macOS 27 without x86 support - how will we benefit?

[DrWojtek]

Since they won’t have to take x86 into account, effectively dropping a potentially huge baggage and being able to focus on code for HW that is 6 years old at maximum, manufactured and designed by Apple themselves, this is a pretty sweet seat.
What improvements could we get from this?

 

[Nermal]

The Intel and Arm code is kept separate, so when you're on an Arm machine it's not even looking at the Intel code (and vice versa). The vast majority of the OS and its apps are written in high-level languages that in most cases will not contain anything architecture-specific, so stripping out the Intel code will have next to no impact apart from freeing up a little storage space.

The parts that are architecture-specific will have already been optimised for Arm given how long the machines have been around.

 

[DrWojtek]

The Intel and Arm code is kept separate, so when you're on an Arm machine it's not even looking at the Intel code (and vice versa). The vast majority of the OS and its apps are written in high-level languages that in most cases will not contain anything architecture-specific, so stripping out the Intel code will have next to no impact apart from freeing up a little storage space.

The parts that are architecture-specific will have already been optimised for Arm given how long the machines have been around.

Yes, I was more thinking, there might be more features that gets implemented that was too costly on the Intel versions, or flows that can be optimized now that ARM is the only underlying architecture.

 

[maflynn]

What improvements could we get from this?

I don't think we'll see direct improvements once the x86 code is removed, but rather a more streamlines code base and/or developers focused on a single code base without needing to worry about compatibility issues, or quirks with x86 code base

 

[Basic75]

Yes, I was more thinking, there might be more features that gets implemented that was too costly on the Intel versions, or flows that can be optimized now that ARM is the only underlying architecture.

They already implemented ARM-only features in macOS even while still supporting Intel processors.

 

[DrWojtek]

It just seems bad marketing that the first non-Intel OS is a boring one. This is the time you release something cool to encourage upgrade.

 

[Basic75]

It just seems bad marketing that the first non-Intel OS is a boring one. This is the time you release something cool to encourage upgrade.

If the M1 outperforming the fastest i9 MacBook Pro while using a lot less energy did not encourage you to upgrade you probably don't need much from your computer anyway. It also makes perfect sense for the first version of macOS to drop Intel support to be "boring", they can remove anything x86 related and take the opportunity to do other cleanup and bug fix tasks.

 

[DrWojtek]

If the M1 outperforming the fastest i9 MacBook Pro while using a lot less energy did not encourage you to upgrade you probably don't need much from your computer anyway. It also makes perfect sense for the first version of macOS to drop Intel support to be "boring", they can remove anything x86 related and take the opportunity to do other cleanup and bug fix tasks.

Im on an M4 Air

 

[Basic75]

Im on an M4 Air

So what's the problem? Which amazing feature are you missing in macOS that can't be done without Apple Silicon?

 

[maxoakland]

Yes, I was more thinking, there might be more features that gets implemented that was too costly on the Intel versions, or flows that can be optimized now that ARM is the only underlying architecture.

That would be nice but I don't see Apple putting any special effort into the Mac. You can tell they really don't care anymore, at least not on the software side. A bigger change will be when Craig Federighi retires or gets fired. He's in charge of Apple's software and has shown he's content with "minimum viable product" thinking. Meaning the least amount of work that people will accept is what they'll put in

 

[cjsuk]

I worry about this one. Apple pretty much run the LLVM compiler project. If they lose interest in x86-64 this will be a big loss for everyone.

 

[bogdanw]

There will be small gains in security, especially once Rosetta will be eliminated. But that’s not happening in macOS 27.

Code injection vulnerability in Rosetta 2
“Dirty Bin Cache: A New Code Injection Poisoning Binary Translation Cache”

 

[keksikuningas]

I worry about this one. Apple pretty much run the LLVM compiler project. If they lose interest in x86-64 this will be a big loss for everyone.

i never considered this. they still seem to use x86_64 stuff inside Apple (using Windows on 2019 Mac Pros for some stuff or having x86_64 servers in use) so I think some initiative is still there?