Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
EXPOSING APPLE'S NEW M1 - A THREAT TO HOW WE USE OUR COMPUTERS.


DEBUNKING M1'S SPEED - Apple is lying.
Imagine how much faster the M1 chip would be if it didn’t have to waste all of that time and resources sending all your personal data to Apple!
 
  • Haha
Reactions: Reason077
EXPOSING APPLE'S NEW M1 - A THREAT TO HOW WE USE OUR COMPUTERS.


DEBUNKING M1'S SPEED - Apple is lying.


Both of these have been pretty well debunked.

Apple checks for revoked developer certificates by sending a hash of the certificate to their servers.
1) This protects you from malware.
2) Apple can' tell what apps you launch from this, the vast majority of developer certificates are used for multiple apps, they can only tell you launched an app from that developer.
3) It doesn't even check every time, the OS caches the result and doesn't check a certificate if it's been recently checked.
4) The only "personal information" is your IP address, which a rough location can be estimated from. You are spewing your IP address across the internet when you visit any website. It sucks that this works outside of an active VPN if you use one to hide your IP address, but you can also permanently turn off the checks in terminal.

The Tech Times article is just a summary of a PC World Editorial that had no factual foundation. The editor at PC World listed 3 of the highest performing Windows PCs, which clearly reside in the top 2% and offered no benchmarks to show that even they could keep up with an M1.

The benchmarks are all over the internet, the 10 watt M1 is crushing virtually every laptop and most desktop PCs, and even keeping up with top end PCs using 65+ watt CPUs and 100+ watt GPUs.
 
I'd bet MS will do this and knock it out of the park. They've been superbly managed over the last 7 years or so.
 
Parallels demonstrated this at WWDC so yes!
They demonstrated a VM, not dual-boot.
I don’t see why Microsoft wouldn’t license out ARM windows. The market would be huge especially when the pro machines (i.e. Mac Pro with apple silicon come into the market).
Agreed. The benefits seem to outweigh the costs. Maybe it won't be a particularly well-optimized version of Windows (I'm guessing it wouldn't, say, support the Neural Engine; it might not even support GPU acceleration), but getting basic support for running Windows apps working in a VM seems like something Microsoft would want.
 
  • Like
Reactions: Reason077
No-one has mentioned cellular data connection yet. Windows 10 for ARM specifications require laptops to have cellular data connection, yet the M1 laptops have none! So MS would need to add "except for Windows 10 ARM on Mac" into all their documentation if they'd decide to support Macs - probably not worth it for them. (Nice to see that Microsoft also promotes instant-on and beyond-all-day battery life for their ARM based OS!)
 
This is interesting. I didn’t even know there was an ARM version of Windows.

If I get an M1 MBP, I’ll still have my spiffy new 6-core Intel mini for virtualization, but it’ll be interesting to see what the future holds here.

There is an ARM version of Windows but it had issues and there were application compatibility issues.

It must grind MS a little bit to see Apple migrate its desktop OS and application suite without any performance or compatibility hits.
 
Why is that an argument? Microsoft explicitly allows Windows to run on non-Microsoft hardware. It's their core business model.
Microsoft core business is Azure and Microsoft 365. Windows has become one of many service delivery platforms, and is still their main developer platform. On Apple devices, you can consume their services using native iOS and macOS apps. What would Microsoft gain by supporting Windows on Apple silicon devices? Oh, you need to run a legacy Windows app? They've got you covered ;)
 
You seem to be arguing why it is in Apple’s interest to allow other operating systems to boot, not why it is in Microsoft’s interest.

Good point. However, I bet Microsoft is interested to see how well it would work given their lackluster results with Windows ARM.
 
That is what can be done at the moment (technically speaking). Natively is another matter.
I guess my point would be right on Intel Mac's, the OOTB, 1st party solution is Bootcamp which requires restarting and is an OR decision at startup. Do you want MacOS or Windows?

With ARM Mac's and an x86 copy of Windows, will it still follow that path or will it just be a program in MacOS like Parallels is right now that lets you run both at the same time.

And with an ARM copy of Windows, what will it look like? Will it go back to the OR choice at startup or can you run ARM Windows at any time you want in MacOS like Parallels.
 
The WWDC AS/Linux demo used Parallels. How exactly does Parallels work? I had previously thought Parallels wasn't an emulator, and thus the OS's it virtualizes need to be able to run natively. Thus I thought the fact that Parallels could run Linux on an AS Mac meant Linux can be run natively there.
....
However, I just found this article, indicating Parallels is in fact an emulator (
https://www.parallels.com/blogs/what-is-a-virtual-machine/?amp ), i.e., that rather than creating a virtual partition in which Windows/Linux is running natively, it creates an x86 hardware emulation layer on top of which Windows/Linux can run. This would mean that, when Apple was demonstrating Linux-through-Parallels at WWDC, it actually wasn't running natively, but rather on top of Parallels' x86 hardware emulation layer.

