Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Sorry, you are just wrong. Virtualization is not emulating the software, the hypervisor provides partitioned access to the underlying hardware, and depending on the system, a abstraction layer for some hardware (which is just what Metal is, an abstraction layer for GPUs and ML).

I understand how virtualization works vs. emulation in a technical sense but it makes no difference in the problem I have with the phrasing being used -- they are functionally the same for the end user, just a different technical process -- in any case people should then say "I'm running X OS as a virtual machine on my M1 mac." Does the same job as what I requested. What I'm trying to get across is that for me so long as macOS is still running and the guest OS is running in a window, it's misleading and gives me false hope lol. I want to be able to boot into an OS and not have macOS running at all.

The only way I'd consider buying an Apple Silicon Mac is if I can create two partitions: a small one for macOS and a much bigger one for Linux wherein I can boot into both individually (i.e, I can boot into Linux and macOS is not running at all a la Chameleon or Clover EFI).
 
they are functionally the same for the end user, just a different technical process
That is just not true. A user running a virtualized version of Windows, macOS or Linux is running that software, not emulating it. The underlying CPU is not being emulated. Even the devices are rarely emulated (sometimes they accessed via an abstraction layer, but in that case translated would still be more accurate than emulated).
-- in any case people should then say "I'm running X OS as a virtual machine on my M1 mac." Does the same job as what I requested.
And has the benefit of being true, unlike what you requested.
What I'm trying to get across is that for me so long as macOS is still running and the guest OS is running in a window, it's misleading and gives me false hope lol.
I understand that you do not want to run a visualized OS, but using the correct technical description is not misleading, and using the completely incorrect term you requested people to use (emulation) is quite misleading.

I want to be able to boot into an OS and not have macOS running at all.

The only way I'd consider buying an Apple Silicon Mac is if I can create two partitions: a small one for macOS and a much bigger one for Linux wherein I can boot into both individually (i.e, I can boot into Linux and macOS is not running at all a la Chameleon or Clover EFI).
Understand your desire, however, based on what Apple has said to date, you are unlikely to get your wish. Guess that means that you will have bought (or will be buying) your last Mac fairly soon.
 
Sounds like a good step forward that -could- see this not be another dead-end OS like Windows RT and Windows 10S. If only they would try more of an iOS approach to Windows on tablets ... or, y'know, if they had upsized what was quite a good phone OS, instead of killing it.

Hopefully, how incredible the Silicon Mac leap was, is a bit of a wakeup call across the board, but it probably won't be as much as it should be.
 
That is just not true. A user running a virtualized version of Windows, macOS or Linux is running that software, not emulating it. The underlying CPU is not being emulated. Even the devices are rarely emulated (sometimes they accessed via an abstraction layer, but in that case translated would still be more accurate than emulated).

And has the benefit of being true, unlike what you requested.

I understand that you do not want to run a visualized OS, but using the correct technical description is not misleading, and using the completely incorrect term you requested people to use (emulation) is quite misleading.


Understand your desire, however, based on what Apple has said to date, you are unlikely to get your wish. Guess that means that you will have bought (or will be buying) your last Mac fairly soon.
It is true, although perhaps you've misinterpreted what I mean by "functionally they are the same." I mean both virtualized machines and emulated machines are running in macOS, the user experience is fundamentally the same in my experience.

Whether I'm using the right technical term for emulation vs. virtualization in my original post, which was intended as a template example of providing context, is irrelevant as my use of "emulating" was the common catchall to refer to both emulation and virtualization tech alike. I'm not advocating for people saying they're "emulating" if they're not actually emulating a different a OS. I'm an advocate only for people providing context and accurately explaining what they're doing. "I'm running Windows ARM on my M1 Mac" may be a technically correct sentence but it's at the very least misleading or meaningless for the vast majority of people on here that don't know if it refers to emulation, virtualization or native booting. So no need to patronize, we likely both want the same thing here.

Real example: someone in this thread was talking about how they're testing out Windows ARM on their M1 Mac but provided no information on how they're testing it. Are they using QEMU? booting into it native? something else?

Yes my current Mac will likely be my last until the community starts to piece together how to support Apple Silicon on ARM builds of Linux. I was thinking about buying a ThinkPad or some laptop I can run Linux on as my main but the Apple Silicon power to performance ratio looks great so I'm holding off for as long as possible. Here's hoping. :D
 
Microsoft should package up a "Microsoft Windows for Mac" software package. It would include Windows 10 and a high quality hypervisor that would let you run Windows 10 ARM alongside MacOS at full performance on the new M1 Macs. it would of course include this emulation technology so x86 apps can be run as well.

This would actually be quite a money making opportunity for MS, to sell Windows to Mac users who absolutely need it for whatever reason. Done right it would have near-native performance and be far superior to previous solutions like VMWare's.
Why would they need to do that? Parallels will release virtualization on Apple Silicon. Microsoft will release Windows for ARM that can be installed as a VM.
 
Key word there is emulation. Rosetta 2 is not an emulator anymore than WINE is...they are both translators. Intel can't do jack regarding a translator because it don't actually use the x86 instruction set.
I don't know why you think emulation of x86 or translation of x86 is any different from Intel's perspective. Regardless, Transmetta's solution was based on translation and Intel went after them. There are lots of interesting papers about their technology.

 
Last edited:
Key word there is emulation. Rosetta 2 is not an emulator anymore than WINE is...they are both translators. Intel can't do jack regarding a translator because it don't actually use the x86 instruction set.
Are you saying that translators don’t really use the language they translate from? Talking about humans here.
 
I don't know why you think emulation of x86 or translation of x86 is any different from Intel's perspective. Regardless, Transmetta's solution was based on translation and Intel went after them. There are lots of interesting papers about their technology.
I think this because I read how a company got around this by black boxing the code ie they knew command x did y on the regular hardware which equated to z on what they had. So they simply had command x make call z rather then call y.
 
If they do not make these available today, making them available to Microsoft would mean making them public (once they are outside Apple’s control, they have to assume they will be public). In addition, making them available to Microsoft, would make it more difficult for them to restrict access to them in other contexts (read iOS/iPadOS/watchOS). If they keep them completely internal, their anti-trust argument is stronger than if they provide selective access to one other player.
The interfaces would be public, but not the implementation. I don't see any sort of secrecy concerns here. Would there be more people in the circle of trust? Yeah. Would Microsoft and others start to know more detail about what's in the SoC? Yeah. Would they know enough to replicate it in any way? No.

I do agree that Apple would be risking accusations of selective access. Probably not anti-trust, but certainly accusations and bad PR suggesting they're playing favorites or keeping some functionality for themselves. I believe they have the right to do so, but sharing some but not all makes a black and white issue grey and there are a lot of people who can't handle grey.

I am not sure in what world you live, but that is not how it would play out. If an Apple Silicon Mac had 24 hours of battery life under macOS and 12 under Windows, most tech reviewers would describe the machine as having 12 hours of battery life. Same for every other aspect. Your argument is the same one that is used by people to justify Apple allowing side loading on iOS/iPadOS/tvOS/watchOS: “If someone got a virus from a side loaded product, that would be their fault not Apple’s”. However, history shows us that is not true. When developers used a pirated version of Xcode to develop iOS apps that introduced problems the reporting was that Xcode had a virus.

Also, people who bought Apple Silicon Systems to run Windows, would describe their experience to others without differentiating: “I have a mac and it crashes all the time.” ”I have a Mac and I only get 12 hours of battery life”.
Yeah, still not worried about this. If nothing else, I think Windows still has a worse reputation than MacOS once you get outside of MacRumors troll country. People would run the comparisons and over time and the pattern would become clear. There's an amusing amount paranoia around Apple, so there may be some people who think the reason Windows is slow is because Apple is withholding access to key functionality, but I think the difference will be seen between the OSs and libraries, not the underlying hardware.
In 2006, their market cap was about $200 Billion. They had 1 major product line and were starting to see serious revenue from the iPod. Today their market cap is around $2 Trillion and they have the iPhone, iPad (a business that is by itself larger than Apple was then), Apple Watch, etc. and over $200 billion in the bank. Seems quite a bit stronger.
Sure, but I don't think the existance of Apple Watch and iPad impact what OSs Apple allows to run on Mac. They were a minor player in the PC market in 2006, and they're a minor player now. One difference is that they have an opportunity to become an at least somewhat larger player now because they're showing a performance edge.

