The basic premise of Microsoft’s Arm64EC is that a single virtual address space can contain a mixture of ARM64 code and X64 code; the ARM64 code executes natively, whereas the X64 code is transparently converted to ARM64 code by a combination of JIT and AOT compilation, and ARM64 X64 transitions can happen at any function call/return boundary.
I wish Windows on ARM would get more traction, because I want more ARM laptops to run Linux on. It seems clear by now that Linux OEMs are not at all interested in, or capable of, making and selling ARM hardware on their own, despite Linux being in an excellent position to make using ARM on a laptop or desktop almost entirely transparent without even needing to resort to translation layers or similar tools.
This is pretty cool but let’s remember Microsoft did this before with x86 and Alpha. It was called FX!32.
Problem is, last I heard, backlash during the introduction of UEFI means x86_64 UEFI is required to at least pay lip service to allowing you to install your own Secure Boot keys, but Microsoft still requires OEMs who want to ship Windows on ARM to lock Secure Boot to ON and only install Microsoft’s keys, iOS-style, so you can’t run anything but official Microsoft-signed Windows without a jailbreak.
@ssokolow
“Microsoft still requires OEMs who want to ship Windows on ARM to lock Secure Boot to ON and only install Microsoft’s keys“
That’s not true. I dual boot linux and openbsd on my Thinkpad x13s. Secureboot is a switch in the UEFI settings.
Good to hear that the “last I heard” is the key line then. Here’s hoping that, paired with progress on right to repair, we’re seeing things head in a better direction for the tradition of people learning alternative OSes on hand-me-down tech.
False
Thom Holwerda,
I agree. I want a good choice of arm laptops too. When it comes time to pull the trigger I end up with x86 instead because too much is proprietary with ARM and everything is easier and much better supported with x86.
We’ve been suffering from a lack of compatibility standards on ARM. While microsoft concerns me for trying to lock things down (ie secure boot restrictions), I don’t feel anyone else has done as much as microsoft to push standardization. Microsoft won’t accept vendors doing their own thing, they insist on a common base like UEFI.
In terms of userspace software, linux has made it trivial to support ARM. Developers can build their software for ARM mostly without caring that the underlying architecture isn’t x86. If we could only hammer down ubiquitous mainline kernel support across ARM devices, the rest would fall into place and it would become a breeze to support all our favorite distros on ARM. Alas, I don’t know who, other than microsoft, would be willing/able to bring ARM manufacturers in line behind common standards. I wish apple were more interested in taking the lead on interoperability standards that everyone could support, but for better or worse they just want to go it alone without industry standards as far as I see.
The big problem, is that there aren’t any decent ARM laptop chips other than Apples. I really want ARM chips almost as powerful as Apple’s but discrete RAM for upgrades. I don’t want to determine the RAM when I purchase a system with zero ability to upgrade with max sizes I was using five years ago. Don’t know what my next platform will be, either a less efficient x86_64 or a memory constrained arm. Such a lame choice to have to make.
Bill Shooter of Bul,
+1 on upgradable computers.
Besides the lack of ports, I see the lack of upgradability as a bummer too. I’m especially bothered by apple’s adoption of soldered / proprietary storage. Not only because consumers cannot simply upgrade or repair it anymore, but because it has a limited lifecycle after which, well too bad. They want you to buy a whole new computer instead of being able to simply swap in new standard off the shelf parts.
I know, right! I’d really like to have an ARM laptop, but between what you are saying and the difficulty of getting mainline linux to work on it (so that my specialized distro will work) have convinced me to go x86 in the past.
ARM is a dead end.
I’m far more interested in seening where RISC-V can go without being held back by license fees.
cb88 (I get a permission denied form EVERY time I login),
Understood. Although I agree royalty free is nice to have, to me that’s almost inconsequential compared to the far bigger problems with the ARM ecosystem, which is just how difficult it is for operating systems to support ARM devices from one product to another. I am most interested in RISC-V being able to fix ARM’s standardization issues. This would provide a massive improvement over ARM, especially if RISC-V could be targeted as “easily” as x86 can. It remains to be seen whether RISC-V products will achieve a high degree of interoperability or if RISC-V vendors will insist on recreating the mess we have with ARM.
I have doubts about RISC-V on the high end.l I don’t see it outside microcontrollers and the least demanding devices.
Indeed, RISC-V was never intended to be a desktop/laptop CPU. It’s not only underpowered, it’s based on a rather old ISA model, and lacks modern extensions needed for decent desktop/laptop computing. Of course, these extension could be added, but already RISC-V lacks standardization across the board (on purpose). It’s great for embedded stuff, but won’t (ever, imho) be a good deskt0p/laptop chip.