ARM-only. Apple could for compatibility reason add a modified Intel chip but let the ARM does the heavy lifting, i.e. be the main chip the OS runs on.

 

[Janichsan]

Where you're wrong is I never said anything about "professional applications". An ARM chip simply needs to be powerful enough to satisfy most users, i.e. the non-Pros, for the switch to be feasible. This has been my argument since the very beginning.

Your quote, my emphasis:

iPad Pro itself lacks the user base to induce developers to write serious professional grade apps specifically for what has been marketed as a Pro tablet but with Macs switching to ARM, developers will finally have a good reason to write apps for both.

How you missed it after having "read the whole thread" is anybody's guess…

SICKBOY's position is that the switch won't happen, …

…because ARM CPUs don't have the necessary performance to actually replace Intel CPUs, as he has stated several times.

But, again, as I said earlier, it's already happening and the article is just more proof to that.

No, it isn't. Sorry, but you really do not seem to have understood that article. To say it again: the article is about little more than an ARM based co-processor for some few specialised tasks, similar to the one already in the current Touch Bar rMBPs. This is still a far cry from having a completely ARM based Mac.

 

[lysingur]

How you missed it after having "read the whole thread" is anybody's guess…

Except the angle you're arguing from is computational power (that ARM chips won't be powerful enough to run professional applications) but what I was arguing when I made that comment is economic viability. While many see the iPad Pro as an awkwardly placed product, notably vis-à-vis the MacBook Air, I see it as the direction toward which Apple would like to take the future of computing. The lack of professional applications on the iOS, especially for the iPad Pro, must have been something Apple expects in the short-term and will rectify in the future. In other words, I already take for granted that (Apple is confident that) ARM will be powerful enough to run professional applications but what you're arguing is professional applications are the reason that the switch won't happen.

Granted, I did say something about professional applications, but not how you understood it. What you failed to understand is that not all professional applications need a powerful chip. Final Cut Pro isn't the only pro app out there. So the switch-based-on-computational-power argument fails because it doesn't take into account that economic viability is the determining factor in any major switch of CPU architecture. ARM chips have economic viability because 1) they already meet the needs of most, i.e. casual, users, 2) have a strong platform to grow from, i.e. iOS, and 3) it's Apple's own chip.

Another thing you fail to see is things are moving to the clouds, including power hungry applications such as games and video editing programs. All this makes the switch to ARM less painful than the switch from PowerPC to Intel because extra computational power can be acquired elsewhere.

The lack of professional applications on the iOS platform only figures in my line of argument to the extent that the switch to ARM on the Mac is already part of Apple's plan. They introduced the iPad Pro as a sort of pioneer product knowing full well that ARM-based Macs combined with iPad Pro will further grow the ARM-based ecosystem and finally induce those recalcitrant few to port professional applications for both.

Finally, if you've been paying attention, you'd know that Intel just can't do it at the level they were capable of anymore. Apple is going to have to come up with something so Macs keep getting faster. ARM chips might not be there right now but the room and potential for speed increase are better on the ARM side.

Expect to see more attention paid to the iPad Pro in the coming years.

o, it isn't. Sorry, but you really do not seem to have understood that article. To say it again: the article is about little more than an ARM based co-processor for some few specialised tasks, similar to the one already in the current Touch Bar rMBPs. This is still a far cry from having a completely ARM based Mac.

I understand the article perfectly. There is now an ARM-based chip in a Mac.

 

[Janichsan]

Finally, if you've been paying attention, you'd know that Intel just can't do it at the level they were capable of anymore. Apple is going to have to come up with something so Macs keep getting faster. ARM chips might not be there right now but the room and potential for speed increase are better on the ARM side.

Well, and if you had been paying attention, you'd know the reason Intel can't increase the performance of their chips much more is not because of faults of their architecture, or simply unwillingness, but because they are approaching the limits of what is physically possible with the current silicon based technology. And you might have noticed that Apple's CPUs operate with the same technology. Could the performance of ARM CPUs increase to a level comparable of high-end Intel CPUs? Surely. But not much further, because they would run into the very same problems.

 

[lysingur]

Well, and if you had been paying attention, you'd know the reason Intel can't increase the performance of their chips much more is not because of faults of their architecture, or simply unwillingness, but because they are approaching the limits of what is physically possible with the current silicon based technology. And you might have noticed that Apple's CPUs operate with the same technology. Could the performance of ARM CPUs increase to a level comparable of high-end Intel CPUs? Surely. But not much further, because they would run into the very same problems.

The fact that Si-based chips are hitting the physical ceiling just lends more weight to the idea that Apple should switch since Intel chips are pretty much the only thing that stands in the way of a fully optimised user experience on the Mac. The engineers at Cupertino surely understand that we're stuck with Si for at least 4 to 5 years before the next new material comes along that can be productised so opting for ARM architecture allows them to squeeze every flop out of this technology. Low power consumption of the ARM chip probably means Apple can achieve comparable computational power in the short run by using multiple cores (>12) while further improving battery life (since more cores can be turned off). In other words, T310 is just the beginning.

The iOS is proof of what can be achieved when every aspect of an ecosystem is controlled by a single company. The advantages are both economic and technological. It only makes sense that Apple would want to extend such advantages over to the macOS and the Mac. They're already making the macOS experience more like iOS with every major OS update and the groundwork for further integration has already been laid, e.g. Swift, Metal, and APFS. Imagine how much streamlining can be had if the whole Apple ecosystem is based on one single CPU architecture (mind you Android apps can soon be run on Chrome OS). Apple would be stupid not to make the switch.

 

[lysingur]

https://www.bloomberg.com/news/arti...an-move-from-intel-to-own-mac-chips-from-2020

Swift, Metal, APFS, USB-C/Thunderbolt 3, and most importantly ARM-based CPU. Cook's master plan is now clear - an Apple ecosystem revolving around not just personal computers but a digital mobile lifestyle. This explains why Macs seem to have diminished in importance in recent years and why a switch to ARM-based chips makes perfect sense. It is simply too costly to maintain a platform that generates much less profit but is based a completely separate CPU architecture. It is very likely that with Apple's embracing of a wireless future, Lightning port would eventually be phased out and free iCloud storage augmented.

I can't say I'm not excited for this future. Less cable clutter has already been realized with the advent of USB-C/Thunderbolt 3. By the time AirPods 3 rolls around, not only would there be one less cable to deal with, their sound quality should be indistinguishable from wired earphones that we have today. Multi-core ARM chips (>12) means more energy efficiency and much longer battery life. Already cloud services are integrated across Apple's entire product line and now a complete integration of CPU architecture of smartphones, tablets, laptops and desktops by 2020, Apple surely is playing the long game.