I predict Boot Camp & Windows 11 are coming to Apple Silicon.

[leman]

How many more of me exist? I don't know, so I don't know if that's a big enough slice of the pie that Apple is interested in claiming. I wonder how many other folks like me exist but who haven't been indoctrinated in the Apple Ecosystem and might convert over given the exposure/opportunity.

The problem is that developing and maintaining a full set of drivers for Windows is a significant long term effort, which likely overshadows all positive marketing Apple would get from it. Especially given that we have readily available VM solutions that satisfy most of the Windows compatibility demand.

If ARM architecture really is as gangbusters and amazing as the frenzy surrounding it seems to be, it feels somewhat innevitable that consumer tech makes a full transition.

ARM architecture doesn’t matter much, implementation does. Apple Silicon is not fast and power efficient because it’s ARM, but because Apple is ahead of the chip-designing game and spares no expenses when it comes to build fast CPUs.

 

[Bodhitree]

Yes, Microsoft will realize that Windows on ARM isn't catching up and they will shut it down like Windows Phone OS. That's the future.

You need to see the long-term industry wide trends. Computing devices are getting smaller, from desktops to laptops to mobiles, and at the smaller sizes performance per watt becomes much more important. ARM has a clear advantage there, and you’d have to consider the future is bright.

We will have to see whether the Nuvia-powered designs for Qualcomm SoCs will move the market, convince business to move to Windows on ARM. Of course the software has to be there, but it’s at its core about size, heat, device lifespan and battery life.

Apple is at the leading edge of this move, they are attempting to prove that it is possible to build laptops that run cool, quiet, fast and with great battery life, and that last for 8-10 years.

 

[genexx]

I am using Windows 11 ARM via UTM on M2 and getting always the same updates as on the Real PC Laptop.
This Windows 11 ARM is also very fast and i have no downside in using even an OLD VB6 Proggy with just registering a dll and this is connecting to Maria DB SQL running on XAMPP via ODBC 32Bit Maria SQL Adapter and saving an Excel Datasheet.
Office 2021 as well as Browsers and many other Proggys just work.

As i do not run Games i have no Problems and all is up and running freakin fast.

I would never choose to Reboot to get Windows instead of having both up and Running @ the same Time.
Before the M2 i always connected to an Lenovo AMD 4700U Laptop via RemoteDesktop but not anymore.

 

[Gudi]

You need to see the long-term industry wide trends. Computing devices are getting smaller, from desktops to laptops to mobiles, and at the smaller sizes performance per watt becomes much more important. ARM has a clear advantage there, and you’d have to consider the future is bright.

Nobody doubts that the ARM architecture is the future or rather the present of mobile devices. But Windows on ARM will go nowhere, because of Windows. Android won the battle for being an iOS alternative for OEMs. Windows will forever be stuck on their Intel monopoly, for as long as it lasts and then it will die.

 

[Bodhitree]

Which means it rather depends on how well Intel does in keeping x86 as a dominant hardware architecture. At the moment they are struggling with a number of disadvantages such as being several process nodes behind, but it looks like they will catch up. That’s just one step but they may be able to make the architecture more power performant. You can already get full x86-64 processors in tablets running Windows.

 

[theluggage]

ARM architecture doesn’t matter much, implementation does. Apple Silicon is not fast and power efficient because it’s ARM, but because Apple is ahead of the chip-designing game and spares no expenses when it comes to build fast CPUs.

I think its a bit of both - x86 has a complex instruction set and a huge backward-compatibility albatross around its neck, so in comparable x86 and ARM implementations, ARM is likely to be simpler (or have more room to add optimisations).

ARM was dominating the mobile market long before Apple Silicon - or even the iPhone - came along, partly through power efficiency, partly because x86's only unassailable advantage was software compatibility - which isn't such an advantage on a mobile device that just isn't physically suited to running PC software. Even before iOS/Android there were things like EPOC/Symbian with roots in ARM.

