Apple to Launch Arm-Powered MacBook in the next 18 Months

[Darmok N Jalad]

I don’t see Apple struggling with an ARM MacBook. I get the feeling they are going this way to source better ULV chips for thinner, simpler designs, especially since Intel has really struggled to provide such chips.

As for software, I’m curious if Apple will depend more on ports from iOS instead of x86. They’ve already laid out the framework for apps to bridge the gap, and external storage and primitive mouse support was added in iPadOS 13. We own a few iPads, and I could totally see the value of a MacBook that sits in the gap. Price is the mystery. Will Apple price this to be an entry level device so it can work in education? I bet Apple can source its own chips for cheaper than what Intel can provide, and they can reuse their chip in iPad as well. It’s probably not going to be a pro device, so I can see it being cheaper than iPad Pro.

 

[notb]

I'm willing to bet this rumour is correct. Apple has been converging its phone/tablet and desktop operating systems while simultaneously putting massive resources into its ARM implementation for the last half-decade.

It's likely that they'll add something to their current product lines (overpriced laptops/overpriced workstations): overpriced Apple equivalents of the Chromebook. After all, the majority who buy Apple laptops don't intend to use them for anything more than productivity, and MS already has Windows for ARM which means they have Office for ARM. No need for x86/x64 emulation if you're using the thing as a status symbol more than doing actual work on it.

EVEN IF Apple somehow manages to supply all the most popular "productivity" programs (that's hundreds of large suites - who will rewrite them?), that's just half of the problem.
Macs are very popular among programmers - something Apple is extremely proud of and will cherish.

Problem: x86 virtualization.
I just can see this happening together with ARM-based coding software.

Which means they would have to emulate the whole workflow. So you have that shiny ARM-based MacBook, but everything you use is emulated for x86. Disaster - on both stability and performance/efficiency. So you're losing all that ARM was meant to bring, while also compromising stability and simplicity, which are Apple's key selling points.

I don’t see Apple struggling with an ARM MacBook. I get the feeling they are going this way to source better ULV chips for thinner, simpler designs, especially since Intel has really struggled to provide such chips.

No offense, but have you seen how Intel-based ultrabooks look in 2019/2020?

As for software, I’m curious if Apple will depend more on ports from iOS instead of x86. They’ve already laid out the framework for apps to bridge the gap, and external storage and primitive mouse support was added in iPadOS 13. We own a few iPads, and I could totally see the value of a MacBook that sits in the gap. Price is the mystery. Will Apple price this to be an entry level device so it can work in education? I bet Apple can source its own chips for cheaper than what Intel can provide, and they can reuse their chip in iPad as well. It’s probably not going to be a pro device, so I can see it being cheaper than iPad Pro.

But why why why?
Why would Apple want to make a cheap laptop for education?
Why would they want to put a product on a new price level?
It all makes very little sense financially. And Apple is a high margin company.

This ARM MacBook will either come in the MacBook Pro lineup (extremely unlikely) - where it will cost exactly as much as x86 MacBooks, or in the MacBook Air lineup - where it'll cost as much as Air does now.
Apple's product lineup is already very cramped (for their standards). Adding more products seems unlikely.

 

[Darmok N Jalad]

EVEN IF Apple somehow manages to supply all the most popular "productivity" programs (that's hundreds of large suites - who will rewrite them?), that's just half of the problem.
Macs are very popular among programmers - something Apple is extremely proud of and will cherish.

Problem: x86 virtualization.
I just can see this happening together with ARM-based coding software.

Which means they would have to emulate the whole workflow. So you have that shiny ARM-based MacBook, but everything you use is emulated for x86. Disaster - on both stability and performance/efficiency. So you're losing all that ARM was meant to bring, while also compromising stability and simplicity, which are Apple's key selling points.

No offense, but have you seen how Intel-based ultrabooks look in 2019/2020?

But why why why?
Why would Apple want to make a cheap laptop for education?
Why would they want to put a product on a new price level?
It all makes very little sense financially. And Apple is a high margin company.

This ARM MacBook will either come in the MacBook Pro lineup (extremely unlikely) - where it will cost exactly as much as x86 MacBooks, or in the MacBook Air lineup - where it'll cost as much as Air does now.
Apple's product lineup is already very cramped (for their standards). Adding more products seems unlikely.

