At the WWDC 2020, the Mecca for Apple developers, Tim Cook concluded the packed keynote address with a bold announcement — the transition to Apple-designed chips for the Mac. Although this has been expected for some time now, it still overshadowed all the major software updates — iOS 14, macOS Big Sur, and watchOS 7. But, why?
The Intel Affair
In 2005, in a similar announcement that shook the tech world, Steve Jobs revealed that Macs, after a decade of running on PowerPC processors, will transition to Intel processors. He credited this decision to Intel’s superior product roadmap. And for the better part of the last 15 years, Intel did not disappoint.
But in recent years, owing to a combination of Intel slacking and Apple making extraordinary progress in chip design, this partnership turned sour.
Intel, the erstwhile flag-bearer of the semiconductor industry, has been falling behind its competitors. While the company initially announced that it will ship processors made using the advanced 10nm process in 2016, it was only able to meet this goal in 2019. Meanwhile, its competitors like AMD had already started making processors using the more advanced 7nm process. AMD has also been able to price-cut Intel, as well as offer better performance, in recent years.
Intel’s focus on desktop chips as opposed to smartphone chips has further added to its struggles. The company is lagging behind in making energy-efficient chips, which are in high demand now.
Adding to all this, Intel’s production shortages became a growing concern for Apple.
Apple, on the other hand, has been making remarkable leaps in chip design. Since 2010, when Apple first started using its homegrown A-series chips for the iPhone and iPad, the company has seen a 100-fold increase in performance. This has given the company the experience and confidence to design its own chip, dubbed “Apple Silicon,” for the Mac.
What is the “Apple Silicon”?
So far what we’ve been referring to as the Intel processor is technically the Central Processing Unit (CPU), and in some cases, this comes with an in-built graphics processor. Other essential components like memory modules, controllers, and ancillary processors are separate.
But the Apple processor is something known as a System on a Chip (SoC). Along with the CPU, the Apple Silicon found in iPhones and iPads contains a graphics processor, memory module, audio processor, camera processor, ML engine, among many other modules.
By putting everything together on a single tightly knit silicon, rather than multiple dies, an SoC offers cost, size, and energy-efficiency benefits.
But the drawback of this approach is a lack of flexibility in changing individual components. When it comes to a smartphone, this is not an issue, but desktop users, especially power users, prefer having the option of changing only the graphics processor or RAM to better suit their needs. For example, an avid gamer will want a graphics processor more powerful than what comes with the Apple Silicon, but this customization is tricky, if not impossible, with an SoC. We are unsure about how Apple is going to address this problem.
What is an ARM-based chip?
You’ve probably read or heard that the Apple Silicon is ARM-based. This classification is arguably the most significant aspect of this chip. But what does it mean?
ARM Holdings is a UK company that designs CPUs and also licenses out its instruction set architecture. An instruction set architecture tells the CPU how to execute code. The CPU in Apple chips implement the ARM architecture.
Qualcomm, the leading mobile processor manufacturer, which supplies chips for Samsung Galaxies and Google Pixels, also makes ARM-based chips. Intel chips, on the other hand, are based on the x86 architecture.
The technical distinction between the two is that ARM chips implement RISC (Reduced Instruction Set Computing), while x86 chips implement CISC (Complex Instruction Set Computing).
You do not have to know what this means, but as a result, ARM chips are more power-efficient and produce less heat, while x86 chips offer better performance and produce more heat.
Due to this difference mobile devices and embedded systems almost always use ARM chips, whereas desktops and servers prefer x86-based chips. Laptops also use x86 chips, but with some restrictions on performance because of the limited power availability.
Why is Apple moving away from high-performance chips?
Well, it’s not.
While Apple chips are ARM-based, they are much more than that. Apple has perfected these chips to a point where it believes it has caught up with the performance of x86-based Intel chips, while, at the same time, consuming lesser power. This gives Apple a win-win situation that is hard to ignore.
But is there evidence that ARM-based chips can rival x86-based Intel chips in performance?
Yes and no.