The other unique thing about ARM is that competing chip makers can license the tech on a "pick'n'mix" basis - which is why Apple were able to make their own tailored chips with the ARM instruction set - so there are lots of independent players like Apple, Qualcomm, Amazon, Ampere, NVIDIA etc. working on ARM whereas x86 is basically "Coke or Pepsi" - er, sorry, Intel or AMD.

So x86 doesn't just have to stay ahead of Apple, it has to beat all of the other players in the ARM business. Apple don't really have a dog in the high-performance computing or server race (as that's a tiny fraction of their market that they can afford to lose) but NVIDIA and Amazon certainly do - and, like the mobile market, the days of servers/HPC being dependent on running Windows are coming to an end with Linux taking a huge share - Linux has supported ARM for years and Linux/Unix developers have a culture of writing portable code.

I think there may be a couple of reasons why Microsoft aren't falling over themselves to support bare-metal ARM Windows on Mac (apart from the fact that it would not be the trivial job it was with x86 Boot Camp). First is that, right now, an M2 running native Windows would probably leave a Surface X choking on its dust. Yay for Apple, but a problem for Microsoft who's Windows business model still relies heavily on licensing it to OEMs, and while OEMs can build Qualcomm-based WoA PCs, they won't be able to make M2 based PCs... embarrassing for MS.

Second - MS rely heavily on Windows' backwards compatibility to keep their Windows business going. Apple was able to "smooth" the x86 to ARM transition by being pretty ruthless about dropping support for older software (including completely axing 32 bit support) - MS probably can't afford to do that - either by decree, or by offering a really enticing and powerful WoA option that might encourage people to re-tool (like a 16" MBP or a Mac Studio running WoA natively). I think at this stage MS are hedging their bets by just having the Surface X as a just-about-credible tablet alternative that is of little interest to most Windows users.

I think we're long past "peak x86" and "peak Windows" (once you look beyond traditional PCs) but "Wintel" was so dominant in the 90s/00s it won't be going away quickly. Long term, though, I think they'll fade into irrelevance as most software becomes processor independent, and only the writers of operating systems, language runtimes and web browsers have to worry about processor architecture. We're not there yet but it isn't far off.

 

[dmccloud]

I think its a bit of both - x86 has a complex instruction set and a huge backward-compatibility albatross around its neck, so in comparable x86 and ARM implementations, ARM is likely to be simpler (or have more room to add optimisations).

ARM was dominating the mobile market long before Apple Silicon - or even the iPhone - came along, partly through power efficiency, partly because x86's only unassailable advantage was software compatibility - which isn't such an advantage on a mobile device that just isn't physically suited to running PC software. Even before iOS/Android there were things like EPOC/Symbian with roots in ARM.

The other unique thing about ARM is that competing chip makers can license the tech on a "pick'n'mix" basis - which is why Apple were able to make their own tailored chips with the ARM instruction set - so there are lots of independent players like Apple, Qualcomm, Amazon, Ampere, NVIDIA etc. working on ARM whereas x86 is basically "Coke or Pepsi" - er, sorry, Intel or AMD.

There is an important difference between Apple's ISA license and other companies' architecture licenses. Apple just licenses teh instruction set itself, and has the ability to create custom silicon that runs said instructions. Other companies license the architecture (i.e. core designs) and integrate those into an SoC.

So x86 doesn't just have to stay ahead of Apple, it has to beat all of the other players in the ARM business. Apple don't really have a dog in the high-performance computing or server race (as that's a tiny fraction of their market that they can afford to lose) but NVIDIA and Amazon certainly do - and, like the mobile market, the days of servers/HPC being dependent on running Windows are coming to an end with Linux taking a huge share - Linux has supported ARM for years and Linux/Unix developers have a culture of writing portable code.

I think there may be a couple of reasons why Microsoft aren't falling over themselves to support bare-metal ARM Windows on Mac (apart from the fact that it would not be the trivial job it was with x86 Boot Camp). First is that, right now, an M2 running native Windows would probably leave a Surface X choking on its dust. Yay for Apple, but a problem for Microsoft who's Windows business model still relies heavily on licensing it to OEMs, and while OEMs can build Qualcomm-based WoA PCs, they won't be able to make M2 based PCs... embarrassing for MS.

