Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Apple to Phase Out Rosetta 2 Starting With macOS 28 as Intel Era Ends
- Thread starter MacRumors
- Start date
- Sort by reaction score
I can imagine it because we do similar.
You put in a business proposal for capex that includes expected life cycle of the device, which (in the case of high end workstation computers) is 3-5 years.
Those buying Mac Pro for business productivity in 2019-2021 have already been retiring them at end of capex life cycle since 2022-2024. Ebay is flooded with them and their asset book value (i.e., value to the business for financial/tax depreciation purposes) is likely approximately $0.
Any serious company spending hundreds of thousands of dollars on high end hardware is doing it with a budget for replacement inside the taxable depreciation period of the device. i.e., when these machines were bought, most of them were (or should have been) budgeted for replacement in the above time frame of 3-5 years. Hence the flood of them on Ebay and refurbishers.
Investing in current hardware keeps your company secure, keeps your staff happy and productive, and reduces unexpected losses due to hardware failure.
Last edited:
Any company that required that needed that kind of computing power, and had that much money to spend in 2019 has most certainly made their money back on them a LONG time ago (or gone bankrupt).
Typical lifespan of computers in a corporate environment is 4-5 years anyhow, they will have made their choices on upgrades already and written off the old machines.
Yup, and this should be a clear indicator to people (it was to me) what is going to happen to intel software support.
Obviously.
1. Your hardware stills works, just don't get new features but you can keep using the current versions indefinitely.
2. What kind of professional was still using Pentium 4 in 2017?
One incredible benefit to Rosetta 2 was the ability to run a full x86 VM emulated in docker and thus run much linux x86 only software on an ARM mac at native speed. One example is FPGA tools like vivado and quartus which are never going to be ported to MacOS and unlikely even ported to ARM. Being able to run this stuff on an ARM mac was both incredible and game-changing.
The work is surely at this point basically done. Rosetta 2 was a very hard problem which Apple solved. I don't really see why keeping it going is a drain. Perhaps Apple would open source the majority of it or make it a paid app-store download to fund the maintenance.
I feel it would be sad to let such a piece of technology which works so amazingly well just die because ..
The work is surely at this point basically done. Rosetta 2 was a very hard problem which Apple solved. I don't really see why keeping it going is a drain. Perhaps Apple would open source the majority of it or make it a paid app-store download to fund the maintenance.
I feel it would be sad to let such a piece of technology which works so amazingly well just die because ..
Right, and Apple was still selling Intel machines brand-new through mid-2023. Which means that, even by your standard, they could well be in use through mid-2028.
And macOS 26 will get security upgrades through 2028. So ride that out to the end and stop. Just because the OS stops getting major new upgrades doesn't mean the machine drops into obsolete status.
By 2023 you'd would have to be clueless not to notice that Apple was pretty much dropping Intel machines at the 5 year mark over the previous year. The Mac Pro was also well past its prime. 2023 and still selling a PCI-e 3.0 slots. You are not buying something leading edge. Technology wise, the product was 'fading' when you bought it.
Pragmatically, the a substantive subset of the entry level Mac Pro market was transitioned in 2022 with the Mac Studio Ultra. That is the subset for the $40-50K Mac Pro variants , but a substantive part of the user base was leaving by the end of 2024 well before any Intel Mac Pro was purchased in 2023.
The only other Intel machine they were selling in 2023 was the "Alternative Mini" that was better at a few corner cases like virtualizing Intel macOS images in 32GB of RAM. ( When the Mini Pro model came out the Intel model faded.). Again a 'dead man walking' product since the core of the Mini line up was transitioned back in 2020. The vast bulk of that customer base was gone.
I wonder if this move could cause Apple legal issues. I bought two MacBooks in the past 12 months (M3 Pro & M4 Pro), but these two machines are going to lose a function that they were sold with in a couple of years. I understand Apple doesn't always make new features available on current/old devices, but the mentality around this has been "don't expect new features on current/old products", but does anyone see this as losing a feature that was supported when then purchased it? Reminds me of the PS3 Linux situation.
Ths is one of those "only in America" spins on it. Sue Apple because the software developer who owns the code to the app failed to do their job and compile a new version of their app. Don't sue the person most immediately responsible for the fault.... sue the party with the deepest pockets.
Apple always position Roseta 2 as transitionary aid .
2020 MacRumors story about Rosetta start up time.
Rosetta Mac App Translation Process Can Take 20 Seconds on First Launch [Updated]
Given that Apple's new M1 chip is designed based on ARM architecture, apps built for Intel's x86 architecture will need to be run through...
www.macrumors.com
of note there.
" Apple's developer documentation acknowledges this matter, noting that the Rosetta 2 translation process "takes time" and that users "might perceive that translated apps launch or run more slowly at times" as a result:
"...
That developer doc is about the "Porting you app" process.
The link in the porting document that goes to Rosetta
'... Rosetta is a translation process that allows users to run apps that contain x86_64 instructions on Apple silicon. Rosetta is meant to ease the transition to Apple silicon, giving you time to create a universal binary for your app. It is not a substitute for creating a native version of your app. ..."
https://developer.apple.com/documentation/apple-silicon/about-the-rosetta-translation-environment
The notion that Apple promised something that would do x86 forever fails "Contracts 101". Apple always presented this a scaffolding to transition to Apple Silcon apps. Made up , artificial promises don't matter.
