I like seeing new developments and all, but I continue to see Windows on Arm as dead. I'm sure, like Windows mobile, it'll stick around for a while, maybe get installed on a few million kiosks and other sorts of computers where being ultra-restricted and most stuff won't run is a feature, not a bug... But as a functional replacement for general work environments (both desktop and laptop)? No thanks. That is a very hard pass from me. Let me know when they exceed mid range x86-64 CPUs in terms of *both* singlethreaded and mulithreaded performance and have sufficient software compatibility that most random programs "just work".
You should read up CISC vs RISC instruction sets, CISC will ALWAYS be require more clock cycles than RISC. This is why an ARM64 core will ALWYAS be more efficient than an IA64 core. Any improvement in silicon manufacturing will yield GREATER improvements on ARM64 than IA64, it's a battle that IA64, by design, cannot win. Sure, IA64s have higher top clock speeds (eg. Core i9's 9Ghz vs Apple M4s 4Ghz), but that's because you need to perform MORE operations per watt on IA64, even then, the i9 can only run at t
The problem with that analysis is that the term IA64 specifically means the architecture used by the Intel Itanium chips, and not the x86 architecture with AMDs 64-bit extensions. These instruction sets are completely different and unrelated. You could argue that Intel and AMD should just drop the 32-bit, 16-bit, and 8-bit modes/instructions at some point, and you'd probably be right, but that still won't change the definition of IA64.
> If you find a good way (or Microsoft rebuilt the parts) to remove the 8088 liability
That's the kicker, they have A LOT of legacy software they (implicitly) support. However, they could invest and build a solid emulator to placate those with legacy software. Box64 already lets you run x86 Linux software on arm64 Linux, there's no reason MS couldn't do the same. Apple used a 2 pronged approach with Rosetta and a "fat" binaries that packaged multiple CPU targets. In theory, MS could adopt their existing M
And all the technical reasons why Linux is a better desktop OS than Windows or Mac are valid too. Doesn't change the fact that technical superiority alone is insufficient to drive adoption of a technology.
Except that the Apple M series chips are only good at Geekbench and Cinebench. Any real world application and it just isn't as fast as AMD's and Intel's chips. Especially gaming because IPC in gaming is not the same as synthetic benchmarks. Even the M4 which is scoring much higher in Geekbench is doing so because of a new extension called SME, which increases matrix multiplications which basically means AI. But yet somehow Geekbench gave Apple another 300 points for it. As for the differences between C
Not all software may be able to run on ARM, but as Microsoft improves their emulation layer and more businesses buy into the architecture, we should see a change. Apple showed with Rosetta that you can do high performance emulation of an architecture, when running applications designed for your own OS.
If you are thinking of games, then they may end up being the laggards in getting the basics support.
Iâ(TM)ll second that from experience. I have a M1 MacBook that runs a Windows ARM VM. Iâ(TM)m as big of Apple fan as youâ(TM)ll find, but I will say (for my use cases) Windows 11 ARMâ(TM)s emulation is very good. Almost Rosetta good.
Now, my use cases are small and niche. Windows ecosystem is vast and complex and I doubt that everyone will be as happy as I am. But both Apple and Microsoft are showing it can be done well. Appleâ(TM)s Game Porting Toolkit also shows that there are path
Have you used windows on arm? I received mine a couple weeks ago before my vacation. The experience was positive for the 2 days I used it. It is refreshing.
Arm doesn't have enough of the server and desktop market for Microsoft to risk restricting themselves to one vendor. Wrong question. And they already lost the handheld market - wrong product.
This is just part of why the premise is silly. The idea of ARM CPUs "taking off" with an OS tightly associated with closed-source x86 legacy code is highly questionable, not only would it incur performance overhead from emulation but it could add another layer of risk of breaking compatibility with legacy apps in the x86-on-ARM emulation layer.
If I had to make a bet about the future of Windows on ARM, I'd bet that it will be a niche product that eventually dies a quiet death like every other attempt to run Windows on an alternate architecture. Closed-source and multi-arch don't mix. Windows is wedded to x86.
In my opinion, the Legacy code is becoming more and more of a niche use. At work I have to deal with industrial PCs still running Windows XP Embedded that run software which is impossible to update. The system ISO files won't run in any modern hardware. If those machines break, we'll need to replace them. Same for some old testing machines with Sun computers running Solaris... last time we got lucky and could find second hand hardware on eBay, but the next time it breaks, we'll have to dump a lot of cash in
Therefore, I don't think the compatibility issue it's a problem for legacy. Nowadays, either you run it on legacy hardware, and when it breaks, you're done, or you can run it on a VM (and if that's the case, you don't care if the CPU is an x86, an a64, or a potato).