Second - MS rely heavily on Windows' backwards compatibility to keep their Windows business going. Apple was able to "smooth" the x86 to ARM transition by being pretty ruthless about dropping support for older software (including completely axing 32 bit support) - MS probably can't afford to do that - either by decree, or by offering a really enticing and powerful WoA option that might encourage people to re-tool (like a 16" MBP or a Mac Studio running WoA natively). I think at this stage MS are hedging their bets by just having the Surface X as a just-about-credible tablet alternative that is of little interest to most Windows users.

I think we're long past "peak x86" and "peak Windows" (once you look beyond traditional PCs) but "Wintel" was so dominant in the 90s/00s it won't be going away quickly. Long term, though, I think they'll fade into irrelevance as most software becomes processor independent, and only the writers of operating systems, language runtimes and web browsers have to worry about processor architecture. We're not there yet but it isn't far off.

Given how Microsoft keeps changing Windows by adding even more junk and ads to the OS, it's only a matter of time before there is a mass defection from that OS to alternatives. My guess is that a lot of the gamers would go Linux because of existing support for WINE and gaming in general as well as driver support for modern videocards.

 

[theluggage]

There is an important difference between Apple's ISA license and other companies' architecture licenses. Apple just licenses teh instruction set itself, and has the ability to create custom silicon that runs said instructions. Other companies license the architecture (i.e. core designs) and integrate those into an SoC.

I don't think that's exclusive to Apple - ARM offer different levels of license.

Companies can also obtain an ARM architectural licence for designing their own CPU cores using the ARM instruction sets. These cores must comply fully with the ARM architecture. Companies that have designed cores that implement an ARM architecture include Apple, AppliedMicro (now: Ampere Computing), Broadcom, Cavium (now: Marvell), Digital Equipment Corporation, Intel, Nvidia, Qualcomm, Samsung Electronics, Fujitsu, and NUVIA Inc. (acquired by Qualcomm in 2021).

ARM architecture family - Wikipedia

en.wikipedia.org

(see also the entries for the other types of license preceding that)

Given how Microsoft keeps changing Windows by adding even more junk and ads to the OS, it's only a matter of time before there is a mass defection from that OS to alternatives.

Yeah, that's what everybody said around the time of Windows ME, Windows Vista, the Office ribbon interface, Windows 8...

 

[WilliApple]

If they haven't allowed boot camp to run Android on iPhone officially, then there is no way they will be developing Boot Camp for Apple Silicon.

 

[Chuckeee]

I have a more basic question.

“A” (if not “The”) motivation for Windows 11 ARM is to support Microsoft Surface tablets that use SQ1/SQ2 ARM Processors, that variant of Qualcomm’s Snapdragon processors.

My question is: how much commonality is there between Snapdragon ARM and the M1/M2/M3 (or even A16/A17) ARM processors? Are all of the ARM based CPU & GPU basically the same architecture with same/similar instructions sets?

 

[leman]

I have a more basic question.

“A” (if not “The”) motivation for Windows 11 ARM is to support Microsoft Surface tablets that use SQ1/SQ2 ARM Processors, that variant of Qualcomm’s Snapdragon processors.

My question is: how much commonality is there between Snapdragon ARM and the M1/M2/M3 (or even A16/A17) ARM processors? Are all of the ARM based CPU & GPU basically the same architecture with same/similar instructions sets?

What do you mean by “ARM based”
in your post? ARM can refer to the instruction set specification, or it can refer to CPU designs made by ARM that implement this instruction set.

Snapdragon and Apple CPUs use the same basic instruction set and can in principle execute the same code. Snapdragon uses CPU cores developed by ARM and M series uses CPU cores developed by Apple. M series low-level stuff (like things like CPU communication protocols/power management etc) are fully custom Apple and not compatible with other ARM CPUs. The GPU is fully custom and has nothing in common with other GPUs used in ARM SoCs.

To boot on M-series, Windows would likely need modifications to its kernel, a new OS loading module, and a full set of custom device drivers.

 

[leman]

I think its a bit of both - x86 has a complex instruction set and a huge backward-compatibility albatross around its neck, so in comparable x86 and ARM implementations, ARM is likely to be simpler (or have more room to add optimisations).

ARM was dominating the mobile market long before Apple Silicon - or even the iPhone - came along, partly through power efficiency, partly because x86's only unassailable advantage was software compatibility - which isn't such an advantage on a mobile device that just isn't physically suited to running PC software. Even before iOS/Android there were things like EPOC/Symbian with roots in ARM.

I am not sure that empirical data really confirms this analysis. It is entirely possible that x86 is fundamentally at a disadvantage when it comes to instruction decoding, the rest though? Both Intel and AMD seem to be able to innovate and break performance walls just fine. New compact Zen cores are a good example.

Probably the main reason why ARM dominates mobile is because Intel was not interested in that market. Their design strategy took them the opposite way. Now they are paying the price.

 

[Chuckeee]

What do you mean by “ARM based”
in your post?

As compared to x86

 

[leman]

As compared to x86

In terms of CPU instruction set, yes, it’s the same, if you disregard Apples custom instructions and protocols.

Maybe a more correct answer is that Apple CPUs follow ARM specification for the basic instruction set, but don’t always follow ARM’s system specifications ( stuff related to how the OS interfaces with hardware). It took the ASAHI Linux team a lot of work and many patches to the Linux kernel to boot M1 natively, and the inly way they could do it reasonably quickly is because they were already building upon years of iOS reverse engineering efforts.

 

[deconstruct60]

“A” (if not “The”) motivation for Windows 11 ARM is to support Microsoft Surface tablets that use SQ1/SQ2 ARM Processors, that variant of Qualcomm’s Snapdragon processors.

That is missing a big chunk of history. A hefty chunk of Microsoft wanting a 'Windows on Qualcomm's Arm" comes from the "Windows Phone" era. Transitions into "Windows RT" and then Windows 10/11. Microsoft has been whittling away at this as a 'side project' for far longer than Apple has. It was far less of a 'side project' for Apple once they decided to 'bet the farm' on it. Microsoft is extremely unlikely going to do that ( they are partially in the business os selling Windows licences to 'everybody' ; not to a handful of specific products. )

My question is: how much commonality is there between Snapdragon ARM and the M1/M2/M3 (or even A16/A17) ARM processors? Are all of the ARM based CPU & GPU basically the same architecture with same/similar instructions sets?

Neither Qualcomm's Snapdragon's or Apple M1 use Arm's GPU ( 'Mali' ). There is no unified coherence in Arm SoC ecosystem past what is classically called the 'CPU' (cores). [ even in the CPU cores there is a wide variance as there are multiple classes of Arm cores that target specific submarkets; each with their own license. There isn't a one get everything. ] The issue is that things really are a CPU (only ) package + GPU (only) package + ... . As things collapsed into a single package (SoC) Arm's mix-and-match flexibility allows a wide range of GPU options ( from none ... . to extremely proprietary).

 

[deconstruct60]

Probably the main reason why ARM dominates mobile is because Intel was not interested in that market. Their design strategy took them the opposite way. Now they are paying the price.

Intel was interested. They just had really bad offerings. The Intel CEO pitched Jobs on using Intel for the iPad when asked for a SoC for phone. But at the time


"... Contrary to popular belief, Intel wasn't caught completely off-guard by the rise of smartphones or the popularity of small, Internet-connected devices. Atom development began in 2004; the Silverthorne core that Intel debuted in 2008 had a TDP of just 2-3W at a time when most mobile Core 2 Duo processors were stuck in 35W territory ..."

How Intel Lost the Mobile Market, Part 2: The Rise and Neglect of Atom - ExtremeTech

In Part 2 we explore the specific decisions Intel made, the rise and neglect of Atom, and why the company's superior foundry technology wasn't enough to conquer the mobile market.

www.extremetech.com

This is roughly the same era where Intel was so self infatuated with x86 that they tried a GPU based on x86 cores.

https://en.wikipedia.org/wiki/Larrabee_(microarchitecture)

Part of Intel's problem was tried to make everything they did revovle solely around the single cash cow they had.

"... Intel's decision to sell its ARM division and XScale processor line in 2006 has been widely derided as a critical error. It's a simple, common-sense explanation with just one flaw: It mistakes symptoms for cause. ..."

If Intel would have attached the Imagination Tech GPU to the XScale Arm cores they had they would have walked away with all of Apple's inital business for several generations. But nooooooooo the margins were not 'fat enough' for them. Tried to make Apple 'eat' Intel's GPUs and x86 too.

Intel did scramble to do a PowerVR Atom but it didn't arrive until 2009-2010 which was way too late. Even if they had a better 'CPU' they still were committed to an even less competitive GPU in their SoC.

The core issue was that Intel didn't look at themselves as foundry maker , but as 'x86 maker'. Way too much drinking their own Santa Clara kool-aid. Poured 'x86' over everything possible like 'ketchup'.
(Microsoft had similar issues. The 'start button' on a phone screen .... really? )


P.S. some more x86 ketchup on their extreme mobile offering.

Intel launches Wireless MMX for XScale - EE Times

SAN JOSE, Calif. -- Intel has announced an addition to its XScale processor architecture in the form of Wireless MMX technology, a set of multimedia instructions taken in large part from the Intel Architecture. This should enable software engineers t

www.eetimes.com

 

[theotherphil]

I'm pretty sure Apple said less than 1% of Intel Macs actually used BootCamp. I can't see Windows running natively on Apple Silicon is worth it for Apple or MS

 

[dmccloud]

I don't think that's exclusive to Apple - ARM offer different levels of license.


(see also the entries for the other types of license preceding that)

Qualcomm is a strange hybrid, because they actually have dual licenses (one to use existing core designs, one for the ISA). However, Qualcomm is still locked into a legal battle with ARM regarding the NUVIA acquisition, so things could change overnight with that relationship.

While Qualcomm does brand cores with their own naming scheme, even the Kryo cores used in the SQ1 and SQ2 SoCs for Microsoft and Qualcomm's own Snapdragon 8 and 8 Gen 2 SoCs are actually Cortex A76 designs (performance cores) and Cortex A55 cores (for efficiency cores). To be honest, Qualcomm's modifications to ARM reference designs is really more about the interconnects between CPU, GPU, cellular, and other functions than it is about designing an entirely new architecture using the ARM ISA (which is how Apple has approached things for years).

There is one other difference between Apple's ISA license and Qualcomm's though. Apple has the ability to add their own instructions on top of the stock ARM ISA, and some of those have been backported into the actual ARM ISA. There is a recent interview with ARMs CEO where he touches upon the uniqueness of Apple's license with ARM.

Also, Apple apparently extended their ARM licensing agreement to run through at least sometime in the 2040s. From ARMs amended IPO filing with the SEC dated September 5, 2023:

Further, we have entered into a new long-term agreement with Apple that extends beyond 2040, continuing our longstanding relationship of collaboration with Apple and Apple’s access to the Arm architecture.

Yeah, that's what everybody said around the time of Windows ME, Windows Vista, the Office ribbon interface, Windows 8...

Microsoft has traditionally had a good release followed by a crap release dating back to Windows 98, where ME, Vista, 8, (and for many people 11) just sucked in one way or another. ME was a solution with no problem to resolve, and was quickly replaced by the far superior XP. Vista was a case of change for the sake of change, and Windows 8 suffered from Microsoft's desire to force a smartphone/tablet interface onto the desktop, which is still a terrible idea even today. With Windows 11, their biggest obstacle to adoption is the TPM requirement, because it has made many systems unable to run the new OS even though their CPU, GPU, etc. are far above the minimum or recommended specs for the OS. This has created a lot of consternation among the Windows user base because it's effectively creating forced obsolescence among a lot of PCs Microsoft has arbitrarily decided can't run Windows 11.