Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Just to clarify, this is translation, not emulation as implied in the article, right?
There's multiple layers going on here:

  • Wine itself is just a reimplementation of (much of) Win32. It's like running a clone of Windows, except you leave out its lower levels and focus mostly on running apps.
  • Rosetta 2 is a hybrid: it tries to translate an app from x86 to ARM ahead-of-time (so much so that you can even take an x86 plug-in — say, for Photoshop — and run it inside an ARM app, because by the time it's running, it's no longer x86), but that doesn't work for all code. The remainder is emulated just-in-time.
So, I do think there is an emulation component.
 
Obviously, you have no clue how applications are used by enterprises. They have thousands of custom apps developed at some point in time at great cost. It is prohibitively expensive for them to redesign these apps. They have to be able to use them for a long time. That's one reason why enterprises do not use (and may never use) Macs.
Oh I know all this but it boils down to the Sunk Cost Fallacy.

I had this problem in one of the places I worked. They were using a Unix version of this program that had migrated to windows ages ago but because they had gotten the program before DOS existed (1981) they refused to make the jump because all that money would be "wasted" even though it was 2008.

The company changed hands and the new owner did what should been done ages ago and finally upgraded. An audit that had taken 3 hours thanks to this antiquated monstrosity we called an accounting program now took about 30 minutes (if that long) and it was far easier to see and identify the source of irregularities. Funny thing is my manager who had seemed to have been trying to get me canned (I was the auditor and the first person to kind of understand the old accounting program) left about 2 months later.

I saw something similar when doing my thesis research regarding museum computation in the mid 1990s. One of the museums was using an 8088 to cataloging the collection. Everybody knew it was too old (the joke there was it should be in the collection not cataloging it :) ) the manager of the museum wouldn't upgrade because the software they used had been so expensive. Given Minsmy (sp?) was $100,000 I don't want to think what they had payed back then. My thesis got turned into a presentation that I gave at two museum conferences and was on how to use off the shelf software rather then internal home-brew or insanely expensive speciality software. What had been planned to a 90 minute presentation turned into a 2 hour one thanks to all the question about how to determine what was the best off the shelf software. I even sent a copy to a museum in Hawaii who had heard about the presentations but couldn't attend as it wasn't "in their wheelhouse" so to speak.

Past a certain point you have to wake and realize that hanging on to old software or using home-brew stuff may be hurting you in ways you don't even realize.
 
Last edited:
Looks great if you want to play games at 10fps
I mean, most people don’t get a MacBook to game on in the first place.
And then you probably wouldn’t expect, say, Doom Eternal to run at 60fps on this.

That said, I tried CrossOver tonight on an early 2014 MBA with Unreal Tournament 2004 and I got between 40 and 60fps on that old machine at 1280x800 and everything maxed out. Not an extremely demanding game but then... [see first paragraph]
 
  • Love
Reactions: Voyageur
There's multiple layers going on here:

  • Wine itself is just a reimplementation of (much of) Win32. It's like running a clone of Windows, except you leave out its lower levels and focus mostly on running apps.
  • Rosetta 2 is a hybrid: it tries to translate an app from x86 to ARM ahead-of-time (so much so that you can even take an x86 plug-in — say, for Photoshop — and run it inside an ARM app, because by the time it's running, it's no longer x86), but that doesn't work for all code. The remainder is emulated just-in-time.
So, I do think there is an emulation component.
No there isn't. "It translates the apps when you install them, so they can launch immediately and can be instantly responsive. And Rosetta 2 can also translate code on the fly when needed." WWDC 2020 Special Event Keynote — Apple 1h39m37s mark The full details are at About the Rosetta Translation Environment.

Just-in-time compilation is not emulation though some emulators do use it (as it speeds things up)

"Dynamic binary translation differs from simple emulation (eliminating the emulator's main read-decode-execute loop—a major performance bottleneck), paying for this by large overhead during translation time."
 
Last edited:
  • Like
Reactions: DocNo and Spectrum
Obviously, you have no clue how applications are used by enterprises. They have thousands of custom apps developed at some point in time at great cost. It is prohibitively expensive for them to redesign these apps. They have to be able to use them for a long time. That's one reason why enterprises do not use (and may never use) Macs.
This is true. And as William Dodd said; “I do believe that facts count. Even if we hate them.”

Don't get me wrong, I'm deeply impressed with what Apple is doing with the M1 chip. I think that not only is the chip itself more capable than anyone expected, but that they've almost certainly written some amazingly tight and optimized code. The kind of code we haven't see since great programmers squeezed every last drop of performance out of Atari and Amiga systems; when every single cycle counted and the programmers knew it. The last time I remember being truly blown away by optimization was the BeOS, when we saw Mac hardware run 5X faster, through the magic of good code. I think Apple probably did an amazing job optimizing Big Sur, the default apps, and Rosetta 2 for the M1.

