Probably baked into the cake is that Mac App Store apps will download the correct "thin" binary for your Mac's architecture thus saving space over a fat binary.
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
ARM based MacBook Pro discussion [merged]
- Thread starter Omega Mac
- Start date
-
- Tags
- arm intel macbook pro macos sierra
- Sort by reaction score
Quite right. Technology is much more advanced now: companies aren't shipping CDROMs or DVDs with software anymore
My job depends on me being able to virtualize other operating systems from time to time and also build docker images to run on our production Linux servers. Both of those tasks will be impossible on a hypothetical ARM Mac. If apple shift to arm, it’s the end of the road for me. I’ll miss macOS. ARM means the effective death of VMware Fusion, Parallels. as well as Docker Desktop for Mac.
Linux is a workable alternative, and I’m keeping a close eye on Microsoft’s efforts with their Linux compatibility layer. I’m not enthused about either, but it’s nice to have options I guess.
I understand, but there isn't going to be this straight up jump to ARM across every product overnight. If a full move to ARM happens it will only be when ARM chips are sufficiently fast enough to run VMs of other operating systems. During a transition there will be a mix of ARM and x86. There will also be Universal Binaries that can run an app on macOS x86, macOS ARM and iPadOS. The interface and app features will adapt depending on which OS it runs on.
The problem isn't "run other Operating Systems". ARM (can) CPUs already have hardware support for virtualisation, but that just gets you a virtualised ARM machine. It doesn't help in the slightest if the software you need to run is x86/x86-64 - you'd be back to software virtualisation of an Intel (or AMD) CPU. We have literally seen what that looks like, and it is not a viable solution.
[automerge]1584212903[/automerge]
That is some straight up wishful thinking.
Assuming the popular theory it's going to be a MacBook Air like machine first, I think we can safely say the A12X already comfortably outperforms Intel-Y, so anything it ships with should be able to at least match it under emulation. As soon as Apple make it clear that their own chips are the future of the platform, the majority of development efforts will shift to making Arm compatible versions, so the ecosystem should begin populating quite quickly. For anything that is no longer updated, well just ordinary MacOS updates can break compatibility over time. Either way, over the long term your only choice is to park your computer on a compatible OS and keep it there until it's completely clapped out, or move on to a modern alternative.
You’ve never used a software emulated virtual machine (ie emulating a different cpu architecture) have you?
that’s literally irrelevant to the topic of running x86 virtual machines
And while a less common use, this is also a very good use of current x86 virtualisation: I use a billing/invoicing application for my business, and it doesn’t support Catalina. It runs perfectly fine in a VM running a pre-Catalina version of macOS.
every time someone mentions the ability to run virtual machines there is always someone there - with clearly no understanding of the use case - to proudly proclaim that Arm can do anything, macOS releases would have broken stuff anyway and that all Mac developers will release updated apps for arm..
My point (as you sneakily omitted) was who's going to be doing that sort of heavy emulation on a MacBook Air? You can bet there's going to be an x86 based Pro available until the transition is so far behind us x86 is completely irrelevant to contemporary software.
Perhaps the reason people keep mentioning it is because people like you try to present a very niche issue as a massive show stopping problem, that completely undermines any benefit of Apple moving to in house processors which would provide a host of other benefits to other users.
You can bet on it can you?
Mac laptops are very popular with developers of web applications. Not being able to run virtualised software that closet mimics a live environment is not a “niche” issue.
I’ve still yet to see this “host of other benefits” that outweighs hardware compatibility with essentially the entire non-phone/tablet computing world.
Yes, it is a niche use case which will continue to be catered to by a lingering x86 machine. Perhaps a whole occasionally updated line of x86 machines if demand is as strong as you say to sustain it. Everyone else will move on to the new platform and be perfectly happy there.
Apple are unarguably the best company to design chips for Apple. Trying to argue that a jack of all trades Intel chip is better is just contrary to what Apple's whole modus operandi shows. Hardware designed hand in hand with software works better, end of story.
