Some of us have been waiting for years for the event Apple wrapped up in under 50 minutes (and likely spent years preparing for). Macs are moving to ARM-based Apple Silicon, and the way the company sees it, the transition should be better for everyone.
What’s the M1 chip?
Apple was pretty aggressive when it introduced its new Apple Silicon Macs, which (as expected) deliver screaming performance, thanks in part to the 5-nanometer (nm) chips Apple has developed on ARM, and thanks also to the work of the company’s silicon design teams.
- The new M1-powered MacBook Pro is now 3x faster than the best Windows notebook “in its class.” (It also has Wi-Fi 6 and Thunderbolt 4.0.)
- The M1-powered Mac mini is 10x faster than the top-selling PC desktop in its class.
- And the new MacBook Air with an M1 chip is faster than 98% of PC laptops sold in the last year, Apple boasted.
That’s the Mac/PC discussion right there. Apple knows it. It ended the event with a short reprise of its famous “Get A Mac” ads starring John Hodgman as PC. The point he made? Now that Apple’s shifted to its own chips, those PCs can no longer keep up.
The transition was no surprise
We’ve all expected Apple to deliver a new Mac chip based in some way on the A-series processors it uses in iPhones and iPads. Apple wasn’t too specific around the relationships between these silicon siblings, but it did detail the following about the M1 chip it’s putting inside these new Macs:
- The chip is a 5nm design, as is the A14 processor used in iPhones, iPads; it’s quite clearly a version of the design custom built for use in Macs, rather than mobile devices. I see it as a more robust sibling to the other chips.
- It’s also the first personal computer built with a 5nm chip, Apple observed.
- M1 is a complete SoC design, with processor, graphics, security and more all packed inside one super-powered 8-core chip.
- M1 uses a unified memory architecture which combines high-bandwidth, low-latency memory into a single pool within a custom package. That means all the tech in the SoC may access the same data without copying it between multiple pools of memory, improving performance and efficiency.
- You get four high-performance and four high-efficiency cores. The first deliver pro performance for single core tasks, and can be used together for boosted multi-thread performance.
- The high-efficiency cores handle all the everyday tasks like web surfing and email. And all eight chips can work together when you need to put your Mac to work.
- Graphics? The GPU inside the M1 has its own eight cores and can run almost 25,000 simultaneous threads. That means smooth 4K video playback, complex 3D and more. You’re basically looking at 2.6 teraflops of throughput.
- It boasts a 16-core Neural engine capable of handling 11 trillion operations per second. (I can’t help but point out that that is the equivalent of 11 XServes in every Mac – think what that can do for machine learning.)
- The performance-per-watt the chip can deliver means the first Apple Silicon MacBook Air has no fan, realizing a long-held ambition of Apple co-founder, Steve Jobs for a fan-free Mac – one that isn’t compromised.
- Apple says the M1 chip can deliver up to 3.5x faster CPU, up to 6x faster GPU, up to 15x machine learning capabilities and twice the battery life of previous systems.
But wait, there’s more: advanced power management, low-power video playback, HDR imaging and video processor, high-bandwidth caches, machine learning accelerators, a high-quality image signal processor (ISP), high-efficiency audio, fourth-generation PCI express and NVMe storage.
What this means to you
The company worked to put a little context around these “speeds and feeds” claims, noting real-life benefits, such as:
- Your Mac wakes from sleep instantly.
- You get the chime again.
- You’ll build in Xcode 2.8x faster and render a complex 3D title in Final Cut Pro up to 5.9x faster on the MacBook Pro.
- Machine learning operations such as image optimization will be up to 15x faster.
- All Apple’s software runs natively on the chip, while existing apps that haven’t been transitioned use Rosetta 2 to run.
- You can run iPhone and iPad apps, including those enterprise-specific apps you use to get business done.
- On the Mac mini, you’ll be able to boost the resolution of an image in Pixelmator Pro up to 15x faster.