Oh it does matter. When running a VM of the same architecture that the hypervisor is running on (as in the typical x86 VM on an x86 host scenario) the CPU's virtualization extensions are used so that the VM's instructions are basically passed-through to the host CPU as-is and there's almost no performance hit. If you're running a VM of a different architecture, then you're back to square one and have to fully emulate the CPU for the VM which carries a massive performance hit.
I remember trying Ubuntu for arm64 some time ago, and despite being open source, the compatibility with applications was abysmal unless you started downloading source code and compiling every little widget and app you wanted to get except for the main ones.
You have a point. However, the software companies (and the OS ones) tend to use (and offer) standard libraries more and more. The more a software uses standard OS libraries, the easier it is to port to another architecture (simply needing the precompiled libraries for the new CPU).
Microsoft released this partial emulation mode where they can use the arm64 native libraries whenever possible, and just emulate the non-standard parts of the code. This reduces the impact of emulation, but in addition, it is a n
ARM is currently suing Qualcomm and has withdrawn their license to use ARM designs. Depending on how this shakes out, the chip industry may be totally re-aligned.
This really needed a "No, ARM is not going to take off for Windows" option... Microsoft has almost no real motivation to make sure this happens, even if platform makers want them to. They just don't. It's not going to take off.
It can be interpreted as "some other company will reap the benefit" - but I don't think that's going to happen, not for Windows.
Some other platform might rise up, Chrome, or Android, or something new - heck, maybe it's the year of desktop Linux - on ARM - who knows! But it won't be Windows.
Whatever incentive hardware makers had for producing ARM-based laptops/desktops evaporated after MS burned them all by competing directly with their ARM Surface.
Whatever incentive software makers had for writing for/porting to ARM from x86/x64 dried up when the hardware companies backed away from it.
The 'long tail' of x86/x64 Windows apps keeps them in business. Users, especially business users, don't want any potential issues with running in emulation.
ARM works for Chromebooks where it's basically acting
The main advantage of x86 is the fact it is tied to the IBM PC platform, one that was too old to properly implement all sorts of horrific hardware lockdowns we have now. It is a happy accident that most computers on the planet are relatively open because people use x86 computers, but as soon everything moves to ARM, you can say goodbye to all that, and have to crack your computer like a console to be able to even install a clean version of Windows on it. Once x86 is over, buying a computer without McAfee and peggle will also be.
If Apple wishes to do so, they can flood the field with their M-series SoCs. These are by far the best ARM implementations on the planet, and may require only minor tweaks to run Windows natively. The tweaks required are page size and memory coherence. Page size could be a simple microcode change and the M-series parts already handle both memory coherence regimes in L3.
...but what I want to know is: are there any good ARM laptops on the market yet, that will run some good versions of Linux ? And not an overpriced Mac with its strange keyboard...
I'm not convinced that any version of Windows is going to "take off". In fact, with the imminent end of Windows 10 support - I think more people will be seeking viable alternatives. Perhaps ones that advertise less and do what an OS is supposed to do, with is get out of the way and efficiently run applications.
Windows 11 just doesn't have the adoption it needs at the moment, and I think a lot of people are considering their alternatives.
Windows survives because of legacy code written for the x86 architecture. If it comes to a moment when companies are forced to upgrade, and they evaluate their options, they definitely won't choose Windows.
As the main developer of a relatively large 3D simulator project in C++ (in development for over 20 years), I've encountered some users who have Windows ARM systems that they wanted it to run on. Since I already had Mac universal binaries for it, porting to ARM/Windows and providing builds for it wasn't hard at all, I basically did it in case the industry moves in that direction. Builds are done on a VM on a Mac system. I don't like Windows anymore (I think it's a huge mess, it's very bloated and slow co
Sounds like x86's entrenchment at the hands of intel and AMD is holding computing back.