Anyway I'm bored of this circular argument now, it's been had throughout this thread and others already. Feel free to have the last word or ignore as you see fit, I'm out.
I said all Macs won't move to ARM? If you have to use VMs you aren't going to use a MacBook Air for example. The fastest MBP and Mac Pro won't move to ARM anytime soon. There is likely no decision made about the top end machines yet. Companies can try to look 3-4 years down the line but they can't predict what is going to happen that far ahead. There are many unpredictable variables in development, physical constraints, supply chains, customer demand, political stability/upheaval.
Besides that, we consider:
- VMs are being run more and more in the cloud.
- Microsoft heavily promoting Windows 10X on ARM. If it is successful many professional applications will move to it in time.
- Bootcamp on ARM based Macs will likely be supported and be for Windows 10X.
Xcode is already merging these resources and moving in that direction.
None of the MacBooks will move to ARM, that makes absolutely no sense. Apple's ARM based "laptop" will be based on IOS/iPadOS, optimized for trackpad/mouse. It will be nothing like a MacBook.
And? Why don't we all just switch to Chromebooks then. Who needs computers.
Windows 10X is a disaster by most all accounts. Exactly why Apple won't be porting classic macOS to ARM.
LOL
"Mac laptops are very popular with developers of web applications. Not being able to run virtualised software that closet mimics a live environment is not a “niche” issue."
Web developers are a very niche market. Maybe not here on MacRumors, but throughout the world, for sure.
Macs are a niche market throughout the world.
And I'm not sure what definition you're using for "web developers". I don't mean (just) a guy sitting in a cafe making some brochure site for boutique coffee beans or what have you.
I'm talking about anything that you access via a browser: hence I said developers of web applications.
VMs in the cloud are pretty useless for just about anything you need to test. It's like saying VMs are for servers, just because servers happen to also use VMs. VMs are used in a variety of places and pointing out one doesn't mean the other uses of VMs suddenly disappear. I'm using VMs a lot, and I'm not doing development. When you're in a mixed environment as a sysadmin, a VM helps you test changes in the network on another type of machine.
But, yeah, sysadmins are probably a niche market. Sum all the niche markets and you get a sizeable chunk of the market.
given that 10x is not out on any hardware and only in emulation that’s an interesting summary!
why couldn’t a MacBook Air run arm? Better battery, thinner, fanless design ...
.. and no pro grade software support.. no Adobe Creative Suite.. no desktop-grade MS Office.. no Virtualization.. no Docker.. just a bunch of App Store junk and Catalyst applets.. NO THANKS!
My Air is already thin, quiet, and has all day battery life.
Microsoft had no choice with Windows 10X, as they needed their own ARM platform. Apple has no such problem - they already have iOS and iPadOS, which is what their current and future ARM computers will run. MacOS isn't getting ported to ARM.
Last edited:
Time will tell, porting software isn't a huge barrier IMHO, and Apple get better (cheaper) supply of CPUs going their own way with better heat mgmnt. I tried a MacBook Air, but sent it back, it was slow and the fans were very noisy compared to my existing Surface 6. It will be interesting to see what happens over the next 12+ months with ARM and MacOS.
People used to think machines like the Toshiba T2000SX were light because they were only 7 pounds instead of 7 kgs. People used to think 3 hours battery life was good. I just switched from an iPhone 7 (nearly all day battery) to a Samsung S10+ (two days battery) and it's great.
As for the pro apps: it remains to be seen, but there were no pro apps for Intel macs on day 1. Photoshop was notoriously slow. If Apple is serious about moving to ARM, it can work with the software developers so that you get a selection of native apps on launch day. It's quite possible.
Of course, you're right about virtualisation/docker (assuming you're talking about running an existing VM/container). You can run Ubuntu Server which has docker and Kubernetes but of course, you need to find ARM binaries for both. I would actually expect an ARM Mac to boost ARM for Linux development, which is already pretty advanced (and not too many people know this), but could get even better.
