If Mac went ARM tomorrow, would you jump?

[Steve121178]

Most likely. I would expect the Mac Pro to run Intel for a long time simply because Apple doesn't make a chip quite like that. MacBook? Definitely going ARM. MacBook Air? Most certainly? MacBook Pro? some of the models maybe? That i9 might be tough to replicate.

I think there's no reason to not have all MacBooks running on ARM CPU's in the near future. Imagine a portable powerhouse notebook with every iOS app able to run natively - it would be incredible! Especially as developers optimise their iOS apps for the desktop/iPad.

The i9 should never have found it's way into a thin & light notebook anyway. It's like putting a 6 litre V12 on a golf cart.

 

[AustinIllini]

I think there's no reason to not have all MacBooks running on ARM CPU's in the near future. Imagine a portable powerhouse notebook with every iOS app able to run natively - it would be incredible! Especially as developers optimise their iOS apps for the desktop/iPad.

The i9 should never have found it's way into a thin & light notebook anyway. It's like putting a 6 litre V12 on a golf cart.

I don't know the specs and performance comparisons between future A chips and future Intel chips, but I'm not sure it's like putting a v12 in a golf cart. I'm not sure Apple has an i9 replacement right now.

 

[theluggage]

that means the DVD of CS4 would be absolute on ARM macs then....

Apple have announced that the next version of MacOS for Intel will drop support for 32-bit Apps. Don't know if CS4 is 32-bit but, given its age, it is quite likely. That Office 2011 DVD will also be toast. Certainly, there's going to be a mass cull of older, unsupported software even on Intel Macs, which will also take out a lot of the apps that feature assembler code or make assumptions about CPU architecture (which, unless you're writing operating systems or drivers, is sloppy programming in the 'new 10s' anyway).

That i9 might be tough to replicate.

How about a chip with 32 or 64 cores?

https://www.huawei.com/us/press-events/news/2019/1/huawei-unveils-highest-performance-arm-based-cpu
https://bit-tech.net/news/tech/cpus/ampere-launches-32-core-33ghz-arm-chip/1/

Or an ARM-based supercomputer:
https://www.extremetech.com/computi...d-supercomputer-launched-as-exascale-heats-up

The way ARM competes with Intel is cramming more cores, plus GPUs, vector processors, video accelerator etc. hardware into the same thermal envelope. For supercomputing applications, its more of a controller chip for dedicated on-chip number-crunching gizmos.

Meanwhile, even if you take the claims that the iPad Pro with A12x is as fast as an i7 MacBook Pro with a pinch of salt, the A12x should already outperform the 12" MB and MacBook Air.

As with the Intel transition, the first thing Apple is likely to do is announce their intentions and make a prototype machine available to developers.

Actually, the "heck, no" contributions to this discussion is awfully like the reactions when the first rumors of the move to Intel surfaced... and I think that was partly down to people thinking that Apple was going to put a Pentium 4 space-heater in the Mac. In fact, Intel were in the process of a major about-face that included junking the P4 "Netburst" architecture and starting again with the older Pentium Pro to produce the game-changing "Core" range.

There is a lot of interest in server- and workstation- class ARM processors at the moment. Of course, ARM started as a workstation chip that could run rings around the (then) Intel 286, but that was at the height of the DOS/Windows stranglehold which has now been greatly weakened by the rise of mobiles, the web and Linux.

 

[AustinIllini]

Apple have announced that the next version of MacOS for Intel will drop support for 32-bit Apps. Don't know if CS4 is 32-bit but, given its age, it is quite likely. That Office 2011 DVD will also be toast. Certainly, there's going to be a mass cull of older, unsupported software even on Intel Macs, which will also take out a lot of the apps that feature assembler code or make assumptions about CPU architecture (which, unless you're writing operating systems or drivers, is sloppy programming in the 'new 10s' anyway).



How about a chip with 32 or 64 cores?

https://www.huawei.com/us/press-events/news/2019/1/huawei-unveils-highest-performance-arm-based-cpu
https://bit-tech.net/news/tech/cpus/ampere-launches-32-core-33ghz-arm-chip/1/

Or an ARM-based supercomputer:
https://www.extremetech.com/computi...d-supercomputer-launched-as-exascale-heats-up

The way ARM competes with Intel is cramming more cores, plus GPUs, vector processors, video accelerator etc. hardware into the same thermal envelope. For supercomputing applications, its more of a controller chip for dedicated on-chip number-crunching gizmos.