Well, you don't have to upgrade to the new OS if you don't want to do that.
Considering that Apple phased out PPC support in like two weeks, this really isn't terrible and I suspect that the vast majority of users have moved on to Apple silicon, or are perfectly fine without the latest OS.
CIV VI on Steam is an old x86 app and hasn't been ported to ARM64. It needs Rosetta.
Anyone who bought a $50,000 Intel Mac Pro after the release of the M1 had a very specific use case in mind and a timeline, could not wait for a new Mac Pro to be announced that may, or may not have fit their software/hardware requirements, and well aware of the transition in progress.
I would guess that most Mac Pro's sold after 2020 were either replacements or additions to an existing infrastructure which required the kind of Power and Ram capacity.
The 2018 Mac Mini appears to be have been kept in the product line for the same reason, if you have a school or organization using a few thousand of them, you might not be able suddenly switch all your software to a new architecture (if the software is even available yet) they might want to keep muddling through until the dust settles.
They had announced the a two year transition period for all products to move to Apple Silicon back in 2020 - an IT person in charge of purchases that big who missed it doesn't deserve a job.
*yes they missed the target by a year
Some developers might exit. But I find it hard to believe that Adobe, Logitech, and Microsoft will abandon the platform.
That's the conclusion I came to after posting. I guess my last MacBook will be my current one. This sucks.
If the developers haven't managed to migrate their software in 7 years, it's time to look at different software.
Or Just never update past MacOS 27.
If want to run old , abandoned software that is stuck in time , then a virtual machine images tend to work quite well with that objective. There is no hardware requirement to run your forward looking software in the same OS image as the completely backward looking software. In fact, the completely backwards looking software primarily assumes older OS as a foundation.
Long term Apple could drop the low level hardware x86 semantics support from the M-series chips. (extra baggage that doesn't add much value, if 99.999% of all aggregate software run on the packages is all Arm based. Transitors that could make a high fraction of the 99.999% code run faster, but is just 'dead' in the primary mode. ) , but pretty decent chance it will stick around for the intermediate term (at least normal user upgrade cycle past M4 . M4 -> M8-M10 zone. ) .
Last edited:
Docker isn't necessarily x86 only. Arm Docker containers don't necessarily need Rosetta 2.
Rosetta 2 really has two parts (well two and 1/2) . The compiler and some libraries to interface to the host OS environment. Over the immediate term, it is likely just the software , 'wrap aound' libraries that is the issue. The bulk of macOS libraries that a broad range of Intel apps use is being dropped ( no security fixes or maintenance).
if the wrap around x86=64 code for a fully contained docket image is 99.9999% paid for by someone other than Apple , it could survive. It is a stuck in time x86-64 opcode coverage, but some may not care.
The '1/2' is the small, narrow amount of hardware support. ( not strictly necessary, but helps with performance). At some point down the road Apple may drop that, for better transistor resource allocation to something that has more widespread uplift. [ Apple has said a reduced footprint Rosetta 2 is going to be around for a while so the hardware support probably isn't going anywhere any time soon. Porting libraries from Windows probably do avoid an very large fraction of macOS library calls. As long as that subset code "happens to work" in comatose deprecated mode, the costs are likely lower. ]
If Apple can shift 100% of the costs for doing docker x86-64 onto somebody else, then it could stick around longer than the reduced footprint Intel game app would. Pretty good chance that isn't 100% , but does die off slower that the general purpose macOS app version. (e.g, probably will take several more years for Apple to dump all of their web services servers that run x86-64 Linux. But also likely substantively fading every year. )
Probably not a grounded conspiracy theory .
Steam doesn’t desperately need a x86-64 only solution.
Steam Beta Adds Native Apple Silicon Support for Mac
Valve has quietly released a Steam Client Beta that runs natively on Apple Silicon, finally ending its reliance on the Rosetta 2 translation layer. ...
www.macrumors.com
This beta probably was not slapped together in the last couple of days or weeks . Apple many months ago has likely been telling a subset of developers ( via NDA ) that the “snooze button” on getting to work porting was about to run out
The Apple games app is not a macOS only thing . It primarily as to due with Apple one big synergistic platforms strategic objectives. It being bound to the move to tag all the OS instances with the exact same version number if the more relevant “ non-coincidence “
“ … Apple today announced an all-new Games app, introducing a unified destination for gaming on Apple devices.
…”
Apple's New 'Games' App Brings Game Center, Library, and Arcade Together
Apple today announced an all-new Games app, introducing a unified destination for gaming on Apple devices. The app features a personalised home...
www.macrumors.com
Apple already was a competitor to Steam. The bigger problem was that it was split across two or three apps. This is primarily an attempt to make it simpler . Somewhat similar to merging elements of Launchpad into Spotlight . Bigger app with more modes that were covered separately. The other problem is that arcade probably hasn’t been getting as much traction as Apple had hope. So it needs a bit of a ‘reboot’ .
The other major issue is steam , not Apple. If Apple is competing with you sitting around doing nothing is a fast path to getting ‘Sherlocked’ . Tapping the snooze button for 5 years is a dumb as a rock strategy. Tortoise and hare fable … hare so ovver confident that goes to sleep in middle of a race.
Steam puts greater emphasis on being multiple vendor platform , so the competition is not 100% . However , the intersection/ overlap is high enough to be foolish to hand wave away the competitive factors.
Last edited: