New Windows x86 translation layer - implications for running Windows software on Apple Silicon?

I assume it'll just be there in Windows for ARM like the current version is, but should be more performant. If that's the case then it'd just run in the virtualized environment automatically. After all, Rosetta works fine in virtualized macOS on Apple Silicon.

Being able to run Windows in a container and fire up some old x86 games that have (compared to the horsepower of modern hardware) ~no minimum requirements will be nice!

I wonder whether Prism's performance is tied to aspects of Qualcomm's ARM implementation, just as Rosetta is to Apple's.

Bonusround said:

I wonder whether Prism's performance is tied to aspects of Qualcomm's ARM implementation, just as Rosetta is to Apple's.

Click to expand...

Not to pretend I understand all the ways chips and software can work together, but wouldn't that be something Microsoft would want to avoid? There's already at least one competitor to Qualcomm's PC SoC announced, and I'm sure there will be more. A translation layer that only worked on Qualcomm's chips, or worked significantly better only on Qualcomm's chips, would be a massive fail for growth of the platform. As always, diversity of hardware is Windows' greatest strength and its biggest challenge.

Rosetta is a convenience on Macs, but the monoculture of Apple development means it's really just a failsafe for the random bit of software that doesn't get updated in a timely fashion. Rosetta could disappear tomorrow and the pain to the platform would be real, but manageable. Comparatively, I'd say Microsoft will be relying on its translation layer to make ARM PCs minimally viable to a much greater extent and for a much longer period of time.

cateye said:

Not to pretend I understand all the ways chips and software can work together, but wouldn't that be something Microsoft would want to avoid? There's already at least one competitor to Qualcomm's PC SoC announced, and I'm sure there will be more. A translation layer that only worked on Qualcomm's chips, or worked significantly better only on Qualcomm's chips, would be a pretty massive fail for growth of the platform.

Rosetta is a convenience on Macs, but the monoculture of Apple development means it's really just a failsafe for the random bit of software that doesn't get updated in a timely fashion. Rosetta could disappear tomorrow and the pain to the platform would be real, but manageable. Comparatively, I'd say Microsoft will be relying on its translation layer to make ARM PCs minimally viable to a much greater extent and for a much longer period of time.

Click to expand...

Agree with all of that. Windows on ARM will be translating x86 app code for as long as API compatibility exists. (which for Windows is nearly everlasting)

Microsoft's translation layer certainly needs to work on all ARM PCs. But wouldn't it be a tragedy of the commons if they didn't at least optimize for more capable ARM PCs?

"Yes, but..."

I guess it depends if Microsoft believes its translation layer has a baseline of performance applicable to any ARM PC design (including the holy grail of ARM Windows running on Apple Silicon?) that is adequate to make it viable. Then yes, gaining "more" performance by buying a more-better ARM PC will work, not unlike buying a computer with integrated graphics versus a discrete GPU.

We are so spoiled by Rosetta. I remember during the PPC to Intel transition, Rosetta worked but was not without significant jank. Lots of stuff either didn't work or was noticeably slower. The utter seamlessness of running Intel software on AppleSilicon is a marvel of engineering.