Meanwhile, even if you take the claims that the iPad Pro with A12x is as fast as an i7 MacBook Pro with a pinch of salt, the A12x should already outperform the 12" MB and MacBook Air.

As with the Intel transition, the first thing Apple is likely to do is announce their intentions and make a prototype machine available to developers.

Actually, the "heck, no" contributions to this discussion is awfully like the reactions when the first rumors of the move to Intel surfaced... and I think that was partly down to people thinking that Apple was going to put a Pentium 4 space-heater in the Mac. In fact, Intel were in the process of a major about-face that included junking the P4 "Netburst" architecture and starting again with the older Pentium Pro to produce the game-changing "Core" range.

There is a lot of interest in server- and workstation- class ARM processors at the moment. Of course, ARM started as a workstation chip that could run rings around the (then) Intel 286, but that was at the height of the DOS/Windows stranglehold which has now been greatly weakened by the rise of mobiles, the web and Linux.

Honestly, I was just speculating. I don't have any expertise on this front but I can imagine it's not super easy to convert 1 to 1.

"Heck no" is fine if you're worried about compatibility issues earlier on in the product cycle. I just bought a Mac, but in 2023-2024 I'll likely try to buy another one and I will certainly want the A chips.

 

[Steve121178]

I don't know the specs and performance comparisons between future A chips and future Intel chips, but I'm not sure it's like putting a v12 in a golf cart. I'm not sure Apple has an i9 replacement right now.

I thought it was well documented that the i7 performed better as the i9 throttled due to completely inadequate thermals/cooling.

It's all relative though. The iOS apps should run far more efficiently on beefed up ARM CPU's than they would if they were optimised for x86 architecture (their Intel equivalents). The CPU/GPU in the iPad Pro 2018 was easily able to keep up with a recent generation Intel Core i5 chip if you believe the benchmarks. I'm no expert, just making some observations.

At the end of the day Apple can't afford to get it wrong and I don't think they will. Their work so far on ARM chips has been nothing short of remarkable, so Apple clearly have the right people on the job.

 

[RecentlyConverted]

For me to switch to an Arm Mac

- It must run Microsoft Office (a version at least as good as the one we have now).
- A substantial number of developers transition their apps to run in it.
- Open, i.e. I can buy apps anywhere not just the App Store. If it’s closed like iOS forget it!
- It doesn’t lead to a dumming down of Mac (just running enhanced iPhone apps.
- A better user experience- speed, functionality etc.

 

[cardfan]

For me to switch to an Arm Mac

- It must run Microsoft Office (a version at least as good as the one we have now).
- A substantial number of developers transition their apps to run in it.
- Open, i.e. I can buy apps anywhere not just the App Store. If it’s closed like iOS forget it!
- It doesn’t lead to a dumming down of Mac (just running enhanced iPhone apps.
- A better user experience- speed, functionality etc.

You’ll be disappointed then. IMO they need to license macOS.

 

[RecentlyConverted]

For me, my MacOS exit plan - buy last intel iMac, use it as now until end of support. Then detach it from internet, do all my work on it. AirDrop documents to an iOS device running latest iOS and use that to email documents etc produced on the iMac and any other web interactions. Or just go back to windows which is getting better all the time.

 

[cram501]

Meanwhile, even if you take the claims that the iPad Pro with A12x is as fast as an i7 MacBook Pro with a pinch of salt, the A12x should already outperform the 12" MB and MacBook Air.

Until we get a laptop or desktop with an Apple ARM chip for comparison, there is no way to know if this is true. The comparison is not Apples to Apples. As much as Apple wants it to be true, an iPad is not the equivalent of a desktop. It is useful and I like my iPad Pro but I do not believe it is close to being powerful enough for my work.

Things will get interesting over time as more investment goes into building the ARM chips. My problem with the ARM transition, besides speculative performance, are Apple priorities. As they've shown with the MBP over the last 4-5 years, their priorities do not align with mine. When they produce a laptop that is an iPad with keyboard that won't work for me or they combine MacOs and iOS into an OS that won't work for me, I'll have no choice but to look elsewhere.

