Apple Silicon is the future of Apple.
Say hello to shiny new metal.
June 22, 2020 was “a truly historic day for the Mac” when Tim Cook, CEO of Apple, made an announcement at WWDC 2020 which many would say has been a long time coming.*
“… the Mac is transitioning to our own Apple Silicon.” — Tim Cook, WWDC 2020 keynote
That’s what Apple is calling their in-house ARM-based SoCs for the Mac, a drastic shift that heralds a new era for app developers within the Apple ecosystem.
Enter Apple Silicon
Apple is no stranger to custom SoCs — they’ve been designing custom chips since the original iPad and iPhone 4. Almost every piece of hardware they’ve released since then has included some type of Apple-designed chip: A-series for iPhones and iPads, S-series for Apple Watch, and T-series for the secure enclaves on various devices, to name a few. Apple knows how to design its own Mac chips without compromising the performance we’ve come to expect from those machines.
By migrating away from a third-party towards in-house design, Apple can exert even greater control over their supply chain and internal timelines. They’ll also be able to iterate faster, allowing ample time to resolve issues as they arise.
Another major benefit to in-house design is the ability to unify software and hardware — a facet that other companies have taken heed of and replicated (e.g. Microsoft Surface and Google Pixel devices, first released 2012 and 2013, respectively). This unification allows Apple to tailor their software for a specific set of hardware, and this optimization impacts both power efficiency and performance.
Power efficiency is defined as the ratio of the output power to the input power. Since the output power is related to the total power draw of all the internal components of a phone and such information is hard to come by, for simplicity’s sake I’ll use a normalized battery life in my comparisons. The battery is normalized against the battery capacity, thus giving a figure that tells you how long a phone lasts per unit of battery capacity (i.e. minutes per milliamp-hour).
With that in mind, consider the following charts that I compiled from the PhoneBuff Battery Test results as of June 30, 2020. As per PhoneBuff, “the test itself consists of 1-hour simulated sessions in a variety of apps … with a 16-hour standby test.”
The first chart is in order of descending time lasted during the test, and shows both the duration and battery capacity of the phones. The iPhone 11 Pro Max comes out on top in the PhoneBuff Battery Test, which is standard across the board with little human involvement (it’s mostly automated). The second chart shows the normalized battery life of the phones, in descending order. A quarter of the phones in the list are iPhones, five of which have the five best power efficiencies as seen in the chart above.
The performance edge that Apple’s tight integration of hardware and software provides is readily apparent here as well. Apple’s A-series chips also dominate the mobile chip market, as evidenced in the Geekbench 5 benchmarks. As I’m writing this article (June 30, 2020), every iPhone and iPad launched since the iPhone 8 has surpassed every Android device tested in terms of single-core performance. This includes devices using the Snapdragon 865, Qualcomm’s flagship SoC for 2020. For multi-core performance, only Apple’s latest SoC (A13 Bionic) stands up to the competition.
If you include Metal, Apple’s low-level graphics API which targets the A-series chips on iOS, then the A13 Bionic benchmarks nearly double what the Snapdragon 865 did. This is a valid comparison since near-direct GPU access like Metal provides is only possible with true software/hardware synergy. If competitors tried to replicate Metal, they would not only need to rewrite a significant portion of Android’s source code, they would also need to rely on their hardware vendors to provide near-direct access to the GPU. Only Samsung, who also develop their own SoCs, comes close to achieving this kind of functionality.
Empowering Developers
Apple Silicon also brings great news for the Apple ecosystem and its developers. With the Mac running on the same architecture as their other devices, apps may ultimately gain the ability to be compiled once for distribution across all Apple platforms.
Until that time comes, Apple revealed two major boons for developers during WWDC 2020: Universal 2, which bundles binaries that target both ARM and x86; and the ability to run iOS and iPadOS apps natively in macOS Big Sur.
Universal 2 will allow developers to compile twice (once for ARM, once for x86), bundle once, and have a single package ready for distribution. This alleviates logistical issues with maintaining separate packages for different architectures, and almost definitely removes any user confusion on which package to download, since only one is available. Having a simpler pathway to distribution allows developers to focus more time on their app rather than logistics.
Allowing iOS and iPadOS apps to run natively on macOS allows mobile app developers to target an even larger market with no work at all. Mobile apps should appear in the Mac App Store from day one, and should take advantage of all that Macs have to offer, including discrete GPUs (as long as the apps take advantage of the Metal API). This should streamline the development process for developers that maintain separate iOS/iPadOS and macOS apps, and allows those without macOS apps to tap into the laptop and desktop user base.
Apple’s transition to their own silicon for Macs was inspired largely due to their software-meets-hardware philosophy. The result is a large step forward for the Mac, and provides incredible opportunities for app developers targeting the Apple ecosystem.
* Apple was one of the founding members of Arm Ltd, the company responsible for the design of the ARM architecture. In a sense, Apple has finally come full circle — welcome home.
