Quote:
Originally Posted by Kwixo
There is really no need for snark or name calling here. After all, this is just a minor disagreement on unreleased technology that nobody has firm answers on. Rosetta 2 won’t face the same limitations technologically as its predecessor. We’ve already seen benchmarks from the devkit running on native x86 code in the non optimized OS with a chip that won’t even be in a desktop that showed decent results. It’s fair to assume those results were colored with the Rosetta emulation as benchmark software is still x86 and not optimized for arm. It’s disingenuous at this point and time to assume that Apple will handicap its users for intel machines, not just for architectural switching reasons, but because they are still releasing intel machines in their product pipeline. It is not to apples advantage, or their users, to end support in 3-4 years as some of the pundits have suggested.
|
^^^^This.
Apple has told devs Intel OS deployment "for the foreseeable future"
. There won't be a full transition for two years, best case. Those buying high end Macs today aren't going to be end of life once the transition is complete. Will. Not. Happen. It will be at least a few years after the transition likely longer. With minimal impact they can keep building Intel versions of Mac OS for as long as they like. The limiting factor at that point will be using whatever new services and features that will be incorporated in Apple Silicon that Intel only based apps won't be able to exploit.
One reason they can do this is software development has significantly changed over the last decade and a half since PPC to Intel. Targeting different architectures from a common code base is a normal thing now particularly with *nix variants. It could be done back then but not with the relative ease it has today. Mac OS (and the watches, TVs, phones and tablets) is Darwin (BSD with a Mach kernel variant called XNU). The APIs and UI are abstracted from the core of the OS. That's how you can build for different platforms. The core OS is built for the target architecture while the "Maciness" is the same as it's always been (more or less, over simplification). Comparisons to previous transitions are a good historical perspective but no so much when it comes to advances in software engineering and deployment.
Rosetta 1 was a real time instruction translator. That's why it was so sluggish and not everything would work with it. Rosetta 2 is primarily an install based program that for all intents and purposes translates a native version of the program thus not requiring real time translation. For instances where that's not possible there is a just in time complier. That's similar to the JITs used everyday on modern UI designs on mobile devices. It's far faster and superior to Rosetta 1. My guess would be Rosetta 2 will EOL sooner than Intel OS X as they push devs to at least UB2 native apps. That push is already happening. Wouldn't surprise me if they EOL Rosetta a bit after the full transition. There is already a dev note urging audio code to port to UB2 right now.
This isn't so much Apple transitioning to ARM as it is Apple using some ARM technology. It's Apple licensing some ARM patents, most notably the core structure. The bulk of the design is from Apple. It's not a ARM reference design though you do use ARM assembler and libraries the system is entirely designed by Apple. The scaling for the ARM based RISC design isn't necessarily going to come from the ARM part. It's likely going to be from the SoC components designed by Apple.
Apple has confirmed there will be Thunderbolt on the Apple Silicon Macs.
For Pro Tools users I'd be more concerned with Avid having a fully compatible Apple Silicon version by the time the transition is completed. It's not so much a slam on Avid as it is there is likely to be some pretty significant changes in something like Pro Tools or Media Composer.
Like we did in the olden days and to an extent we do today lock that old rig down and don't use it for anything else until you're sure the transition is OK.