Newsletter Issue #997
June 24th, 2020SO IS YOUR INTEL MAC GOING TO SELF-DESTRUCT?
In announcing a two-year transition from Intel Inside to Apple Silicon, has Apple killed potential sales of new Macs?
Is that even possible? Or are the tech pundits getting just a little bit beside themselves with logic and reason?
Now the news about Apple’s third processor switch wasn’t surprising. It had been predicted for several years, and Intel’s delays in reducing die size and making CPUs faster and more power efficient haven’t helped. It meant that your spiffy new Mac wasn’t very much better than the one for the previous year, or the year before that.
In turn, Apple boasts that A-series CPU performance has increased 100-fold in the past decade. I’m not aware that today’s MacBook Pro’s performance is 100-fold faster than the 2010 model, the one that’s still in regular use over here.
To be realistic, the fastest Apple Silicon we have now, the A12Z, will essentially match the speed of a 2020 MacBook Pro, with faster graphics benchmarks. That’s for a chip being used on an iPad Pro, in an extremely resource constrained environment.
Now imagine how such a chip might fare without far fewer resource constraints within the 16GB environment of the developer kit. Essentially the kit is a glorified iPad Pro with extra ports outfitted in a Mac mini case.
How much faster will it perform with more thermal space in a product with a built-in cooling fan? Apple’s developer constraints mandate that benchmarks not be posted. But, of course, they will appear in print on the day those kits reach customers.
So where does that put someone who needs to want or needs to buy a new Mac right now?
After all, the first models with Apple Silicon won’t ship till the end of the year, and the transition won’t be completed, according to Apple, until the middle of 2022. So when will your preferred model make the cut?
And should you be buying a Mac with processor technology that’s on the chopping block?
Part of the answer may well depend on what you need it for. So, until Apple Silicon versions are out, your Mac apps will run in an emulated environment with Rosetta2.
Now Apple makes non-specific but promising claims for emulation performance, perhaps because Macs with Apple Silicon may run so much faster than the Intel models they replace. So the performance loss from emulation won’t seem so bad, or won’t even be apparent.
This was made clear when Apple ran a high-powered 3D app, Maya, under Rosetta2 during the WWDC keynote. That Apple Silicon has helper hardware for audio and video processing won’t hurt.
But what about virtual machines?
In its developer notes, Apple says Rosetta2 won’t work with kernel extensions or with visualization apps. There will be no Boot Camp either.
Now some tech pundits have misread the limitations to assume that such virtualization apps as Parallels Desktop and VMWare Fusion must be history, but that’s not true. Apple also demonstrated Linux virtualization using an ARM-based version of Parallels Desktop.
What this indicates as clear as day is that there will compatible versions of one or both apps. How well they’ll work with Windows obviously won’t be known until someone is able to demonstrate such a configuration. The promise of high-performance emulation, which will get even faster as Apple Silicon is enhanced over the years, should convey its own message.
Many of you recall running Windows in emulation on a Power Mac, and waiting long minutes to open apps and perform a few simple tasks. That was then, this is now. For most people, I expect that running Windows in a virtual machine may well deliver satisfactory performance for most people.
I’m just guessing here, but Apple is making the point about good Rosetta2 performance. They didn’t have that boast in the early days of the PowerPC or Intel Macs, though the original Rosetta did just fine for most of you. I remember it being pretty decent.
Remember that the promise of Apple Silicon is not just better battery life, but much better performance. The implication is that the loss from emulation will be more than compensated for by Apple Silicon enhancements.
Apple also promises that it will be relatively easy for developers to recompile their apps as Universal binaries for Apple Silicon and Intel. Apple made that promise with the PowerPC-to-Intel transition, but there were notable exceptions.
It helps that Apple has been working with Adobe and Microsoft to update their apps as quickly as possible. No doubt Parallels Desktop will make the transition too, based on the demonstration of Linux virtualization.
Many apps, especially those that haven’t been upgraded in a while, may never make it to Apple Silicon, so you’ll have to depend on Rosetta2, at least until Apple decides to retire it, as they did with the original Rosetta after a few years.
That leaves the inevitable decision about the sale of new Macs. Will they fall off the cliff because they are destined to become obsolete? Or will folks decide that they’d rather not put up with potential incompatibilities with existing apps and take the plunge?
I would expect that the Mac mini, MacBook Air, MacBook Pro and regular iMac will receive upgrades first, within the first year. The iMac Pro and the Mac Pro may receive Intel upgrades during that period, and won’t be replaced until 2022.
It may even be that Apple will look into new options for professional users with its own chips. Don’t forget that ARM CPUs power some of the most powerful workstations and servers on the planet, so it’s not as if Apple is going to cripple pros while they make the transition.
It’s also quite possible that the migration to Apple Silicon will occur faster than Tim Cook predicted. Don’t forget that Steve Jobs promised that the original Intel transition would be complete by the end of 2006, but the final Mac to switch, the original Mac Pro, arrived in August of that year.
So it’s very possible that Apple will be done by the end of 2021. But they are allowing themselves a quarter or two to deal with unexpected glitches, or possible customer resistance that might require additional product engineering.
Regardless, macOS will be updated for Intel Macs for at least five years or longer. What you buy today won’t suddenly become obsolete, although some apps will not run at full efficiency because they take advantage of the new features in Apple Silicon.
Let me take it personal.
I own a pair of older Macs. My MacBook Pro arrived in 2010. It works fine, but has been orphaned with macOS upgrades for several years. My Late 2014 iMac will run Big Sur and may be supported by OS or two after that. It is my primary work machine that’s used seven days a week, so it will have to be replaced eventually.
It could be on borrowed time, however. It received a Fusion Drive transplant last year when the original mechanisms failed. Apple cleaned it out and thoroughly tested the unit during two visits to an Apple Store.
But I fully expected to have to buy new hardware in the not-distant future. I will be eyeing the first iMac with Apple Silicon to see if that’s the way to go.
And buying a huge penny jar to begin to save for it.
THE FINAL WORD
The Tech Night Owl Newsletter is a weekly information service of Making The Impossible.
Publisher/Editor: Gene Steinberg
Managing Editor: Grayson Steinberg
Marketing and Public Relations: Barbara Kaplan
Copyright 1999-2020 Making The Impossible. All Rights Reserved.
Privacy Policy: Your personal information is safe with us. We will positively never give out your name and/or e-mail address to anybody else, and that’s a promise!
Print This Issue