This is unlikely; Apple have been moving consistently further away from Linux in recent years, and this is under x86. Getting Linux to work on any recent Apple product is tortuous, to say the least.
Proprietary hardware (eg SSDs, Wifi, T2 chip), the inexorable purge and ossification of GPL tools in MacOS, the closing down of non-App store avenues for installation; Apple have shown no intention of getting cosier with Linux. If anything they seem to be moving more towards a closed, Apple-only ecosystem in both software and hardware, and ARM - if it happens - is likely to be part of this strategy.
I don't see any Apple building bridges with Linux or anything outside the Apple ecosystem because of ARM.
Anything is "quite possible" in theory and in forum speculations / wish lists..But let's get real. Microsoft doesn't even have native Office port for their own Windows 10X. Are they really going to port macOS Office to ARM? Not anytime soon. Same goes for other major pro grade software makers - Adobe, AutoCAD et all. Best case scenario - you will be able to run iOS versions of those apps, which are all feature-crippled and not fully compatible with x86 apps.
Sure, you're going to have some Indies releasing a bunch of App Store grade note taking / task managers / weather type apps. That doesn't cut it, of you truly want to have a pro-grade Mac computer.
Much more pragrmartic and realistic approach for Apple would be to continue to evolve and enhance iPadOS and its app ecosystem, to act as an alternative to a classic MacBook/macOS. And in parallel - maintain Macs/macOS as an x86 Intel platform. Which is.. EXACTLY what they are doing now
I didn't mean it as in: running Linux on the ARM mac. I meant it in a different way. If you want to develop for ARM Linux at the moment, you need an ARM platform but they're not very mainstream. You could take a Raspberry Pi, but despite the fact that it's a "big seller", the total sold is in the order of magnitude of macs sold PER YEAR.
So if you get a platform that sells as well as the mac, but you can run a native ARM Linux VM on it, it would help ARM Linux development.
It's never a good idea to point at Microsoft as an example of a company taking an approach and failing. You might as well say that chat apps aren't viable because Google tried it 10 times already. Microsoft is never decisive. Apple is, or at least used to be. With the Intel switch, they made up their mind and then planned it as an unavoidable future. With Microsoft, the ARM was very tentative, to say the least.
I'm not saying Apple will switch mac to ARM. I don't know. Just saying that it's much easier than many people think, both technically as well as in the area of leveraging their power over the ecosystem. The Intel switch was do-or-die because the hardware sucked so much that hardly anybody was using it anymore. So they couldn't push the software vendors because it was basically a dying platform. At this moment, mac is very much alive and hip, so software vendors are much more likely to go along pre-launch.
We are just going to have to agree on disagree on both of your points.
First - it is anything but "technically easy" to migrate complex pro-grade ecosystems between x86 and ARM micro-architectures. Again, I am not talking about simple macOS utilities from App Store, which you can mostly just re-compile - I do not care about those. I am talking about complex pro-grade suites like Adobe Creative Suite, Microsoft Office, AutoCAD and others. These would take years of complex and costly development to port over. You simply do not understand the technical challenges and intricacies that are involved here, if you're going to claim "oh this is easy, just click an Xcode button".
Second - Apple does not have any real power over developers, that's another myth. There are hardly any profits in developing and selling macOS software these days. Apple has hard time even convincing major companies to develop macOS software. There wasn't even a native macOS Twitter client, until Apple offered Catalyst cross-platform hack as an option. Apple is not going to convince developers to maintain yet another platform "macOS for ARM". There is hardly any money in it, and no one is interested.
Apple knows this too, they aren't stupid.
Yeah, we can agree to disagree alright. But do note that I wrote "much easier than many people think", not "easy". I've got more than passing knowledge of software development and porting, having ported a few OSses to a few processors in my days. While I don't claim to have done an extensive inside knowledge of Adobe or Office, I'd say that after the software frameworks have been ported over, it will be much less than years. The step from Intel to ARM is much smaller technically than from PowerPC to Intel (endianness etc) and software development has also evolved quite a bit, so agreeing to disagree, I stand by my point