So I'm confused.

That Parallels blog page .... go to the bottom and follow the link there.

" ... Want more? Deep dive into the technology of virtualization and learn here: “What is Virtualization?” .."


Running an completely operating system on top of a virtual machine is more than solely executing CPU opcode calls. Applications invoke GPUs. They make low level calls to other devices.

If the CPU provides built in hardware to make virtualization more transparent/lightweight task but the other components don't then what will end up with is a hybrid of a complete virtual machine. The application CPU opcodes can run on the virtual CPU mode provided by the CPU hardware and the rest ... is done more the "old fashioned" emulation way.

Technically, Parallels can weave back and forth in using the terms at the 10,000 ft level when what they are burying is the fact the implementation has to deal multiple aspects of a complete systems. When Parallels is using emulation they are trying to use the term in a more accessible and generalized sense. If you use emulation is the broadest way possible, you can sweep in virtualization. The virtualization article goes more into the details. In a context of "emulation" versus "virtualization" usually the definition of "emulation" shifts to the more narrow one to describe the non hardware assisted techniques. (and "virtualization" pragmatically these days on contemporary processors takes on the hardware assist aspects also. ).

That's in part why Apple's Hypervisor solution is relatively behind the times. There are GPUs now that can provision out a hardware accelerated virtualize access. There are hypervisors that can assign devices to individual hosted OS images for almost 99.9% management.


For the modern versions ( not way back in pre-2006-9 era ) , the way that VMWare Fusion and Parallel VMs worked was by putting a type 2 hypervisor into the macOS kernel by adding an extension(s). Apple is killing off extensions that directly modify the kernel space (and kernel code). So the virtual machines going forward are layered on top of Apple's "hypervisor framework". But Apple's framework is limited. There are still value add parts that the VM vendors will have to add to make them work as before. Some of that has to do with emulating GPUs, boot environments , etc.
 
No-one has mentioned cellular data connection yet. Windows 10 for ARM specifications require laptops to have cellular data connection, yet the M1 laptops have none! So MS would need to add "except for Windows 10 ARM on Mac" into all their documentation if they'd decide to support Macs - probably not worth it for them. (Nice to see that Microsoft also promotes instant-on and beyond-all-day battery life for their ARM based OS!)

Requiring a cellular data connection is a terrible idea. Qualcomm licensing fees are ridiculously expensive for PCs, and they are forcing that cost on the 90% of laptop users who don't need or want it.

This is why Apple is doing the intelligent thing and waiting until their own modem team is ready before even thinking about baking cellular modems into Apple Silicon. Other PC makers don't have that luxury.

And Raspberry Pi has instant-on and beyond all day battery life, Microsoft should be focusing on building ARM chips that actually have decent performance and getting developers to build software that runs on their ARM based Windows. Otherwise they will be stuck serving only their tiny installed base.
 
Only $1.5 billion according to The Verge, and not their biggest quarter. Down from $1.86 billion in Q2 of FY 2020 according to PC World.

Year over year is usually more information. Unless the product introduction dates have moved around more than a bit. New product introductions can easily move "best quarter" to different parts of the year. A quarter X with 90% coverage with a new production introduction is unlikely to match the another quarter with the last dregs an old as possible production line.

Microsoft's fiscal year runs from July 1 to June 30. Q2FY20 puts them Q4 calendar year 2019.

Which was a big intro quarter.....



Microsoft has not finished Q2 FY 2021. We are currently in FY 2021. The article mentions

( PC world article published January 2019. )
"...” to $1.86 billion, Microsoft chief financial officer Amy Hood said during an analyst call covering Microsoft’s second fiscal quarter of 2019. ..."

that's '19 not '20 .


Which was another "big" Q4 Calendar year 2018 event.

https://news.microsoft.com/october-2-2018/


From the Verge article.

".. Surface revenue has jumped by 37 percent this quarter to $1.5 billion. That’s a big increase for a quarter that hasn’t seen any new Surface devices introduced. Microsoft only just introduced a new Surface Laptop Go device and updated Surface Pro X earlier this month, but those will count to next quarter’s revenue. ..."

The other comparisons in the Verge article are "year over year". I did't look it up but I suspect they are not flip-flopping year inside of a calendar year comparisons. For that quarter in the year given the calendar Q4 introduction cadence ... it has a pretty good chance of being "biggest".