It is, back in the 90's there were a lot of competing architectures, some very good, but that all basically died out at the hands of Intel and market consolidation around x86; I've seen lots of good technology just die off. If MIPS had taken off, Windows might've been primarily for that platform due to the fact that Microsoft developed a lot of NT on the Jazz architecture (custom-built MIPS systems with ARC firmware), some of that can still be seen in SGI systems (which use ARCS firmware, a modified ARC, a
MS will try to turn the table on them and all of them will ignore MS because they are big enough to simply ignore MS. Smaller brands will be the ones taking the risk by ordering the cpu, putting windows there and them get cut off by MS when they release their own surface arm computer, possible even with a no name arm cpu supplier
MS is known to trash this kind of projects by being greedy and trying to grab everything for them... only to later crash and burn by the lack of third part support
I'm curious... who is telling Slashdot that ARM on Windows is going to take off? Microsoft? Qualcomm?
Gamers aren't interested in these ARM chips because of compatibility issues, and I'd imagine that stodgy old-school IT departments are worried about compatibility issues with their legacy applications. These are the same people who are now just warming up to using AMD Ryzen processors for the first time after only using Intel chips for the past 30 years.
Apple has a very high share of the ARM laptop market, considerably higher than everyone else put together. While not everyone who owns a MacBook runs Windows on it, some certainly do, and I'm sure that proportion vastly out-numbers the total number of people who own ARM laptops from other suppliers.
Windows on arm? Why?
Either you are locked in on proprietary x86 software on Windows or you are not. If you are not there is zero reason to use windows at all.
I see people here claim that you only need a browser - fine. Then you don't need Windows either - choose something else.
If you need a specific x86/64 software on Windows - run it on x86/64 with the Windows compiled for that architecture
Just like how this was the "Year of Linux"? (Score:3)
I like seeing new developments and all, but I continue to see Windows on Arm as dead. I'm sure, like Windows mobile, it'll stick around for a while, maybe get installed on a few million kiosks and other sorts of computers where being ultra-restricted and most stuff won't run is a feature, not a bug... But as a functional replacement for general work environments (both desktop and laptop)? No thanks. That is a very hard pass from me. Let me know when they exceed mid range x86-64 CPUs in terms of *both* singlethreaded and mulithreaded performance and have sufficient software compatibility that most random programs "just work".
Re: (Score:2)
I take you don't think Arm's power efficiency is enough of an advantage even in the laptop market.
Re: (Score:3)
When Arm can't run all the software and will poorly run some software and x64 power efficiency is continuously improving, not.
Re: (Score:2)
You should read up CISC vs RISC instruction sets, CISC will ALWAYS be require more clock cycles than RISC. This is why an ARM64 core will ALWYAS be more efficient than an IA64 core. Any improvement in silicon manufacturing will yield GREATER improvements on ARM64 than IA64, it's a battle that IA64, by design, cannot win. Sure, IA64s have higher top clock speeds (eg. Core i9's 9Ghz vs Apple M4s 4Ghz), but that's because you need to perform MORE operations per watt on IA64, even then, the i9 can only run at t
Re: (Score:1)
Re: Just like how this was the "Year of Linux"? (Score:2)
The problem with that analysis is that the term IA64 specifically means the architecture used by the Intel Itanium chips, and not the x86 architecture with AMDs 64-bit extensions. These instruction sets are completely different and unrelated. You could argue that Intel and AMD should just drop the 32-bit, 16-bit, and 8-bit modes/instructions at some point, and you'd probably be right, but that still won't change the definition of IA64.
Re: (Score:2)
> If you find a good way (or Microsoft rebuilt the parts) to remove the 8088 liability
That's the kicker, they have A LOT of legacy software they (implicitly) support. However, they could invest and build a solid emulator to placate those with legacy software. Box64 already lets you run x86 Linux software on arm64 Linux, there's no reason MS couldn't do the same. Apple used a 2 pronged approach with Rosetta and a "fat" binaries that packaged multiple CPU targets. In theory, MS could adopt their existing M
Re: Just like how this was the "Year of Linux"? (Score:3)
And all the technical reasons why Linux is a better desktop OS than Windows or Mac are valid too. Doesn't change the fact that technical superiority alone is insufficient to drive adoption of a technology.
Re: (Score:1)
Except that the Apple M series chips are only good at Geekbench and Cinebench. Any real world application and it just isn't as fast as AMD's and Intel's chips. Especially gaming because IPC in gaming is not the same as synthetic benchmarks. Even the M4 which is scoring much higher in Geekbench is doing so because of a new extension called SME, which increases matrix multiplications which basically means AI. But yet somehow Geekbench gave Apple another 300 points for it. As for the differences between C
Re: (Score:2)
Is Itanic, I mean Itanium, back?