They may produce a hell of a set of Macs. I could be surprised and happy. In the last few years they have shown that their priority is not only form over function but form to the detriment of function in their search for thin-ness. That does not give me hope but we'll see.

 

[theluggage]

Honestly, I was just speculating. I don't have any expertise on this front but I can imagine it's not super easy to convert 1 to 1.

It depends entirely on the application. If an application is written entirely in C/C++/ObjC/Swift (or, these days, Javascript or Python) and uses the in-language facilities for checking things like the sizes of data types, where relevant, and uses OS frameworks (Metal, Core Audio, Grand Central, Accelerate etc.) rather than rolling its own graphics/hard maths code then ideally it should be no more work than rolling out a point release. That won't cover everything - but modern software is far more likely to follow those rules than it was in 2005 when Apple switched to Intel. Following App Store rules will have helped that, the 64-bit transition will have helped further, plus lots of developers have already produced iOS versions/variations of their apps so they'll no the ropes and may have made sure that the 'engine' of their software is the same between platforms. Developers from a Unix/Linux background (rather than old DOS hackers) will be used to a tradition of writing portable code - most of the well-know Open Source packages already support multiple processor types, including ARM.

Apple have done this twice before, remember.

- It must run Microsoft Office (a version at least as good as the one we have now).

Unlikely on day 1, other than via emulation. Longer-term, impossible to call without insider info on how the current Office for Mac has been written, how much of it has already been re-written to get rid of any ancient DOS-era code and how much is shared with the Windows version (which Microsoft probably are porting to ARM). I believe the current version of Office for ARM is a 'lite' version, which doesn't meet your criteria.

- A substantial number of developers transition their apps to run in it.

A substantial (but by no means all) number of developers will just need to tick the 'ARM' box and re-build.

- Open, i.e. I can buy apps anywhere not just the App Store. If it’s closed like iOS forget it!

There's no technical feature of ARM that makes that more likely - if Apple think that a switch to ARM would be a good time to sneak this through they should google 'Windows RT'. If Apple want to do this, they don't need ARM.

- It doesn’t lead to a dumming down of Mac (just running enhanced iPhone apps.

True - but the flipside is that there are more iPhone users than Mac users, so it could increase the range of Mac apps: if you're planning an iPhone app but can expand your potential customers by 10% with only modest effort by adding Mac compatibility, why not?

- A better user experience- speed, functionality etc.

That's up to Apple. But it already looks like an A12x could thrash a 12" MB or an Air.[/QUOTE]

 

[MBAir2010]

For me, my MacOS exit plan - buy last intel iMac, use it as now until end of support. Then detach it from internet, do all my work on it. AirDrop documents to an iOS device running latest iOS and use that to email documents etc produced on the iMac and any other web interactions. Or just go back to windows which is getting better all the time.

I use linux besides osx and linux seems a better option because the pureness and quick response time. linux apps are fun to use nowaday because the users are expanding, windows is still too commercial for my computing soul

 

[joema2]

Until we get a laptop or desktop with an Apple ARM chip for comparison, there is no way to know if this is true....As much as Apple wants it to be true, an iPad is not the equivalent of a desktop. It is useful and I like my iPad Pro but I do not believe it is close to being powerful enough for my work...

The latest-generation iPad Pro uses the A12X CPU which is faster than many laptops. Its GeekBench single and multi-core scores aren't much slower than my 2017 i7 iMac 27.

And -- that's on a mobile-optimized part which wasn't intended to compete head-to-head with a laptop or desktop.

I also agree my iPad Pro is not powerful enough for my work, but it's not due to lack of CPU power -- it's the current limitations in the iOS, app and mobile web ecosystem. E.g, cutting/pasting entry fields on some web forms doesn't work right, not all web sites respond to "request desktop", so you get stuck with the dumbed-down mobile version. There are no full-fledged high-end desktop apps like Photoshop or Premiere Pro for iOS but this will gradually change.