I don't think "we have iPhones now, we don't need Windows on Mac" is a coherent enough argument from Apple's perspective. They've indicated they're fine with other OSes running on their hardware, I don't see a reason to think they wouldn't be.

First, it is not just about having more people, it means splitting the attention of those that are currently working on these products/projects. Second, if it was that easy to just hire more people, Apple would still rather hire them for projects that will net them more revenue and control.
Easy to say, but that is not how these things work in the real world. The liaison team will take resources and they have their own interests. They will lobby for changes that will benefit them and they customer. If they did what you describe, they would essentially be Microsoft in the 90’s. Using internal APIs and other things that they did not make available to others. When they make these things available to no one externally, they are in a stronger position to deny others access in the iOS/iPadOS/tvOS/watchOS realm.
Having more people means not splitting people's attention.

There is already a team focused on making Macs more compatible with Windows. Bootcamp provides Windows drivers for Mac hardware, and they managed to do that alongside MacOS development and the M1 and MacOS for M1 developments. I think they can walk and chew gum on this.
It also creates a problem for Microsoft in that it would create a fork (Windows and Windows on Apple Silicon). Anyone who wanted to take advantage of the Apple Silicon custom hardware would then lock those applications onto Apple machines, increasing sales of Apple’s machines and making native macOS ports more likely. None of these things is good for Microsoft
They've already forked. Post 1 on this thread is about Windows on Arm.

Part of what already makes Windows such a hot mess is that they don't lock anything to particular hardware. They're not locked to Nvidia graphics, or Intel, or AMD/ATI. They're not locked to specific audio or camera hardware. They don't restrict what drivers you can install to access custom hardware and accelerators on the PCIe busses.

I think you might be making a big thing out of what is just a normal Tuesday for OS developers.

Why not? Everyone said they would not make their own computers and yet now they are. The only way to control one’s destiny is to do it. To quote the over quoted: “People who are really serious about software should make their own hardware.”
I still think it's dumb for them to be making their own computers and competing with their customers. Just like I think the Pixel is a dumb idea. Fortunately they're bad enough at it that their customers aren't feeling particularly threatened.

I should be more precise in my language. I think it's dumb for Microsoft's customers to buy their OS from a company that is willing to compete with them directly, just like I think it's dumb for Android customers to do so. Therefore, I think it's dumb for Microsoft and Google to put their customers in a position where they find themselves making dumb decisions.

That said, the reason I don't think they'll make their own chips isn't for business reasons, but because it's really, really, hard. Designing a motherboard is a relative cakewalk. Apple spent years bringing the team they have together and developing the experience they have and they funded that exercise by selling hundreds of millions of iOS devices. Best I can tell, Apple ships more processors a year than Intel does. There must be a billion devices out there right now with Apple Silicon in them, and Apple ramped that up by competing with mediocre entries in the smartphone space until they had matured their designs to the point they could cross over to PCs. Where would Microsoft practice their design chops until they were competitive?

So yeah, if it became existential to Microsoft, they could design a CPU, but as far as I can tell they'd be starting from standing start and it would take years before they were competitive with existing offerings. There's a few companies they might buy to bootstrap the effort, but they'd still be starting from behind.
 
Microsoft should package up a "Microsoft Windows for Mac" software package...

This would actually be quite a money making opportunity for MS, to sell Windows to Mac users who absolutely need it for whatever reason...

MS makes almost no money on the Macs to begin with...and has not for several decades. First, Apple only has, at best, year after year for 30+ years, 7% of the Personal Computer space. That's such a tiny segment of the Personal Computer market. Think about it: if one piece of the market supplies 93%+ of your potential customers and the other segment of your market supplies 7%, who are you going to write your apps for?

Second, out of the few million Apple Macs that are sold each year (10-20 million source: https://www.statista.com/statistics/276308/global-apple-mac-sales-since-fiscal-year-2002/ ) , only a minority purchase MS Office. A miniscule share want to run Windows on a Mac. So to say that this "would actually be quite a money making opportunity for MS" is nonsense and most likely not worth Microsoft's time other than to promote that MS now supports ARM in some capacity to make some folks happy. "But what about all the iPads and iOS devices that are ARM?!" folks will say. My reply is that practically nobody on a touch screen iOS device is using MS Office. So again, this is not going to make much money for Microsoft.

Microsoft is going to concentrate on where the sales and profits live: non-Apple world. Dell shipped 260 MILLION machines in 2019...Lenovo shipped 60+ million. Then factor in Toshiba and all the other Wintel players. Again, I'm sure MS will happily accept some revenue from Mac Arm folks, but it's tiny peanuts compared to the Wintel vendors out there.
 
1. I don't have to use DOS, while apple people talk about "terminal" :p
2. To search is easy but to search in MACOS, er aaa let open "terminal" locate.

Why do I use macOS iMac?

1. I hate laptop keyboards either it was windows or mac. I got broken keyboard magic 2 and a lot of broken keyboard laptop after one year
2. I build iOS apps. If not I prefer to stay in ubuntu/fedora/windows.
1. Terminal is used by developers. Guess what windows also has terminal now via windows subsystem for Linux. You really seem to have no idea what you are taking about here...

2. haven’t heard of spotlight?

1. How is MacOS related to keyboards? It’s hardware vs software. Not related.

2. I doubt you actually build iOS apps. As you don’t know terminal vs dos vs powershell vs wsl. Again how is it related to MacOS?
 
  • Like
Reactions: ipxsystems
You really think so? Macs have about 10% market share, IF 10% of Mac users were interested in this, and I doubt it’s that many ... 1% of the PC market, do t think that’s interesting for Microsoft...
But it’s an interesting step, maybe an opportunity for Arm in the PC market overall...

It depends on how you look at it. Certainly, of the 18 million or so Mac computers sold per year, if 10% of that figure (1,800,000) wanted to run Windows software, it would hold a small amount of interest for Microsoft to port Windows over to Mac.

However, supposing that 10% per annum, keep a Mac for, let’s say 4 years, a figure Apple has hinted at in the past as an average. That becomes a rolling customer base of 7,200,000 - although the real figure would almost certainly be slightly higher. Which is a much more interesting proposition.
 
3 years too late ??

Worse than that as this is only available to Windows Insiders so will probably be another year before it's included in a general release Windows update. My guess is it will perform as badly as x86 emulation. As a Microsoft fanboy it really saddens me they have had such terrible results with ARM, and this has been going on for a LONG time. I'd definitely switch over to Apple hardware but unfortunately Apple does not have a touchscreen device that would run Windows, and of course Microsoft does not license an ARM version of Windows.
 
I think a big part of this question is “why do you want to do it?”
And we have different answers from different people.
1. I want to boot into Windows x86. -not likely
2. I want to boot into Windows ARM and run all Windows apps. - possible but hampered
3. I want to run windows in window and use all apps. - likely coming
4. I want to run windows apps (non-games) in the MacOS environment. - likely possible

People who want A should buy Intel Mac or a PC.
For B, seems more of a hobby project
For C, that has generally been the USE case for VirtualPC, then Parallels and Fusion, and they will soldier on

For D, Fusion allows it to seem this way in unity mode, and it may be possible. Or there might be a binary translator similar to Rosetta2 created to do this better, but there are a lot of hurdles there due more to the Windows interface than hardware call translations.
 
It depends on how you look at it. Certainly, of the 18 million or so Mac computers sold per year, if 10% of that figure (1,800,000) wanted to run Windows software, it would hold a small amount of interest for Microsoft to port Windows over to Mac.

However, supposing that 10% per annum, keep a Mac for, let’s say 4 years, a figure Apple has hinted at in the past as an average. That becomes a rolling customer base of 7,200,000 - although the real figure would almost certainly be slightly higher. Which is a much more interesting proposition.
I don’t think it’s close to that high.

As for companies developing and maintaining an OS for such a niche?

Taco Bell discontinued a product they had sold for a generation, The Mexican Pizza, with somewhere between $250-$500M in annual sales, because it was a bit more cumbersome to make. And there were no updates (other than taking off the scallions) and no security to maintain on a food item.

Companies don’t like niche products that require extra labor for a marginal payoff, no matter how popular the product is, was or might be.
 
1. Terminal is used by developers. Guess what windows also has terminal now via windows subsystem for Linux. You really seem to have no idea what you are taking about here...

2. haven’t heard of spotlight?

1. How is MacOS related to keyboards? It’s hardware vs software. Not related.

2. I doubt you actually build iOS apps. As you don’t know terminal vs dos vs powershell vs wsl. Again how is it related to MacOS?

1. We don't call terminal, either call PowerShell or command prompt. Yes, I do install the Linux (ubuntu) one ..
In the old-time, I used edit and when a new era emerges most will use notepad filename to edit file.

2. Spotlight not for the folder.
E.g I clicked on the search function and I want to find index.php example

What will I see is a hundred of index.php instead of supposing normal is one not in a subfolder. If you using "locate" at least it will give the path of index.php which totally confuses you.

3. I write a lot of code, which makes the keyboard broke also.

4. Either you doubt or not important, I don't know you and I no need to share my swift file to you.
 
It depends on how you look at it. Certainly, of the 18 million or so Mac computers sold per year, if 10% of that figure (1,800,000) wanted to run Windows software, it would hold a small amount of interest for Microsoft to port Windows over to Mac.
While I think that 10% is a very high estimate, at retail, that would be $180,000,000, per year. How interesting that would be Microsoft would really depend on two things: How much it would cost them create and maintain a port, and the cost to them of people replacing Microsoft (e.g. Visio and Project) and non-Microsoft applications that run on Windows with alternatives that run on other platforms (including web versions).

The first is easy to calculate, the second is harder.

Under Steve Balmer, Microsoft had a version of Office for iOS ready to go, but he would not let them ship it until the had a version for Window 8 touch screen devices. What happened is that many people who were using iPads as primary computing devices discovered that they did not need to have Microsoft Office (Apple’s iWork and Google Docs were good enough). Instead of having the desired effect of forcing people to use Windows, they lost users of Office.

The risk to Microsoft is that, if Apple Silicon Macintosh systems are compelling enough, people will find (or force to be created), macOS/iOS/iPadOS applications to replace ones that were previously Windows only. Each time that happens, there is the risk that those users will be permanently lost to them.

Calculating that risk is hard.
However, supposing that 10% per annum, keep a Mac for, let’s say 4 years, a figure Apple has hinted at in the past as an average. That becomes a rolling customer base of 7,200,000 - although the real figure would almost certainly be slightly higher. Which is a much more interesting proposition.
That makes sense for calculating subscription service revenue, but not for calculating individual purchase revenue. Windows 10 licenses are a single purchase (or are part of a Microsoft 365 subscription). The single purchase is the only revenue that would require Windows to be running on Apple Silicon Macintosh Systems (as the subscription also includes Office for the Mac).
 
  • Like
Reactions: alien3dx
I'm looking sad how Apple kills the real possibility for gaming with a good performance reached with the lasts 16" inch versions avoiding run natively Windows on ASi machines. All of these yearswating it and then, when they release a good one, they sentence to die that system.
 
Worse than that as this is only available to Windows Insiders so will probably be another year before it's included in a general release Windows update. My guess is it will perform as badly as x86 emulation. As a Microsoft fanboy it really saddens me they have had such terrible results with ARM, and this has been going on for a LONG time. I'd definitely switch over to Apple hardware but unfortunately Apple does not have a touchscreen device that would run Windows, and of course Microsoft does not license an ARM version of Windows.

x86 and x64 emulation performance is excellent under Windows ARM. Did you even bother to check the numbers? Most benchmarks run between 50% and 70% of native performance - which is probably as good as it gets for a dynamic translation solution.
Also x64 code is easier to emulate than x86 code, so it comes with no surprise that x64 emulation is faster - as initial tests show. Main reason is, that x86 code contains significantly more memory references compared to x64 code - which is sub-optimal when running on ARM.
 
Last edited:
I don’t think it’s close to that high.

As for companies developing and maintaining an OS for such a niche?

Taco Bell discontinued a product they had sold for a generation, The Mexican Pizza, with somewhere between $250-$500M in annual sales, because it was a bit more cumbersome to make. And there were no updates (other than taking off the scallions) and no security to maintain on a food item.