Re: Just like how this was the "Year of Linux"? (Score:2)
Re: Just like how this was the "Year of Linux"? (Score:2)
Not all software may be able to run on ARM, but as Microsoft improves their emulation layer and more businesses buy into the architecture, we should see a change. Apple showed with Rosetta that you can do high performance emulation of an architecture, when running applications designed for your own OS.
If you are thinking of games, then they may end up being the laggards in getting the basics support.
Re: Just like how this was the "Year of Linux"? (Score:1)
Iâ(TM)ll second that from experience. I have a M1 MacBook that runs a Windows ARM VM. Iâ(TM)m as big of Apple fan as youâ(TM)ll find, but I will say (for my use cases) Windows 11 ARMâ(TM)s emulation is very good. Almost Rosetta good.
Now, my use cases are small and niche. Windows ecosystem is vast and complex and I doubt that everyone will be as happy as I am. But both Apple and Microsoft are showing it can be done well. Appleâ(TM)s Game Porting Toolkit also shows that there are path
Re: Just like how this was the "Year of Linux"? (Score:2)
In theory Arm is more efficient.
In practice Zen 5 is currently beating the snot put of Apple in perf per watt right now.
So...
Re: Just like how this was the "Year of Linux"? (Score:2)
nVIDIA (Score:5, Insightful)
By virtue of offering the best iGPUs for laptop-class arm chips.
Re: nVIDIA (Score:2)
Orin is 65W and Grace is 140W.
Not enough market share (Score:2)
Arm doesn't have enough of the server and desktop market for Microsoft to risk restricting themselves to one vendor. Wrong question. And they already lost the handheld market - wrong product.
Who, indeed? (Score:3)
Who is going to be the ARM CPU supplier of choice for Windows?
Who is going to be the x86 CPU supplier of choice for Windows? There can be only one. Oh wait...
Re:Who, indeed? (Score:4)
This is just part of why the premise is silly. The idea of ARM CPUs "taking off" with an OS tightly associated with closed-source x86 legacy code is highly questionable, not only would it incur performance overhead from emulation but it could add another layer of risk of breaking compatibility with legacy apps in the x86-on-ARM emulation layer.
If I had to make a bet about the future of Windows on ARM, I'd bet that it will be a niche product that eventually dies a quiet death like every other attempt to run Windows on an alternate architecture. Closed-source and multi-arch don't mix. Windows is wedded to x86.
Re: (Score:1)
Re: (Score:3)
Therefore, I don't think the compatibility issue it's a problem for legacy. Nowadays, either you run it on legacy hardware, and when it breaks, you're done, or you can run it on a VM (and if that's the case, you don't care if the CPU is an x86, an a64, or a potato).
Oh it does matter. When running a VM of the same architecture that the hypervisor is running on (as in the typical x86 VM on an x86 host scenario) the CPU's virtualization extensions are used so that the VM's instructions are basically passed-through to the host CPU as-is and there's almost no performance hit. If you're running a VM of a different architecture, then you're back to square one and have to fully emulate the CPU for the VM which carries a massive performance hit.
I remember trying Ubuntu for arm64 some time ago, and despite being open source, the compatibility with applications was abysmal unless you started downloading source code and compiling every little widget and app you wanted to get except for the main ones.
Yes, this is because Ubuntu on A
Re: (Score:1)
Microsoft released this partial emulation mode where they can use the arm64 native libraries whenever possible, and just emulate the non-standard parts of the code. This reduces the impact of emulation, but in addition, it is a n
Re: (Score:2)
you mean like apple with Mac OS ? apple being the pinnacle of closed source, and last I checked went from x86 to ARM.
yeah a really silly premise.
It won't be Qualcomm if ARM wins its lawsuit (Score:5, Insightful)
ARM is currently suing Qualcomm and has withdrawn their license to use ARM designs. Depending on how this shakes out, the chip industry may be totally re-aligned.
Ampere (Score:1)
Re: (Score:2)
No it isn't (Score:3)
This really needed a "No, ARM is not going to take off for Windows" option... Microsoft has almost no real motivation to make sure this happens, even if platform makers want them to. They just don't. It's not going to take off.
Re: (Score:2)
Isnt that what the Cowboy Neal option is for?
Re: (Score:2)
It can be interpreted as "some other company will reap the benefit" - but I don't think that's going to happen, not for Windows.
Some other platform might rise up, Chrome, or Android, or something new - heck, maybe it's the year of desktop Linux - on ARM - who knows! But it won't be Windows.