But ultimately software compatibility is everything. It absolutely HAS GOT TO RUN the software you need it to run or it is useless to you. That's why we worry about the future of Windows compatibility on ARM Macs. Last night my daughter sat down, switched into Bootcamp and played Dragon Age Origins. A total impossibility on an M1 Mac.

For many it doesn't matter; people with needs that don't exceed the Apple ecosystem at all. For others it means being forced away from Apple hardware, or not considering it in the first place, no matter how much better it may be. Interesting times!

I hope to get my hands on one of the new M1 MacBook Pros through my education channels soon. As a huge Amiga and Atari (and BeOS) fan of old I've always been extremely fond of "super responsive" computers (come to think of it, I also prefer small light sports-cars)... Anyway, I have a feeling I'll be lugging two laptops around; using my MacBook Pro Intel (Macintel?) almost exclusively in Windows and an M1 for Mac (ARMac? Macapple?). Seeing how fast those things react to input; I know I won't be able to go back. 😂😂😂
 
  • Like
Reactions: Val-kyrie
This is true. And as William Dodd said; “I do believe that facts count. Even if we hate them.”

Don't get me wrong, I'm deeply impressed with what Apple is doing with the M1 chip. I think that not only is the chip itself more capable than anyone expected, but that they've almost certainly written some amazingly tight and optimized code. The kind of code we haven't see since great programmers squeezed every last drop of performance out of Atari and Amiga systems; when every single cycle counted and the programmers knew it. The last time I remember being truly blown away by optimization was the BeOS, when we saw Mac hardware run 5X faster, through the magic of good code. I think Apple probably did an amazing job optimizing Big Sur, the default apps, and Rosetta 2 for the M1.

But ultimately software compatibility is everything. It absolutely HAS GOT TO RUN the software you need it to run or it is useless to you. That's why we worry about the future of Windows compatibility on ARM Macs. Last night my daughter sat down, switched into Bootcamp and played Dragon Age Origins. A total impossibility on an M1 Mac.
Using bootcamp is but playing the game may not be. It all depends on what Parallels and VM Fusion do. Sadly that game is "Installs, Will Not Run" for CrossOver which does run on the M1 :-(

It would be funny if somebody showed ARM Windows running in Virtualization mode and went 'We can put the claim Windows doesn't run on the M1 to bed'. It would be the most annoying troll on the internet :)

For many it doesn't matter; people with needs that don't exceed the Apple ecosystem at all. For others it means being forced away from Apple hardware, or not considering it in the first place, no matter how much better it may be. Interesting times!

I hope to get my hands on one of the new M1 MacBook Pros through my education channels soon. As a huge Amiga and Atari (and BeOS) fan of old I've always been extremely fond of "super responsive" computers (come to think of it, I also prefer small light sports-cars)... Anyway, I have a feeling I'll be lugging two laptops around; using my MacBook Pro Intel (Macintel?) almost exclusively in Windows and an M1 for Mac (ARMac? Macapple?). Seeing how fast those things react to input; I know I won't be able to go back. 😂😂😂
As I said it all depends on what Parallels and VM Fusion do. From what I have read using Windows for ARM (they would have to swing a license from Microsoft to have it as a preinstall) is not a solution as everything I have read shows that uses an emulator (Rosetta 2 is a translator through and through) and that thing blows goats on an ARM and while the M1 is fast I don't think it is fast enough to fix that issue.
 
Last edited:
What's the difference between JIT translation and emulation?
The time delay caused you edit not to hit before you posted this. The relevant parts are:

Just-in-time compilation is not emulation though some emulators do use it (as it speeds things up)

"Dynamic binary translation differs from simple emulation (eliminating the emulator's main read-decode-execute loop—a major performance bottleneck), paying for this by large overhead during translation time."
 
Oh I know all this but it boils down to the Sunk Cost Fallacy.

I had this problem in one of the places I worked. They were using a Unix version of this program that had migrated to windows ages ago but because they had gotten the program before DOS existed (1981) they refused to make the jump because all that money would be "wasted" even though it was 2008.

The company changed hands and the new owner did what should been done ages ago and finally upgraded. An audit that had taken 3 hours thanks to this antiquated monstrosity we called an accounting program now took about 30 minutes (if that long) and it was far easier to see and identify the source of irregularities. Funny thing is my manager who had seemed to have been trying to get me canned (I was the auditor and the first person to kind of understand the old accounting program) left about 2 months later.

