WWDC 2020, 1 hour and 26 minutes into the Keynote speech, Apple changed the world (For Mac users).
In just 2 years’ time, new Macs will only be using ARM-based processors designed in-house by Apple. As an engineer this means a lot to me and I thought I’d share my thoughts on a few key points that many people in my industry seem to be both excited and concerned with.
Personally, this is my greatest concern so I feel we should start here.
I use macOS as my primary operating system. I have since 2010, starting with my first Mac. The classic white plastic MacBook. Initially I bought this MacBook to learn Mac OS X to provide support to Mac users at work. Within 2 weeks of using it, I switched entirely to Apple.
However, being in a Windows world (re:enterprise IT) I’ve come to rely on VMware Fusion to let me run Windows virtual machines on my Mac. These wonderful x86_64 based VMs let me use all the Windows only tools that my job required without having to remote in to another computer, or even worse, carry two computers. Additionally Chef’s Test Kitchen tool is something I use frequently for testing various projects.
The move to an ARM-based architecture will remove the ability to run x86_64 based VMs entirely. Apple’s Rosetta 2 documentation explicitly states it will not support virtualized operation systems1. Even if it did, the performance hit would be unusable and certain software could break when running on emulated architecture (Motorola two-way radio CPS comes to mind).
Quite frankly, virtualization alone is enough to make me start wishing Apple switched to AMD Ryzen2 and stayed on x86_64 instead of ditching CISC architecture altogether.
Also, Boot Camp is probably dead in the water for obvious reasons.
One thing I’m excited for is the potential boost in performance. Over the years Apple has extracted a lot of performance out of ARM in its lower power A-series chips for iOS and its variants. Even as far back as 2018 we can see iPads outperforming3 some MacBook Pros.
Apple’s A13 Bionic chip has a TDP of just 6 watts4 and yet it performs even better than those 2018 iPads. I have to admit, I’m excited to see what Apple can do with their custom ARM chips when the chip is built around the TDP of a laptop or desktop.
It’s very possible that moving to ARM is the future, and that like Thunderbolt or removing CD-ROM drives, Apple is just the one bold enough to go all in first.
Just because desktop-class ARM might be able to run circles around the amazing chips we see in today’s mobile devices, doesn’t mean it will be suitable for “real” pro users. By “real” I mean users who buy devices such as the 2019 Mac Pro. This is a machine that can be custom ordered to have a 28-core Xeon W processor in it and I just don’t see how ARM is going to compete with it on a performance level. Even if it can, there are not that many people who fit in the category of customers that need this type of computer.
One type of customer I can think of immediately that might need x86_64 in a beastly computer such as the 2019 Mac Pro, is the scientist. Scientists use a lot of legacy software5 that doesn’t get updated very often (if at all). They rely on it immensely for their work, and even a translation layer such as Rosetta 2 may not be acceptable. It may invalidate their claims as a possible source of error introduction.
Customers like scientists are why I can’t see Apple Silicon being the right choice for the entire product line. It probably would have been better to have ARM run macOS and an x86_64 coprocessor that kicks in when needed. This type of design already sort of exists in the current T2 equipped Intel-based Macs. I would buy a Mac that does this in a heartbeat!
iOS/iPadOS Apps on Mac
This is the first thing I’ve seen that is just a straight up benefit with no downside. There have been plenty of times where iOS has a great app for something, and the desktop version is just “use our inferior website design” instead of a macOS app.
The ability to just see iPadOS and iOS apps in the Mac App Store natively is game changing! Couple this with Apple’s advancements in Catalyst and I can almost look forward to Apple Silicon.
I really don’t have much more to comment here. This one is an absolute win.
Overall, I would say I have cautious pessimism.
I’m hoping for great things to come out of Apple Silicon and I’m dying to get my hands on one of the first ARM Macs for testing. Regarding consumers, I don’t see any major hurdles in the adjustment. Most of them won’t even know Apple made such a major architecture change unless branding and marketing talks about it. This will be a good thing for them.
“OMG Steve, my new Mac can run my favorite iPhone apps and it’s so fast!”
- Average Apple Customer
For pros like engineers, scientists, architects, and digital media editors, I am a bit more concerned. I believe engineers and scientists will be hit the hardest.
I’m sure a company with a 1.5 trillion-dollar market cap knows what it’s doing and can hire the best chip talent the world has to offer…so maybe once the first ARM Macs
hit the market I’ll switch to cautious optimism.
It’s weird to feel like I can’t wait for this, and at the same time feel like I want this never to happen. Either way, I will embrace the future with whatever solutions are required in order to keep doing my work efficiently.
I’m just assuming based on nothing but speculation that Intel’s lack of progress is a huge driver behind this move. ↩
I spent the first 5 years of my career as IT support for a particle physics experiment. I’ve witnessed this first hand every day. ↩