The Apple Silicon found in the latest iPhone and iPad are currently industry-leading chips. When Apple launched the 2018 iPad Pro with the A12X processor, it proudly claimed that the chip was faster than 92 percent of portable PCs, and rivaled the Xbox One S gaming console in graphics performance. Follow-up benchmark tests also found that some of Apple’s latest MacBooks were beaten by the iPad Pro.
This gives us sufficient evidence that the Apple Silicon is ready for at least the MacBook Air lineup and entry-level MacBook Pros. But what about higher-end MacBooks and workstations like the Mac Pro? These are the real workhorses that need the horsepower of Intel chips. Well, the jury is still out there for this segment because there is currently no ARM-based laptop or PC that offers the same performance as Intel’s high-end Core i9 and Xeon range.
Apple tried to address this in its keynote. Prerecorded demos showed a Mac running on the same chip found in the latest iPad Pro sailing though some of the most processor and graphic intensive tasks with ease. It was put to test in video editing on Final Cut Pro, photo editing a 5GB file on Photoshop, rendering effects on Unity game engine, editing animations on Autodesk’s Maya studio, and running graphics-rich games.
Given that all this was done on an iPad processor, and that Apple is most likely working on a more advanced chip for the Mac, there’s enough reason to trust that the Apple Silicon will “give the Mac industry-leading performance per watt.”
Another big boost for ARM came one day after Apple’s keynote. Japan’s new ARM-based Fugaku supercomputer trashed the IBM Summit to take the crown for the world’s fastest supercomputer. This is big news because currently only 4 of the top 500 supercomputers use ARM chips, whereas 469 of them use Intel chips.
What other benefits apart from performance and power?
While the low-power of Apple Silicon is undoubtedly a driving factor, the custom chips also give Apple a different kind of power, the power to be in control of its product end-to-end.
By tightly coupling hardware and software, Apple has been able to offer a remarkable product experience on the iPhone and iPad. Now the same can be done with the Mac. Apple will have a better idea of how its own chip will work, and this translates into more optimized software.
Apple also has more control over the product roadmap. By designing its own chips for the iPhone, it has been able to release a new version every year like clockwork. The timeline for the Mac, however, depended mostly on Intel’s processor upgrade cycle, which was sporadic. With this transition, Apple can change that.
Will old software run on the new Mac?
Yes. Although existing software must be adapted for the new ARM-based Macs to offer the best user experience, in the worst-case scenario, old software can run in emulation mode, where the software thinks it’s running on an Intel chip. But this is usually not preferred for performance reasons.
Apple has already adapted its own OS and applications for the ARM-based Macs, but third-party software developers still have to do so. Fortunately, Apple has made this process easy for them. Microsoft Office and the Adobe Creative Cloud are some big names that have already begun adapting their software.
Going by the difficult history of Windows and ARM, this part of the transition is likely to be the biggest challenge for Apple, but it’s gotten off to the right start. You can read more about how Apple seems to have solved the chicken-or-egg problem here.
There’s also another side effect of this transition. Because the Mac will have the same underlying architecture as the iPhone and iPad, most iOS apps are ready for the new Mac. With nearly 2 million apps in the App Store, this could be a game-changer.
When can you expect these new Macs?
The first ARM-based Macs are expected to ship by the end of 2020, while the complete transition of the entire Mac line-up is expected to take up to two years. In the meantime, Apple is expected to release a few more Intel-based Macs as well. What was Apple thinking? Is it immune to the Osborne effect?
Finally, the question in many of our minds: Will the new Macs be more expensive?
Unlikely. In fact, they might get cheaper because Apple is expected to save about $100 to $150 in Mac component costs by designing its own chips. But it’s unsure if Apple will pass on this benefit to the consumer. After all, R&D costs for these new chips must be factored in.
If Apple is successful in this transition, the implications will stretch far beyond just itself. It’s only a matter of time before laptop manufacturers across the board flock to ARM chips. For power users, this could mean doing their complex work on-the-go. For everyday users, you finally might not have to carry a charger every time you leave your house. And as for me, I might, at last, have an incentive to let go of my faithful 7-year old MacBook Pro.