Powerful as the A12X is, that's not the version that would be used in the hypothetical ARM MacBook. It would be a future 2020 version, probably the successor to the upcoming A13, and likely fabricated using 5nm technology:

https://www.tomshardware.com/news/apple-iphone-rumors-2020-tsmc-5nm-cpu,38692.html

Even the current A12X iPad Pro is faster on various productivity apps than a Dell i7 XPS 13: https://www.laptopmag.com/articles/new-ipad-pro-benchmarks

Until recently the prevailing view in the CPU architectural community was neither ARM (nor any RISC design) can beat Intel. Over the decades this viewpoint has evolved. Up until the late 1990s the general viewpoint was RISC had an advantage. Then Intel's Pentium Pro used an internal RISC design while externally maintaining the original CISC instruction set. Other Intel CPUs followed this design.

Since then Intel's massive fabrication and design resources seemed to indicate it's the CPU design *execution* not the design itself that's important. The few surviving RISC designs such as IBM POWER and Oracle SPARC only seemed to support this -- they really weren't dramatically better, faster or more power efficient.

ARM was better on the low end but the traditional wisdom was if ARM was scaled up to the performance level of an 8-core or higher Intel desktop CPU, the ARM chip would burn just as much power.

Recent results indicates this may not be the case. It's starting to appear that ARM (at least in Apple's hands) can compete and possibly beat Intel -- even on the desktop. This has yet to be realized but current trends are pointing in that direction.

What is different now that enables this has yet to be clearly explained.

 

[cram501]

The latest-generation iPad Pro uses the A12X CPU which is faster than many laptops. Its GeekBench single and multi-core scores aren't much slower than my 2017 i7 iMac 27.

And -- that's on a mobile-optimized part which wasn't intended to compete head-to-head with a laptop or desktop.

I also agree my iPad Pro is not powerful enough for my work, but it's not due to lack of CPU power -- it's the current limitations in the iOS, app and mobile web ecosystem. E.g, cutting/pasting entry fields on some web forms doesn't work right, not all web sites respond to "request desktop", so you get stuck with the dumbed-down mobile version. There are no full-fledged high-end desktop apps like Photoshop or Premiere Pro for iOS but this will gradually change.

Powerful as the A12X is, that's not the version that would be used in the hypothetical ARM MacBook. It would be a future 2020 version, probably the successor to the upcoming A13, and likely fabricated using 5nm technology:

https://www.tomshardware.com/news/apple-iphone-rumors-2020-tsmc-5nm-cpu,38692.html

Even the current A12X iPad Pro is faster on various productivity apps than a Dell i7 XPS 13: https://www.laptopmag.com/articles/new-ipad-pro-benchmarks

Until recently the prevailing view in the CPU architectural community was neither ARM (nor any RISC design) can beat Intel. Over the decades this viewpoint has evolved. Up until the late 1990s the general viewpoint was RISC had an advantage. Then Intel's Pentium Pro used an internal RISC design while externally maintaining the original CISC instruction set. Other Intel CPUs followed this design.

Since then Intel's massive fabrication and design resources seemed to indicate it's the CPU design *execution* not the design itself that's important. The few surviving RISC designs such as IBM POWER and Oracle SPARC only seemed to support this -- they really weren't dramatically better, faster or more power efficient.

ARM was better on the low end but the traditional wisdom was if ARM was scaled up to the performance level of an 8-core or higher Intel desktop CPU, the ARM chip would burn just as much power.

Recent results indicates this may not be the case. It's starting to appear that ARM (at least in Apple's hands) can compete and possibly beat Intel -- even on the desktop. This has yet to be realized but current trends are pointing in that direction.

What is different now that enables this has yet to be clearly explained.

My point isn't that Apple can't produce an Arm chip that competes with intel in a laptop or desktop, it is that it will take longer than everyone anticipates and they haven't produced one yet. When/if they get a commercially viable laptop or desktop, we'll see how it stacks up.

You can't compare an iPad to a laptop until it has the same limitations with the same capabilities on the same preemptive, multitasking operating system. Comparing artificial benchmarks on iOS to MacOs or Windows isn't valid. There are way too many variables. People seem to think that is the end of the discussion and Apple is going to turn out an Intel killer.

What made Intel and X86 the winner was that I could go down and build my PC for a few hundred dollars and have a decent OS running (well, in the mid-80's it cost ~$2000 but was still comparatively inexpensive). The Intel architecture was always weird.