Other factors can creep in ( Covid adaptions driven purchases ) , but comparisons outside of context are going to drift into Apples vs Oranges comparisons.

Too early to tell Microsoft isn't going to see an increase in 2QFY21. Surface Pro X speed bump is a bit weak and Surface Go is a bit "blah" too.
 
There’s an ARM version of Windows that runs fine on a Surface tablet, but I’m not aware of anyone virtualizing it.

Other than Microsoft ....

".... Back in February with Build 19559, we added the ability to install Hyper-V on ARM64 devices such as the Surface Pro X running the Enterprise or Pro editions of Windows 10. Today, we are offering the ARM64 VHDX for Windows 10 Insider Preview Build 19624, so that Insiders can run Windows 10 as a guest OS in Hyper-V. You can download the VHDX here. We will regularly release updated VHDX downloads for newer Insider Preview builds going forward. ..."


Microsoft previously stumbled here (large because Qualcomm stumbled here)

https://www.lightreading.com/enterp...oft-qualcomm-bring-arm-to-azure/d/d-id/730959

but there are several public indicators that they are going to be deploying substantive ARM64 several nodes into Azure. That Microsoft won't run Windows as a service at all on any of those would be a head scratcher. If they can lower the operating cost of those instances they likely will sell more time on them. Couple with the "regular distribution of VHDX " images and they'd be avoiding synergies to avoid the functionality present.


There are several other players. Other C8X systems.

Samsung
https://hothardware.com/reviews/samsung-galaxy-book-s-arm-vs-x86

Lenovo Flex 5G

Qualcomms 8CX gen 2


The have been some not so well supported hacks with QEMU that have had Windows on Arm running.
 
I don't think this is true. HP and Dell already exist and eat the Surface program's lunch. Apple would be the same influence on Windows they are now. Not very big...

You're forgetting one thing: in years gone by (i.e., in reference to your "HP and Dell already exist...", Apple didn't have a 2-3x performance advantage in terms of performance per watt. HP/Dell do not have 2-3x performance per watt improvement either.

Native Windows apps on M1 would have that.

Microsoft will go where the market goes. They're an apps/services company these days, if there is demand for Windows on M1, it will come. Wouldn't even surprise me if they license Apple hardware at some point, if Apple maintain this sort of tech lead.

Not saying it will happen, it just would not surprise me if it did, is all. Because they aren't as tightly bound to intel as they once were.
 
Last edited:
Yes but it was not the full version of windows and not the full version of their primary applications.

See post 267 above yours. There are both Enterprise and Pro versions of Windows on ARM64. How "not full version is that" ? Windows Server ? Microsoft is dropping updates.

".., 2020-11 Servicing Stack Update for Windows Server, version 2004 for ARM64-based Systems (KB4586864) ..."
https://www.catalog.update.microsof...updateid=4fd5881f-0632-4d79-ba0b-f8c90e814737


If you are trying to reference Windows RT that was long , long ago. That hasn't been the baseline for Windows on ARM for years.

Microsoft's applications run fine. (e.g, recently Teams. Office 365 Enterprise )


Third party plug-ins (and extensions) and odd ball software that is still 32-bit is the bigger hang up. ( MS runs Office 365 inside of a wrapper to interface funky plug-in 23 that is still x86. ) The larger gap that Microsoft needs to close is x86_64 emulation/translation for software that hasn't been ported. ( I suspect will run into similar issues that Apple did with AVX and apps/extensions tied to hardware virtualization . And Apple just labels plug-in that don't get with the migration movement as 'broken" and just don't go through giant gyrations to accommodate them. At this point MS should get on that boar. Slacker plug-in developers are just boat anchors after a while. Just twisting yourself up into knots for people who just probably aren't going to do the work. )
 
Last edited:
No-one has mentioned cellular data connection yet. Windows 10 for ARM specifications require laptops to have cellular data connection, yet the M1 laptops have none!

Require? Pragmatically, the only ARM SoC vendor for the laptop form factor space for Windows 10 has been Qualcomm. I don't think that is actually a requirement. It is more so a sales pitch to extol a feature of Qualcomm's SoC that come with that feature bolted on (no choice). Yeah, Qualcomm requires you buy the radio. And because the performance gap is usually relatively big Microsoft needs a "yeah it is slower but has this other aspect as a tradeoff" feature. That isn't a requirement of the OS. That is more of a requirement of having to put on some tap dance show to make some sales.


If Microsoft was calling 100% of the specs on the SoC then I doubt it would be a requirement. It isn't on the Windows 10 Server ARM64 version they use .


