Apple’s ARM Transition
Almost one month ago, Apple laid out its roadmap for transitioning its computers to utilize the A-series silicon found within iPads, iPhones and the Apple TV product lines. Back in 2014, when Phil Schiller claimed the A7 processor was a “desktop class” chip, the writing was on the wall and the clock began ticking toward the end of Intel x86 processors within the Mac lineup. The presentation and initial use cases during WWDC 2020 showed some incredible potential which highlights the culmination of advancements within Apple’s processor development. The supporting strategy for change is comparable to the execution of the transition from PowerPC to x86; the translation element of a new Rosetta implement will provide a bridge to legacy software until newer ARM-native packages have replaced the current offerings. The previewed support for virtualization and containerization addresses the potential absence of Boot Camp in the future. The development kits, which consist of an A12Z processor and resources that are installed in a Mac mini enclosure, allow the software development community to begin their efforts using a fairly modern performance target. The use of the same chip from the current iPad Pro line does raise some questions regarding a future inflection point where Apple may empower the iPad Pro line to completely replace the MacBook Air. The iPad Pro, when paired with the Magic Keyboard, would support additional workflows thanks to the tandem of a touchscreen and the Magic Pencil.
There are still unanswered questions which may become clear as time passes. We have no reason to believe that Apple’s SoC solutions, both current and future releases, won’t meet or exceed the needs of established customers. The transition most likely won’t move the needle pertaining to starting retail prices for entry-level configurations. The MacBook Air’s $999 USD starting point and the Mac mini’s $799 starting point will most likely remain intact when these platforms are migrated to using Apple’s silicon. It would not be unrealistic to presume that the final x86 Mac mini will be the last model that supports some level of end-user upgrade capability. Excluding the potential Mac Pro equivalent in this new paradigm, all other units would transition to soldered memory and storage to facilitate refreshed form factors for platforms that may have a fairly stale design. This approach is going to make premium configurations very expensive due to the inability to source core upgrades from third parties. Customers have already suffered a price increase for memory upgrades this year for the MacBook Pro. External connectivity and expansion may be addressed using USB4’s compatibility with Thunderbolt 3 but the number of available ports on finished products has the potential to differ in relation to the current product stack.
It will be interesting to see how the A14 and future SoCs scale out from a compute and rendering perspective. Additionally, it would be nice to see the Windows ecosystem make an equivalent effort after the failed starts of Windows RT and the current Surface Pro X. nVidia’s continued development of solutions for embedded markets pair traditional ARM cores with the company’s graphical prowess in a power-efficient package. Although AMD never released their efforts pertaining to the development of a dedicated ARM-based SoC, the availability of a 2021 Samsung solution that is paired with AMD’s graphics technology may provide a solid idea as to what could have been had the Zen road map not been developed and realized. We’ll defer to the information presented by AdoredTV and Semiaccurate for the state of affairs and near-term prospects for Intel. The recent departure of rockstar Jim Keller doesn’t bode well from the perspective of fixing the company’s glitches. Although Intel’s effort to replace x86 with the now-dead Itanium architecture didn’t move the needle, the thought process behind requiring a successor to existing commodity compute applications does warrant further investment. The exascale initiative being pitched by Intel will remain encumbered in technical debt due to the continued reliance on x86 and manufacturing process inferiority. The potential to change the market and shun legacy standards may be realized in the current decade. Only time will tell. Apple’s first-mover advantage will provide clarity as to whether others will follow suit.