Because X86 had a stranglehold, i"m not sure it was believed other architectures couldn't beat it. It wasn't worth it. If you can't grab a big enough marketshare from x86, it isn't worth the investment. Companies tried and failed. I'm not sure it had anything to do with capability.

Arm is just another architecture that hopefully I'll be able to do my work on assuming Apple creates a machine I think is worth buying. It is better for Apple. Except for possibly better battery life, I don't see any benefits for the consumer. Apple isn't one to sell anything cheaply.

 

[jerwin]

Every couple of months, Apple starts nagging me about upgrading my installed apps to 64 bit. Many I don't care about. The ones I do care about (on at least a nostalgic level) are games. Good games that I occasionally play. Many of these games are old enough to be playable at 5k-- which is part of the reason they still hold appeal.

 

[Ledgem]

Imagine a portable powerhouse notebook with every iOS app able to run natively - it would be incredible! Especially as developers optimise their iOS apps for the desktop/iPad.

That's an interesting way of looking at it. I'd guess most of us view it the other way around: that it's amazing when a desktop-class application finds its way onto iOS, and that the Surface's ability to run desktop-class programs is what's amazing.

For me, iOS apps tend to be either something that already has a Mac program (one exception being Notability, which made the Mac program later), or a front-end for a website that I'd normally access through a browser on the Mac. Aside from Apple's own News, I can't think of many iOS apps I'd be running on my Mac, even if I could... in fact, I can't think of any. I admit that I'm not thinking creatively. I liked the idea of Marzipan, and the promise of being able to run iOS apps on my Macs. I don't like it enough to give up my current macOS applications, though.

Actually, the "heck, no" contributions to this discussion is awfully like the reactions when the first rumors of the move to Intel surfaced... and I think that was partly down to people thinking that Apple was going to put a Pentium 4 space-heater in the Mac.

People don't like change. I used PowerPC-based Macs at work at the time, but my personal devices were Windows-based. How was the transition to Intel? Did PPC programs run slower, or buggy under Rosetta? Those are the types of fears people have.

At least with the move to Intel there were clear benefits. People argued over which processors were superior, but it became clear that the G5, despite being powerful at the time, wasn't particularly optimal... and then Intel kept forging ahead. The other benefit was going with a mainstream architecture, instead of a minority one. I'm not saying that ARM can't blow by Intel, but I'm also not counting Intel out just yet. And while ARM isn't a minority architecture in general, it is a minority architecture on the desktop scene, which introduces risks if Apple were to move their desktop line there.

 

[russofris]

F
- Open, i.e. I can buy apps anywhere not just the App Store. If it’s closed like iOS forget it!

You see, I desire just the opposite. I want a completely locked down, air-tight, encrypted, closed system that nobody can install software on unless it has been rigorously vet'ed by a trusted 3rd party that maintains a complete chain of accountability. I own a Mac so that I (alone) have 100% agency over the files that reside on it.

I realize that this is not the reality at the moment, but it is my opinion that the extremes (completely closed/open) must both exist for any of us to get useful work done.

If I want an open-box, I'll rdesktop to the household PC or spin up a VM in Fusion. Why would I compromise the security of my Mac when I can run any extraneous software I want on your PC (joking to make a point).

 

[cram501]

The system is only as secure as you make it. Hiding behind a walled garden that may reduce security issues just gives you a false sense of security. Security is your problem and your responsibility on your devices, not the vendors. If you do and/or install stupid crap, you're just as insecure.

PC is too important for me to mess with. I'll avoid walled gardens.

 

[RecentlyConverted]

You see, I desire just the opposite. I want a completely locked down, air-tight, encrypted, closed system that nobody can install software on unless it has been rigorously vet'ed by a trusted 3rd party that maintains a complete chain of accountability. I own a Mac so that I (alone) have 100% agency over the files that reside on it.

I realize that this is not the reality at the moment, but it is my opinion that the extremes (completely closed/open) must both exist for any of us to get useful work done.

If I want an open-box, I'll rdesktop to the household PC or spin up a VM in Fusion. Why would I compromise the security of my Mac when I can run any extraneous software I want on your PC (joking to make a point).

So who is the infallible 3rd party, that will ensure all the software protects you and your data, and who only thinks whats best for you and does not limit you usage if what you should wish to do go against Apples income model. Monopolies are rarely good. We got sold on the iOS app store for its privacy, yet more and more stories are coming out about apps that infringe our privacy.

