Native Chrome arrives fashionably late to the Windows on Arm party It was a while coming, but Google has finally made a Windows on Arm-native version of Chrome. The update appeared as a Canary release, which could be unstable and full of bugs, but its arrival is a welcome development for Windows on Arm users waiting for positive news about the operating system. There are browsers available …
to me, the writing is on the wall for Intel/AMD devices especially Laptops. When working all day is easily achievable with an M1 MacBook and that includes playing videos. If that is easy then why bother with I5 or I7 type devices.
Then (and yes, I know that it costs and arm (pun intended) and both legs) the M3 Max can be configured with 128Gb of RAM and 8TB of disk. It shows what is possible with the ARM architecture.
Apple make moving from X86 to Arm easy with their Rosetta tool. If another giant OS maker would think about doing the same then it would no longer be the OS that no one cares about.
but MS does seem to be afraid of taking the next step.
I'm done with X86.
The issue inevitably is the jump from one platform to the other. Apple has managed to get away with it because people accepted buying 2-or-3 updated applications that they would have bought the new version increment anyway (and Rosetta was a good enough stop gap, at least for some of the jumps). Rather more awkward in X86 land where folks routinely run 20 year+ old software that most definitely will not be getting a version bump.
I'm sure I have seen X86 successfully emulated on ARM, albeit with performance issues.
X86 is at an evolutionary dead end though at this point, going ever more parallel; giant slab sized silicon and ramping up power consumption that most users never touch isn't going to solve that.
It's had a good run. Time for a change?
"to me, the writing is on the wall for Intel/AMD devices especially Laptops. When working all day is easily achievable with an M1 MacBook and that includes playing videos. If that is easy then why bother with I5 or I7 type devices."
It should be noted that x86 laptops tend to have longer runtimes and a generally better user performance when not running Windows.
Considering this, I wouldn't hold my breath that Windows on ARM is somehow more efficient than other OSes. The runtime might be better because of the ARM advantage, but the same hardware will most likely have better runtime and performance under a different OS there as well.
Proprietary software is like that. Porting it seems very hard, for some reason. Here we have Linux, which runs on about two dozen different major processor architectures, and a distro like Debian which supports something close to a dozen of those.
And on the other hand we have Windows NT, where every single non-x86 port has failed. And in spite of Microsoft’s repeated attempts to move it to ARM, that isn’t going well either.
Proprietary software is like that. Porting it seems very hard,
Whereas proting open-source software is so easy the MS has dumped their BSD-based System-For-Unix, and replaced it with a linux virtual machine, providing linux binary compatiblilty.
Yes, it's easy to recompile trivial applications for similar platforms.. POSIX and Unix(TM) were the result of trying to unify unix, so that it would be easier to port open-source software between different systems, but git-hub is full of pre-compiled binaries, and for any kind of complex software, the instructions for building from source are arcane, voluminous, and incomplete.
You have to remember that the Windows model depends on binary compatibility: the users of software packages typically don't have access to the source code and you can't swap instruction sets and finesse decades of API changes simply by means of some smart macros and recompilation. You have to emulate x86 code and every supported version of every API has to be reproduced exactly and there's a long tail of deprecated-but-supported functionality.
This is why Microsoft is really in a dilemma over ARM. The main selling point of Windows is its compatibility with the software packages customers already have.
Apple got around this several times, but in each case they were moving to significantly faster processor families (which compensated in part but not in full for emulation overhead) and the range of software is sufficiently limited that they could reasonably expect it to be promptly ported to the new native environment.
I think Microsoft, with its much greater legacy problem, will struggle successfully to move its existing code base to another processor architecture unless it also includes some support for the x86 ISA - which would not only mean custom silicon but potential licensing headaches. They'd probably be better off creating a new future platform (perhaps based on something like .Net if they ever solve the UI problem) that's portable and distinctly different and allows new code to target both the old and the new platforms with minimal changes. It would mean maintaining two product lines, but they have most of the code they would need to do it already.
I like what Windows on Arm represents - and once their silly exclusivity agreement ends, I hold high hopes for it.
Meantime, emulation on Windows is really pretty good. I use it on a VM on an M1 Pro so I have the benefit of that processing power which I'm sure helps, but still, I've been very impressed.
It’s the same old chicken-and-egg situation: customers won’t buy Windows-on-ARM machines because there is precious little software for it, and the proprietary software vendors are reluctant to port their wares to Windows-on-ARM because there are precious few customers asking for it.
As for Windows-on-RISC-V? Forget it.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