Apple has many updated “pro” Macs now, from MBP to IMac Pro to Mac Pro. The bottom is not as solid. They ended the MacBook, which was the passively cooler model. The Air moves up a level and has a fan. Intel has good ultrabook chips, but those are U-series. The Y-series has never seemed to be a good fit in the MacBook, as one criticism was how hot it got. This is where Apple is probably looking for their opening, as all their chips are passively cooled. This model may never be intended for a programmer, but it’s a push for consumer dollars. Sure, education may be a long shot, but I don’t know how threatened Apple is feeling with Chromebooks in schools. It is where the future buyers are, and maybe the iPad just isn’t cutting it. I guess we have a long wait to find out. This product has been rumored for a few years now, so who knows.

 

[R0H1T]

But why why why?
Why would Apple want to make a cheap laptop for education?
Why would they want to put a product on a new price level?
It all makes very little sense financially. And Apple is a high margin company.

Oh the irony, why do you think Apple is switching to ARM ~ hint it's not about x86, it's all about APPLE! Apple loves margins & them switching over to their own chips will increase their margins, in case you also didn't notice they've bough the entire flailing Intel 5G division & will use it for their 5G phones after the QC deal expires. They've done the same with displays (buying stake in Sharp) as well as everything else where it makes financial sense.
There was a report in WSJ or Forbes ~ By Moving Creative In-House, Could Apple Be A Sign Of Things To

And I saw this coming what 6~7 years back when the rumors first started appearing, although it's taken at least a couple of years longer than some of us anticipated.

 

[phanbuey]

apple is by no means the only one moving in this direction:

Apple and MS will have competing ARM lines in the consumer / prosumer space very soon. A fight is coming!

 

[TheGuruStud]

I don't think Apple will be that foolish. Whether the performance and limitations of such emulation layer are acceptable enough is a completely different matter altogether. I think it will be like Rosetta, no rebooting necessary.

It does come to my mind that Adobe users will probably not worry a lot, since Creative Cloud is always running the latest version. Hence, Adobe will probably work on an ARM version of their software suite, which could easily get switched from x86 to ARM during a software update.

For bootcamp, though... Is it really that common to use a non-Apple OS on Macs?



Just like it happened when Macs ran on PowerPC.

Well, I've seen people buy MacBook airs and install Windows 7 since osx sucks lol

 

[efikkan]

By numbers only, the new ARM CPUs designed for MAC is likely outperforming the Intel CPUs Macbook Pros are using:
iOS: https://browser.geekbench.com/ios-benchmarks/
Mac: https://browser.geekbench.com/mac-benchmarks/

This is going to be interesting, though being successful or not depends a lot on the software side.

Benchmarks like this are not comparing "apples" to "apples". Geekbench is known for using varying accelerated features, which makes it pointless as an indicator for "general performance".

I think it'd make a lot of sense for Apple to team up with AMD. Having mixed CPU support (ARM and x86) would make a lot of sense and only AMD can provide both.

AMD have not been successful in their ARM ventures. Don't forget that Zen was intended as a "stop-gap" before K12 was supposed to take over the PC market. K12 is now MiA and its ISA already obsolete, and AMD is instead focusing on improving Zen.

I'm interested on how Apple can implement their own x86-emulation with this. I know they've got Metal for former OpenCL-code on the GPU, but how about for native apps that rely on x86 instructions? If they can make it better than what Qualcomm+Microsoft currently got with the Surface Pro X and implement proper AMD64-emulation, they can see big success with this.

Just forget emulation.
A powerful x86 CPU struggles with emulating a tiny cellphone CPU.
Apple really don't care about compatibility, they are going in the direction of a more closed ecosystem.

The problem is ARM. Personally I don't want a personal computer with an ARM as the main processor. ARM Ltd is just another layer of pricing and control. We're better off with x86_64 imho. ARM is no longer simple. It is more complex than x86 in many ways.

The biggest problem is the fact that ARM can't compete with x86 performance wise. To become competitive, it would have to become more CISC-like. Mobile devices with ARM CPUs rely heavily on accelerated features(through ASICs) while the ARM cores are very weak. This is great for energy efficiency, but gives very poor general performance and the hardware quickly becomes obsolete. Applications also have to be written specifically for each ISA to get the benefits from accelerated features.

But the problem you describe is also a big problem. There are many very customized ARM designs. The OS' needs a custom firmware to even be able to run on them, and writing applications for various ARM designs is a maintenance nightmare (if they want accelerated features).

EVEN IF Apple somehow manages to supply all the most popular "productivity" programs (that's hundreds of large suites - who will rewrite them?), that's just half of the problem.
Macs are very popular among programmers - something Apple is extremely proud of and will cherish.