 

[joema2]

...For me, iOS apps tend to be either something that already has a Mac program (one exception being Notability, which made the Mac program later), or a front-end for a website that I'd normally access through a browser on the Mac. Aside from Apple's own News, I can't think of many iOS apps I'd be running on my Mac, even if I could... in fact, I can't think of any....

I agree to a point -- but it's not about today. The full Marzipan rollout will take two years. As software development moves increasingly toward a mobile-centric focus, Apple is trying to avoid future marginalization of Macs as an app development target. There is no Mac NetFlix app and Twitter is dropping their Mac app. Multiply that by 20x over the next few years.

Note your point about just using web apps -- not necessarily needing a native Mac or ported iOS app. This shows how an ARM transition would be less traumatic than past Mac CPU changes.

Another advantage to a future convergence of Mac/iOS apps is you'd often only need one device. Many times I take both MacBook Air and iPad Pro because I need apps or functionality available only on one or the other.

If some future ARM-powered 12" MacBook had two-day battery life, weighed 1.5 lbs and could run most iOS apps plus all Mac apps, I might take only that sometimes.

Or if a future 12.9" iPad Pro could run Photoshop, Premiere Pro, FCPX, plus all iOS apps, maybe I'd take only that.

Any loss in sales from people buying only one device would probably be compensated by increased market share and avoiding marginalization of the Mac app platform.

 

[cram501]

Or if a future 12.9" iPad Pro could run Photoshop, Premiere Pro, FCPX, plus all iOS apps, maybe I'd take only that.

Any loss in sales from people buying only one device would probably be compensated by increased market share and avoiding marginalization of the Mac app platform.

If Apple reaches that point it will be a complete collapse of the MacOs ecosystem. I have an iPad for that. I'm not sure why you want less utility and flexibility but I will be using something else.

I just can't see how they would make more money reducing their target audience unless they became a lot cheaper. Apple has never been about inexpensive.

 

[s15119]

Not in the market tomorrow. But yeah. If by the next time I'm ready for a desk top or mobile computer and Apple has switched to ARM, I'd be happy to make the change. I think computers will only get better with Apple designing the software, hard ware AND the chip.

 

[skottichan]

Yeah, I'd be pretty interested in an ARM-based iMac if it had equal to or better performance than my current system (Mid-2017 27" i7)

 

[maflynn]

Yeah, I'd be pretty interested in an ARM-based iMac if it had equal to or better performance than my current system (Mid-2017 27" i7)

That's the 64,000 dollar question as they say. Depending on who you ask, Apple's Ax processor family will either out perform Intel core i7 CPUs or it will be slower.

The Ax has been great at running one app at a time, but with these multi-threaded apps all running at the same time, who knows how that will perform.

 

[skaertus]

It is an interesting move, and people should have seen it coming for some time now. An ARM Mac makes sense if it provides better performance and better battery life than an Intel-based CPU. And of course higher margins for Apple (nobody expects Apple to reduce the prices, right?).

Intel has been dropping the ball. Four years using the same 14nm process and prices are still high. Maybe this year Intel finally delivers its mainstream processors with a 10nm process, as the rest of the industry is already moving towards 7nm.

As for me, I am not sure. I like the possibility of running Windows on my Mac. In all Macs I ever had, I installed Windows on BootCamp. It is a (literally) window of freedom, as I hate the idea of being locked in Apple’s “ecosystem”. I am OK with being restricted on a smartphone if it provides a better experience, but not my desktop.

Windows is also running in a few ARM processors, but not all of them (and I doubt Apple will make any effort to make Windows run on its ARM processors). If an ARM processor on a Mac prevents me from running Windows in all its glory, then I am out. Definitely out.

 

[russofris]

So who is the infallible 3rd party,

There isn't one, mainly because infallibility is impossible. Thankfully, us technologists have already solved the problem.

Approved/Reviewed apps are signed by Apple.
Signed apps are procured via the app store.
An issue is found with an application post-release (security/privacy, etc)
The signature on the app is revoked / added to a CRL.

I need a hardened implementation of this exactly as much as I need a completely open and unencumbered ecosystem like Libreboot+GNU.