I saw something similar when doing my thesis research regarding museum computation in the mid 1990s. One of the museums was using an 8088 to cataloging the collection. Everybody knew it was too old (the joke there was it should be in the collection not cataloging it :) ) the manager of the museum wouldn't upgrade because the software they used had been so expensive. Given Minsmy (sp?) was $100,000 I don't want to think what they had payed back then. My thesis got turned into a presentation that I gave at two museum conferences and was on how to use off the shelf software rather then internal home-brew or insanely expensive speciality software. What had been planned to a 90 minute presentation turned into a 2 hour one thanks to all the question about how to determine what was the best off the shelf software. I even sent a copy to a museum in Hawaii who had heard about the presentations but couldn't attend as it wasn't "in their wheelhouse" so to speak.

Past a certain point you have to wake and realize that hanging on to old software or using home-brew stuff may be hurting you in ways you don't even realize.
I am talking about internally developed software for which there is no substitution. In many cases this software is integral part of the manufacturing or design process.
 
So...If Rosetta 2 can run x86 apps via on-the-fly translating faster than on intel x86 chips...why cant we have Parallels native app that translates Windows x86 OS on the fly too?

Everyone is posting these amazing emulation videos but no one is reporting on heat...My MBP can run a lot of stuff but if feels like a heater that I can warm myself too during the winter.
It was pretty skeptical about the transition to ARM based Macs. Once again, I find myself on the wrong side of a debate. What I’m seeing in these last few days is more encouragement to me updating to an M1 (or whatever comes next) Mac.

To be fair, no one anticipated on-the-fly emulation/translation being this good. This never happened. The Rosetta 2 apps are running on emulation faster than native on the older intel chips. This is unheard of AFAIK! Its actually faster in emulation!
 
  • Haha
Reactions: Maximara
Also super low settings. The exploding animation after he kills Pyro plus the super low poly count on the sewer walls while he's chasing scout. The sewer wall is only a couple sides away from being an octagon.

Is cool that they got this to work, but even when I first played TF2 on my circa 2005 mid tier setup - I was playing at higher settings than this.
The mac community has been connoisseurs of bottom feeder iGPU garbage for years, they have forgotten what real graphical grunt looks like. M1 while good on the CPU front, cant hold a candle to modern dGPUs. While they are getting excited M1 can beat a four year old entry level graphics card from Nvidia, and that the M1is so much more energy efficient than an Iris. They should be asking, where the hell is the dGPU? Not foaming at the mouth over an Apple iGPU (as the only option), and that we should be happy it can beat an intel Iris iGPU. DaVinci painting the sistine chapel roof is faster than an intel iGPU.
 
I am talking about internally developed software for which there is no substitution. In many cases this software is integral part of the manufacturing or design process.
Some of the museums were using internally developed software and across the board they were in trouble. They would have to manually transfer everything and if anything ever happened to that machine they were FUBARed to Mount Doom and back. They were like the monkey who reaches into a jar and can't get it back out without letting go of the food it reached in to get. They were, quite literally, in a trap of their own making.

Having internally developed software that is an integral part of the manufacturing or design process is just asking for trouble. Upgrading it can be a nightmare and tends toward spaghetti coding. The old Classic System for Mac was a prime example as it was a wild mish-mash of assumably, Pascal, and C that no one at Apple actually knew how it worked anymore.

The reason I know this is in one of our Apple club meetings we has an Apple representative who told us a little more than he should have about Copeland. Namely that there were these pieces of code that as far as what they could tell were never called but the moment they were removed the system promptly crashed.

According to him large parts of the Mac System where black box and thanks to later programmers using tools like ResEdit the original source code wasn't helpful anymore. In some cases this was because the compiler stopped working several systems ago and certain bugs had been "baked in" (this is why World Builder couldn't be recompiled to be 32-bit clean and why HyperCard was such a PITA to update)
 
Last edited:
But ultimately software compatibility is everything. It absolutely HAS GOT TO RUN the software you need it to run or it is useless to you. That's why we worry about the future of Windows compatibility on ARM Macs. Last night my daughter sat down, switched into Bootcamp and played Dragon Age Origins. A total impossibility on an M1 Mac.

A lot has changed in the past decade. We have:

- The web, where a lot of software can run in the browser, with or without a website
- The cloud, where a lot of programs and services run
- Mobile devices where a lot of programs, services and applications run without Windows
- Inexpensive computing devices, including PCs
- The growth of macOS users, software and compatibility
- Microsoft writing the programs to run natively on macOS and iOS
- A lot of desktop software is available on Windows and macOS, or even Windows, macOS and Linux
- The ability of many to set up a home cloud
- macOS Virtual Machines
- Windows Virtual Machines

I've made the effort to be able to run on either Windows, macOS, or both over the past decades and am there right now. I was running all macOS a year ago but my main system today is Windows with macOS laptops. My philosophy right now is run the application on the platform where it makes the most sense. I am not a gamer but I'd just buy or build a Windows desktop if I were if that was the best platform.
 