Re: (Score:2)
Whatever incentive hardware makers had for producing ARM-based laptops/desktops evaporated after MS burned them all by competing directly with their ARM Surface.
Whatever incentive software makers had for writing for/porting to ARM from x86/x64 dried up when the hardware companies backed away from it.
The 'long tail' of x86/x64 Windows apps keeps them in business. Users, especially business users, don't want any potential issues with running in emulation.
ARM works for Chromebooks where it's basically acting
That's a monkey paw if i seem one (Score:4, Insightful)
The main advantage of x86 is the fact it is tied to the IBM PC platform, one that was too old to properly implement all sorts of horrific hardware lockdowns we have now.
It is a happy accident that most computers on the planet are relatively open because people use x86 computers, but as soon everything moves to ARM, you can say goodbye to all that, and have to crack your computer like a console to be able to even install a clean version of Windows on it.
Once x86 is over, buying a computer without McAfee and peggle will also be.
Apple (Score:3)
Re: (Score:2)
Mod points to you if I had 'em.
I don't care *AT ALL* about Windows (Score:2, Offtopic)
Re: (Score:1)
Re: I don't care *AT ALL* about Windows (Score:1)
Re: (Score:2)
ARM Chromebooks seem to work fine:
https://github.com/hexdump0815... [github.com]
Re: (Score:2)
Re: (Score:2)
What are you talking about?
Re: (Score:2)
"overpriced Mac with its strange keyboard"
They changed the symbols on two keys and destroyed your mind?
Weird.
Apple (Score:2)
Pretty much the only instances of windows on arm i've seen were running virtualized on macbooks...
The premise of the question is wrong. (Score:1)
Why not? (Score:1)
MS will buy Intel (Score:2)
Windows survives because of legacy code written for the x86 architecture. If it comes to a moment when companies are forced to upgrade, and they evaluate their options, they definitely won't choose Windows.
Consumers... forget it, it's a shitshow already.
Developer's perspective (Score:2)
Re: (Score:1)
sparc and my sig were the best
Re: (Score:2)
Re: (Score:2)
Sounds like x86's entrenchment at the hands of intel and AMD is holding computing back.
It is, back in the 90's there were a lot of competing architectures, some very good, but that all basically died out at the hands of Intel and market consolidation around x86; I've seen lots of good technology just die off. If MIPS had taken off, Windows might've been primarily for that platform due to the fact that Microsoft developed a lot of NT on the Jazz architecture (custom-built MIPS systems with ARC firmware), some of that can still be seen in SGI systems (which use ARCS firmware, a modified ARC, a
Arm doesn't have enough power. (Score:2)
Re: (Score:2)
I am sure RISC will surprise us again one day
None (Score:2)
MS will try to turn the table on them and all of them will ignore MS because they are big enough to simply ignore MS. Smaller brands will be the ones taking the risk by ordering the cpu, putting windows there and them get cut off by MS when they release their own surface arm computer, possible even with a no name arm cpu supplier
MS is known to trash this kind of projects by being greedy and trying to grab everything for them... only to later crash and burn by the lack of third part support
Re: (Score:2)
Microsoft has been shipping ARM-based surfaces for quite a while now. The current offering is probably the first one though that is worth considering.
Apple (Score:1)
Who said that this was going to take off? (Score:2)
I'm curious... who is telling Slashdot that ARM on Windows is going to take off? Microsoft? Qualcomm?
Gamers aren't interested in these ARM chips because of compatibility issues, and I'd imagine that stodgy old-school IT departments are worried about compatibility issues with their legacy applications. These are the same people who are now just warming up to using AMD Ryzen processors for the first time after only using Intel chips for the past 30 years.
Mediatek you have failed me too many times (Score:2)
If Mediatek becomes the dominant ARM supplier we will all end up running Windows in perpetual Safe Mode
Most popular option missing (Score:2)
Which is of course Apple.
Apple has a very high share of the ARM laptop market, considerably higher than everyone else put together.
While not everyone who owns a MacBook runs Windows on it, some certainly do, and I'm sure that proportion vastly out-numbers the total number of people who own ARM laptops from other suppliers.
Window on arm? (Score:2)
Anyone but Qualcomm (Score:1)
Given Qualcommâ(TM)s lawsuit with Arm, I doubt itâ(TM)ll be them. Bummer, on the other hand great opportunity for other Arm SoC makers to step up.
How should I know? (Score:2)