Companies don’t like niche products that require extra labor for a marginal payoff, no matter how popular the product is, was or might be.

I agree, I doubt the numbers would be that high, but I was using the same percentage as the post I was replying to.

I’m also in agreement that there’s little interest in a lot of work, for a niche product. Having said that, the cost of a Windows 10 licence is what? $140 or so. If people are legitimately buying a licence, and we all know not everyone would. It could quite feasibly scale to an excess of $1bn. A number many companies can only dream of.

To Microsoft on the other hand, it’s a good chunk of change certainly, but not the be all and end all.

What it would boil down to is fairly simple. They are making an ARM version of Windows already. We know the M1 is technically capable of running an ARM Windows, Apple have confirmed that.

So to Microsoft it becomes, is the almost $2bn revenue from sales of our entire Surface lineup - including Intel, enough to ignore a small adaptation to our existing ARM variant to ignore an additional revenue stream which could be anything from a quarter to a half of our entire Surface revenue?

Moreover, the revenue of Microsoft’s existing ARM variant, when extracted from the rest of the Surface lineup, is at this point probably low enough that the potential income from adding Apple compatibility is quite appealing. Of course, we don’t know exact sales of their ARM variant, they don’t break the figures down. But time will tell I suppose. And obviously, using Surface as an example as that’s what Windows on arm is mostly available on at the moment. As more and more ARM Windows systems from other manufacturers become available, the licence fees for those must be taken into account.

Obviously, the real number of Apple ARM users who wish to run Windows is frankly, anyone’s guess at the moment. The vast majority of Mac buyers likely have no interest in Windows whatsoever. But y’know, Devi’s advocate and all.
 
How much it would cost them create and maintain a port, and the cost to them of people replacing Microsoft (e.g. Visio and Project) and non-Microsoft applications that run on Windows with alternatives that run on other platforms (including web versions).

What port are you talking about? There is a Windows version for ARMv8-A already. You only have to port and maintain the drivers for a particular HW. However Microsoft never provides drivers (aside from some generic drivers, like input, display and mass storage etc.) - they need to come from OEM. Even for Surface Pro X - all the drivers like modem driver, sound driver, WiFi/GPS, DX12 etc. are all from Qualcomm.
What Microsoft is doing - it is providing the DDK (device driver development kit), which is fully supported for ARM64.
 
What port are you talking about?
Someone needs to do the work to support Windows on Arm on Apple Silicon. Whether that is properly called a port or not is semantics.
There is a Windows version for ARMv8-A already. You only have to port and maintain the drivers for a particular HW.
No one has argued this port will be hard, just that it has to be done.
However Microsoft never provides drivers (aside from some generic drivers, like input, display and mass storage etc.) - they need to come from OEM. Even for Surface Pro X - all the drivers like modem driver, sound driver, WiFi/GPS, DX12 etc. are all from Qualcomm.
The question is who would do this work (both writing the code and the final integration)? It could be Microsoft, or a third party like Parallels or VMWare.
What Microsoft is doing - it is providing the DDK (device driver development kit), which is fully supported for ARM64.
Again, someone has to do the integration to make it all work. Apple has not expressed a lot on interest in this, so it likely needs to either be Microsoft or a third party.
 
As for companies developing and maintaining an OS for such a niche?
They already support Windows on Arm. This would just be an integration project.
Taco Bell discontinued a product they had sold for a generation, The Mexican Pizza, with somewhere between $250-$500M in annual sales, because it was a bit more cumbersome to make. And there were no updates (other than taking off the scallions) and no security to maintain on a food item.
Taco Bell discontinued a product that required that they take up space in every restaurant for equipment, maintain a separate storage for the supplies, sourcing for those supplies, required on going training for all staff in all locations.

Assuming a team of 30 people (probably high) including driver developers, QA staff, admin and product marketing, at an average fully burdened cost of $200,000 (also probably high), it would cost $6 million. Using a figure of 5% (half what was being discussed), and a sale price of $100, they would bring in $90 million and earn $84 million an RoI of 14x. Not bad incremental revenue with other potential benefits for them.
Companies don’t like niche products that require extra labor for a marginal payoff, no matter how popular the product is, was or might be.
Software companies have lots of niche products, as their cost of goods is so small, it often makes sense.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.