The mac community has been connoisseurs of bottom feeder iGPU garbage for years, they have forgotten what real graphical grunt looks like. M1 while good on the CPU front, cant hold a candle to modern dGPUs. While they are getting excited M1 can beat a four year old entry level graphics card from Nvidia, and that the M1is so much more energy efficient than an Iris. They should be asking, where the hell is the dGPU? Not foaming at the mouth over an Apple iGPU (as the only option), and that we should be happy it can beat an intel Iris iGPU. DaVinci painting the sistine chapel roof is faster than an intel iGPU.

Are you talking as a user or a shareholder?

I have no doubt that Apple could be a top-tier dGPU company. They could just buy out AMD in cash. Apple has 27 times the revenue of AMD. We look in amazement at what Apple has accomplished with CPUs. Why do you think that they couldn't do the same thing with dGPUs if they wanted to?

One thing about megacap tech companies. They can hire your top talent. They can hire whole teams of people out of your company. They can hire the top talent from your competitors. They can buy out small companies that are making the next generation of your product area. They can buy out the services that you need to make your product.

Sure, I'd like more than an M1. But Apple has made something to improve profits. Something that I like as a shareholder. What they are selling is going to greatly reduce their PC component costs for a mass-market product. Can you imagine what they are playing around with in their labs?
 
The time delay caused you edit not to hit before you posted this. The relevant parts are:

Just-in-time compilation is not emulation though some emulators do use it (as it speeds things up)

"Dynamic binary translation differs from simple emulation (eliminating the emulator's main read-decode-execute loop—a major performance bottleneck), paying for this by large overhead during translation time."
The key part here is “simple” emulation.
 
An attempt at a rough analogy:

Emulation - where you are reading a book written in another language and you have a dictionary where you look up the words as you read.

Translation: someone translates the book to your primary language and you read it in your language.

JOT (Just in time): someone who knows both languages reads the book to you in your language.
 
  • Love
Reactions: Maximara
A lot has changed in the past decade.
Yes, the world today is so different that not all 11 year old programs run even on Windows. World Builder is the longest lasting Mac program I know of (25 years; lasting from System 3.2 to MacOS 10.6.7) and even it had problems (sound stopped working for example)
I've made the effort to be able to run on either Windows, macOS, or both over the past decades and am there right now. I was running all macOS a year ago but my main system today is Windows with macOS laptops. My philosophy right now is run the application on the platform where it makes the most sense. I am not a gamer but I'd just buy or build a Windows desktop if I were if that was the best platform.
I think what is the "best" platform just got a major kick in the pants. It is starting to get where you want to write code to be as platform independent as possible so if there is a major shift you can take advantage of it and are not "locked" into one architecture.

One thing I do not understand is how programmers can write programs that run on Windows and Linux and yet claim they can't write for the Mac because "there isn't enough of a marketshare". The "logic" of supporting Linux which has far more variation than the MacOS and yet a smaller marketshare is so non sequitur that I go 'Are you kidding?' 'Are you freaking kidding?' and seriously wonder if they know what the sam hill they are doing.
 
In my experience, when WINE works, it tends to work well. When it doesn’t work, it tends to be really bad. Fortunately the program I need runs well.
That mirrors my experience. An example of running well: I use MP3Tag on my iMac via WINE. Works perfectly.

On the other end, one of the top issues with WINE (and derivatives of it) that I've experienced is how it handles native display resolution and scaling.

I've used Crossover on my Pixelbook. It was very fragile. It was running surprisingly well for a cycle... but every update to CrOS was a bit of a roll of the dice to see if it would continue to work. After a while, it never returned to the stability it once had.
 
That mirrors my experience. An example of running well: I use MP3Tag on my iMac via WINE. Works perfectly.

On the other end, one of the top issues with WINE (and derivatives of it) that I've experienced is how it handles native display resolution and scaling.

I've used Crossover on my Pixelbook. It was very fragile. It was running surprisingly well for a cycle... but every update to CrOS was a bit of a roll of the dice to see if it would continue to work. After a while, it never returned to the stability it once had.

My experience as well. I used it for Active Trader Pro (which is a big program) on macOS. I am back on Windows and it just runs better on Windows in terms of performance. Every once in a while, though, it broke. It might have been an update from Fidelity or an update from Apple. And then you need an old kit or a new kit.
 
That mirrors my experience. An example of running well: I use MP3Tag on my iMac via WINE. Works perfectly.

On the other end, one of the top issues with WINE (and derivatives of it) that I've experienced is how it handles native display resolution and scaling.

I've used Crossover on my Pixelbook. It was very fragile. It was running surprisingly well for a cycle... but every update to CrOS was a bit of a roll of the dice to see if it would continue to work. After a while, it never returned to the stability it once had.
WINE in general is very Girl with the curl - when she is good she is great; when she is bad she is horrid.
 
  • Like
Reactions: DocNo
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.