Recompiling for ARM itself wouldn't be hard in most cases, but performance would suck, and performance critical ones would have to be rewritten as you said. Even worse, as the ISA keeps changing with nearly every new Apple chip, just imagine the workload for those poor software developers to keep up with Apple's ever-changing eco-system™.

My impression is that Mac usage among programmers is dropping quickly, it's nowhere like it was 5-10 years ago. Today Dell, Lenovo and HP have much better options, and as more and more developers tend to go for Linux as Apple keep selling outdated machines and lowering build quality, I think Macs soon will no longer be relevant for any of the "professionals".

If Apple's current trend continues, with their products becoming more a "fashion statement" than something for productivity, lower build quality to encourage people to upgrade quickly, and a walled garden where everything is lock-in. Since the typical Mac users today are using nothing but a web browser, I can see them at some point having a ARM-based laptop, and then perhaps let the "professional" lineup just slowly fade away.

 

[Darmok N Jalad]

Benchmarks like this are not comparing "apples" to "apples". Geekbench is known for using varying accelerated features, which makes it pointless as an indicator for "general performance".


AMD have not been successful in their ARM ventures. Don't forget that Zen was intended as a "stop-gap" before K12 was supposed to take over the PC market. K12 is now MiA and its ISA already obsolete, and AMD is instead focusing on improving Zen.


Just forget emulation.
A powerful x86 CPU struggles with emulating a tiny cellphone CPU.
Apple really don't care about compatibility, they are going in the direction of a more closed ecosystem.


The biggest problem is the fact that ARM can't compete with x86 performance wise. To become competitive, it would have to become more CISC-like. Mobile devices with ARM CPUs rely heavily on accelerated features(through ASICs) while the ARM cores are very weak. This is great for energy efficiency, but gives very poor general performance and the hardware quickly becomes obsolete. Applications also have to be written specifically for each ISA to get the benefits from accelerated features.

But the problem you describe is also a big problem. There are many very customized ARM designs. The OS' needs a custom firmware to even be able to run on them, and writing applications for various ARM designs is a maintenance nightmare (if they want accelerated features).


Recompiling for ARM itself wouldn't be hard in most cases, but performance would suck, and performance critical ones would have to be rewritten as you said. Even worse, as the ISA keeps changing with nearly every new Apple chip, just imagine the workload for those poor software developers to keep up with Apple's ever-changing eco-system™.

My impression is that Mac usage among programmers is dropping quickly, it's nowhere like it was 5-10 years ago. Today Dell, Lenovo and HP have much better options, and as more and more developers tend to go for Linux as Apple keep selling outdated machines and lowering build quality, I think Macs soon will no longer be relevant for any of the "professionals".

If Apple's current trend continues, with their products becoming more a "fashion statement" than something for productivity, lower build quality to encourage people to upgrade quickly, and a walled garden where everything is lock-in. Since the typical Mac users today are using nothing but a web browser, I can see them at some point having a ARM-based laptop, and then perhaps let the "professional" lineup just slowly fade away.

Well, they did just launch the new Mac Pro, which I would think would have a long run in its current form. I’m wondering if Apple will start putting both ARM and x86 chips inside their Pro line, or maybe even all of them. Could it be that they never really fully transition away from x86, but instead leverage whatever CPU is needed for the program? Imagine a base OS and programs running on ARM, and x86 being idle unless called upon. They already have some of their own architecture in the T2 managing functions of x86 machines, could that be the beginning of something bigger?

 

[efikkan]

Well, they did just launch the new Mac Pro, which I would think would have a long run in its current form.

Which is insanely overpriced and will not reach any substantial marketshare…
In recent years Apple have moved towards custom APIs and deprecating open source ones, making it harder to ship applications there etc. Over time they will loose more and more of the productive applications which were the reason why "professionals" used Macs in the first place.

I’m wondering if Apple will start putting both ARM and x86 chips inside their Pro line, or maybe even all of them. Could it be that they never really fully transition away from x86, but instead leverage whatever CPU is needed for the program? Imagine a base OS and programs running on ARM, and x86 being idle unless called upon. They already have some of their own architecture in the T2 managing functions of x86 machines, could that be the beginning of something bigger?

Well, they did have something like that in the early 90s; an optional card which basically contained a whole PC on it with DOS and Windows, and the possibility to switch between them. link

On a serious note; technically it could be possible to do different CPUs in a single computer, but be aware that the management for the OS would be insanely complex. I don't think it would be worth it; the reason for using ARM would be energy efficiency in laptops.

 

[cyneater]

Apple went from Power PC to X86..
It won't take to much to go from X86 to arm.