Perhaps the most impressive statistic in terms of sheer computational power (to me) was Apple’s claim that the first M1-powered MacBook Pro will play back full quality 8k ProRes video in DaVinci Resolve with zero dropped frames.
Now link your new pro laptop up to Apple’s Pro Display XDR and have a little giggle as you consider the performance you would expect from a notebook attempting high-end video graphics as recently as early 2018.
The first M1-based Macs
The first Apple Silicon Macs include the $999 MacBook Air, 13-in. MacBook Pro (from $1,299) and Mac mini. An Intel version of the 13-in MacBook Pro remains available.
Apple says it is engaged in a two-year (actually, “about two-year”) transition and is expected to introduce an Apple Silicon iMac in 2021 and Mac Pro later down the line.
Pricing remains the same, with the exception of the Mac mini, which is around $100 cheaper. There had been some expectation of a slight price reduction because of savings seen by using Apple rather than Intel chips.
Battery life is astonishing. One developer noted that his Mac’s battery life now exceeds the amount of time he “customarily” spends in between sleeps. In more real terms, Apple claims the longest battery life ever on both its notebook — 20 hours of video playback on the MacBook Pro and 18 hours pn the Air. The Pro also features studio quality microphones and has graphics that’s five times faster than before.
During its presentation, Apple also quietly pointed out that the M1’s storage controller means you’ll see SSD performance twice as fast as before. There’s lots more to say about these machines individually, but the fact that a Mac mini now runs an XDR display and can render a complex Final Cut Pro timeline up to six times faster means I’ll be reading Geekbench performance data with a great deal of interest in the next few months. How high up the grid will these machines get? Rest assured, I’ll be taking a deeper look at them when I get the chance.
Want to see for yourself? All three new Macs are available to order now to ship next week.
That’s Apple, but what about its developers?
Apple has gone through several transitions. It migrated from PowerPC to Intel, and from Mac OS 9 all the way to Big Sur. It even figured out how to deliver real computing experiences on iPhones. No surprise, the usual suspects found something to moan about just before the Apple Silicon Mac launch. You should ignore them.
In the real world, those “major developers” who build the software we use to get things done seem to be having a relatively good time during this particular migration from Intel Macs to M1. At least, that’s what the Apple marketing told us.
In a video, developers from Panic, mmhmm, Adobe, OmniGraffle, Shapr3D, Affinity Publisher, and GOAT all shared comments such as: “Incredibly fast,” “Transition took a day,” “Almost limitless interactivity,” and boasted of seamless workflow across Apple devices.
The biggest claim in there, of course, is the speed and ease with which transition to Apple Silicon can take place – “transition took a day.” Apple knows it needs developers, is working to woo them, and told us Adobe Photoshop will become an Apple Silicon Native App as soon as early next year. I’m relatively convinced we can expect a few more announcements on key apps – though one thing we didn’t hear one solitary word About is the future of Windows on Mac.
The feedback seems to be that for some of the world’s most important applications, the process of migrating apps to Apple Silicon Native status isn’t too tough. We’ll have to wait on real-world feedback to see if this is the case, and to what extent the transition-easing Rosetta 2 layer (which lets Intel Mac apps roam happy and free onApple Silicon) delivers on its promise.
One more thing
The one more thing in this story is that Apple is not entirely unique in its migration to ARM-based processors. Microsoft, Qualcomm and others are all heading in the same direction, coalesced around Snapdragon.
Apple is years ahead in terms of processor design, has made deep investments in the attempt, and already sits in the catbird’s seat when it comes to manufacturing 5nm (and, soon, 3nm) iterations of these processors.
CCS analyst Wayne Lam puts it thusly:
“Apple’s moves will help validate Arm-based chips for personal computing and even in the data centre, meaning the whole Arm ecosystem will benefit. This, rather than the loss of the Mac business, is the longer-term concern for Intel.”