For the general PC market there is basically a large "hole" in the ARM SoC vendor supply chain. There are bunch of folks off chasing high end servers and specific embedded niches. Then there is a large group chasing Smartphones and small tablets. Apple isn't licensing their tech.

Things might get better with ARM's Cortex-X program and the X1 or X2. Perhaps Microsoft will do a direct pull on implementation and do their own ( if Qualcomm can't get something together that leaves the smartphone space. )


Microsoft is also tightrope walking line between agitating Intel (and AMD) [and their host of system maker partners. .. who largely just make x86 boxes ) into action and opening the door for ARM SoC vendors. This "always on" is a contrast the bulk of the other stuff that is the core product. So carving out a new niche is less threatening to the partners. Again not really a requirement of the OS.


So MS would need to add "except for Windows 10 ARM on Mac" into all their documentation if they'd decide to support Macs

Errr, those systems sell as Windows 10 system. No Mac is going to sell as a Windows 10 system. Highly unlikely any Mac is going to boot as a Windows 10 system with M-series and iPhone boot security.

When folks fire up a virtual machine they know they aren't on the raw hardware so there is no huge reason to change the FAQ docs.


- probably not worth it for them. (Nice to see that Microsoft also promotes instant-on and beyond-all-day battery life for their ARM based OS!)

Microsoft get Apple to sell Windows systems? It didn't happen for x86 "boxes" and it sure isn't going to happen now.

Apple go through the hoops to get Windows boot natively on M-series systems. Why? Apple sells more ARM based systems than Microsoft does. It isn't like Apple is hurting for sales volume. It isn't like the x86 transition where they are jumping into someone else's big pond. MS has had 2-3 years to get their ARM laptop business off the ground and ... it really hasn't gone all that far (relative to the rest of their business).

Lots of that is squandered effort tryign to bend into pretzels to keep folks with quirky, ancient 32-bit x86 programs and plug-ins happy. And putting minimal customization effort in so have relatively middling smartphone SoCs to work with.

Why should Apple help Microsoft get out of a hoe they put themselves into. Help windows get bigger market share to the detriment of macOS. That really don't make any sense. Apple provides a virtual "box" that Windows 10 can run inside of. It is up to Microsoft to get Win10 to run in that box well with a evolving complete set of native apps. If MS can't get the apps to run there be a user satisfying environment then that is on them.
 
Xbox GamePass

They'd also gain windows sales on hardware that has better performance per watt than anything else on the market in the 13" notebook space.

Some people may not care for Windows enough to give up that performance. If windows is available, they may purchase it. I suspect there's plenty of users who would love M1 hardware "if only it ran windows"
 
See post 267 above yours. There are both Enterprise and Pro versions of Windows on ARM64. How "not full version is that" ? Windows Server ? Microsoft is dropping updates.

".., 2020-11 Servicing Stack Update for Windows Server, version 2004 for ARM64-based Systems (KB4586864) ..."
https://www.catalog.update.microsof...updateid=4fd5881f-0632-4d79-ba0b-f8c90e814737


If you are trying to reference Windows RT that was long , long ago. That hasn't been the baseline for Windows on ARM for years.

Microsoft's applications run fine. (e.g, recently Teams. Office 365 Enterprise )


Third party plug-ins (and extensions) and odd ball software that is still 32-bit is the bigger hang up. ( MS runs Office 365 inside of a wrapper to interface funky plug-in 23 that is still x86. ) The larger gap that Microsoft needs to close is x86_64 emulation/translation for software that hasn't been ported. ( I suspect will run into similar issues that Apple did with AVX and apps/extensions tied to hardware virtualization . And Apple just labels plug-in that don't get with the migration movement as 'broken" and just don't go through giant gyrations to accommodate them. At this point MS should get on that boar. Slacker plug-in developers are just boat anchors after a while. Just twisting yourself up into knots for people who just probably aren't going to do the work. )

OK you are correct - I stand corrected on that point. However, doing some follow up research there were serious performance issues on ARM for full Windows. They could not get the same level of performance you would get on x86 architecture. This points to the advantage of Apple's ARM approach and points to Microsoft's interest in seeing how the M1 processor would support windows.
 
However, doing some follow up research there were serious performance issues on ARM for full Windows.
Sure, but that doesn't have anything to do with the completeness of Windows. They boil down to:

  • the CPU used (Microsoft SQ2; sort of related to the Qualcomm Snapdragon chips in many Android phones)
  • the emulator
An M1 is much faster. The emulator would also be much faster on it. Fast enough? Hard to say.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.