Don’t Apple’s chips have specific hardware support to make Rosetta 2 as effective as it is? I’ve been really hoping other manufacturers find a way to do something similar.
I don’t know about hardware support, but I found this article on box86.org which seems to be the best alternative to rosetta on Linux. The performance drop on box64 vs native is still much greater than the performance drop in rosetta:
I found nothing, that implies that there would be specific hardware features in m1 for making the translation faster. Only that it does translation mainly ahead-of-time (AOT) and saves “that version” of the app somewhere as cache). I only scrolled through it and did not read it all, so maybe I missed it.
Don’t Apple’s chips have specific hardware support to make Rosetta 2 as effective as it is? I’ve been really hoping other manufacturers find a way to do something similar.
Yes they do, and you can bet that’ll go away as soon as Apple thinks x86 isn’t important (to them).
I don’t know about hardware support, but I found this article on box86.org which seems to be the best alternative to rosetta on Linux. The performance drop on box64 vs native is still much greater than the performance drop in rosetta:
https://box86.org/2022/03/box86-box64-vs-qemu-vs-fex-vs-rosetta2/
Edit: many infos about Rosetta under the hood: https://github.com/FFRI/ProjectChampollion
I found nothing, that implies that there would be specific hardware features in m1 for making the translation faster. Only that it does translation mainly ahead-of-time (AOT) and saves “that version” of the app somewhere as cache). I only scrolled through it and did not read it all, so maybe I missed it.