A year ago this week the era of Apple silicon truly began, as the first reviews of M1 Macs arrived, followed shortly thereafter by
M1 Macs arriving in Apple Stores and in the hands of Mac users
We had hope that the future would be brighter with
Apple-designed processors, but that optimism was tempered by
Apple's recent Mac missteps. There were also a lot of questions
about a processor that had only really seen success in iPhones and
iPads. Would there be unexpected pitfalls of abandoning Intel?
Could Apple pull off its latest Mac chip transition with the same
skill that it showed during the two previous transitions?
Twelve months later, the answers are much clear: We're in the
The Mac is in a safer
Right away it was clear that the M1 would provide to the Mac
what we had all hoped it would: impressive performance and
excellent power efficiency–leading to great battery life on
laptops. Over the last year, as Apple-designed chips have spread to
most Mac models, those facts have remained intact.
In just the last month, one of the biggest questions of the
entire chip transition has been answered. The M1 chip was capable
enough to run lower-end systems that needed about as much
processing power as an iPad Pro, but could Apple's chips scale to
meet the needs of professional Mac users? With the release of the
MacBook Pro with M1 Pro and M1 Max chips, we got the answer.
It's a definitive yes.Apple's
This was not a given. It's easy to say just throw more processor
and graphics cores at the problem, but computer performance doesn't
necessarily scale so easily. Apple's advantage really came into
play with its unified memory architecture, which doesn't just
offer a lot of fast memory for processing operations, but also
offers an enormous pool of fast memory for GPU use.
A few years ago, if you had told me that Mac laptops would be
shipping with 10-core processors with 32-core GPUs, I would believe
you–but I would have been seriously impressed that Intel had
managed to make those processors and that Apple had been able to
ship them in its laptops.
Bottom line: Apple's skill in making chips for iPads and iPhones
does translate to the Mac, after all.
Apple's unified memory
architecture gives the company an advantage over Intel.
There were very few
Shifting to an entirely new processor architecture isn't easy,
but Apple has done it to the Mac three times, and in every case
it's acquitted itself well. That said, there's probably nobody left
at Apple who worked on the PowerPC transition, and even the Intel
transition is probably only a distant memory in the mind of the
most grizzled of veteran Apple engineers.
And yet the streak remains intact. In fact, I'm tempted to call
the past year painless. Compatibility has largely been a
non-issue, starting with Rosetta 2, the code-translation system
that allows Intel-based apps to run on Apple silicon without
trouble. Rosetta got a huge leg up with the speed of the M1 chips,
of course–it's a lot easier to run translated code when it's
running on a really fast processor–but Apple also did a good job in
letting translated apps tie into native code that runs at full
speed. (For example, an Intel-built game using the Apple
silicon-native Metal graphics engine may run faster on an M1 Mac
than it did on an Intel model.)
As someone who relies on a few apps that were what Steve Jobs
used to call laggards–they took a long time to run natively on
Apple silicon, or are still not there–I am happy to report that
they run just fine, to the point where it doesn't matter
that they're not native. (And yet, I am angry at those laggard
developers, because I know their software could run much faster
than it does. One of these days they'll release an update and all
of a sudden, those apps will fly. I continue to wait.)
Even better, it seems like there are very few apps that are
actually laggards. Most of my apps embraced Apple silicon very,
very quickly. That's down to the flexibility and motivation of
Apple's developer community, and to Apple for providing them with
tools to make the transition relatively painless.
doesn't have an answer
Apple's long-time frenemy Qualcomm insists that it is going to
make chips that can match up with Apple silicon–maybe by 2023. Fans of PCs running AMD and
Intel processors cling to the fact that while the new MacBook Pros
might be the fastest laptops around–especially if you unplug them
from the wall–at least there are still more powerful computers on
the desktop. (Let's see what happens when Apple releases its Apple
silicon-based Mac Pro.)
The truth is, Apple caught the tech world flat-footed. They're
all scrambling to catch up. Apple was already more than a year
ahead of Qualcomm, every single year, in terms of smartphone
processor performance. Now it's shown that it can extend that
performance to the Mac–and in the process, use all the tricks it
used to surpass Qualcomm to blow past Intel, too.
But rest assured, they all know now. Qualcomm's next-generation
processor (which might challenge the M series in a couple of years)
will be designed by a company founded by Apple silicon engineers
which was recently bought by Qualcomm. Intel talks about having to
beat Apple at its own game–or, failing that, convince Apple to use
Intel's factories to build Apple-designed chips.
The game continues. The future isn't guaranteed. But Apple has
the drop on the competition, and this past year has shown that
everything we thought Apple's chips might be able to do, they can
The system on s chip that Apple puts in the Mac Pro will be an
important indicator of the capabilities of Apple silicon.
The transition's half over
As with the transition from PowerPC to Intel, Apple's first
steps from Intel to Apple silicon were conservative. The MacBook
Air, Mac mini, and 13-inch MacBook Pro were identical on the
outside–but transformed on the inside.
Then came the next step, with new designs (the iMac and MacBook
Pro) being coupled with upgrades to M-series processors. A redesign
of the MacBook Air is likely on the horizon.
Now that the professional-level version of the M1 chip has
arrived, there really aren't that many chapters left to write in
the story of the transition to Apple silicon. All along, Apple has
said that this is a two-year transition, and it sure feels like it
will meet that self-imposed timetable.
All that is left to revise is the Mac Pro, the larger iMac, and
the high-end Mac mini. All but the Mac Pro could probably be solved
with the existing M1 Max and M1 Pro chips. And with a second
generation of M-series processors due in 2022, who wants to bet
against Apple shipping an M2 Max Mac Pro about 12 months from
Not me. If the last year has taught any of us anything, it's
that Apple silicon delivers. I don't expect the next year to change
that perception one bit.