Every Apple event attracts a lot of attention, but the launch of the 2020 MacBook Air, MacBook Pro, and Mac Mini was about more than new features and sleek industrial design. This year, the big news is hidden deep inside the anodized aluminum chassis of new MacBooks, in the form of the Apple M1 System on a Chip (SoC), which is replacing Intel processors in 13-inch MacBooks and the Mac Mini.
Why is the transition to Apple processors so important? After all, Apple has been designing its smartphone processors since 2010, and this is not the first time it has embraced a new CPU architecture. Apple has gone through a number of different architectures, from Motorola CPUs in its Wozniak days, through PowerPC chips in the ‘90s, to Intel x86 processors in 2005. Now the company is moving to an ARM-based processor of its own design. In contrast, the Windows PC platform has stuck to x86 processors since the first IBM PC launched in 1981.
Today we will be taking a closer look at Apple’s new M1 chip, its implications for software engineers, designers, consumers, and the industry as a whole. We will not provide you with detailed performance reviews or test the compatibility of individual software suites and tools, legacy software, SDKs, and so on for brevity. Since you can easily find benchmarks and reviews on reputable hardware sites, we will focus on the big picture and address some concerns voiced by designers and developers interested in buying Apple’s new M1 MacBooks.
What Makes the Apple M1 Processor Different?
What makes the Apple M1 processor different? The Apple M1 differs from Intel processors used in previous MacBooks in many respects:
- The M1 is an ARM processor, not an x86 processor.
- It integrates more components than an Intel CPU.
- The Apple M1 also integrates RAM in the same package.
- It features Rosetta 2 dynamic binary translation that allows it to run x86 software.
- The chip boasts eight CPU cores, in addition to the integrated GPU.
- It is manufactured using the 5-nanometer process and has 16 billion transistors.
For the average user, most of these specifications won’t mean much, but power users should take note: The ARM-based Apple M1 is supposed to run legacy x86 software using Rosetta, which could cause adverse effects on performance when running x86 applications. Fortunately, due to its speed, the M1 will still outperform older Intel chips in most scenarios, even with legacy x86 apps. Also, some teething problems are to be expected, as exotic tools and applications might not run out of the box, or they may incur a performance penalty. Still, these problems will be ironed out, and most users needn’t worry about them.
For example, Docker users were unsure whether or not their x86-64 images would work properly. For the time being, they are facing serious issues, and Docker is not working properly. Although Adobe is working to optimize its products for the M1, designers relying on third-party plug-ins for Adobe products are experiencing compatibility issues.
This is to be expected on new hardware, as developers need time to catch up to ensure compatibility and port their software for the new hardware. It may take a while.
Designed by Apple, Marketed by Apple
Apple used the event to claim the M1 is the fastest CPU core on the market, but hardware enthusiasts were quick to point out that some of its performance claims were vague and difficult to confirm. We are not going to put every Apple marketing claim to the test as independent reviewers have already done that and the reviews are overwhelmingly positive. The M1 is not the fastest CPU on the planet, but it seems to be the fastest CPU for ultraportable notebooks.
Let’s take a closer look at the design of the Apple M1 and explain what makes it ingenious and how it could profoundly impact the industry.
Apple M1 Design and Features
The Apple M1 chip features four big Firestorm CPU cores for high-load scenarios, backed by four smaller Icestorm CPU cores designed for efficiency. If this sounds familiar, you’ve probably encountered Android phones with a similar ARM CPU layout. ARM calls this layout ARM big.LITTLE and it’s been around since 2014. The CPU uses the AArch64 or ARM64 extension set of the ARM architecture.
Likewise, you are probably used to integrated GPUs as well, as they’ve been used in Intel and AMD chips for years. The GPU used in the Apple M1 has eight cores and takes up just a bit more space on the chip than the eight CPU cores. Apple claims the GPU can deliver 2.6 TFLOPS. To put this in perspective, Nvidia’s GeForce GTX 1050Ti from 2016 manages 2.1 TFLOPS. That’s a desktop graphics card with 3.3 billion transistors that draws up to 75W of power, beaten by integrated graphics on a passively cooled MacBook Air.
The only downside is that there’s no eGPU support and you can’t get discrete graphics. Apple didn’t offer optional discrete GPUs on 13-inch MacBooks powered by Intel processors, either. In case you need a Radeon Pro on your new MacBook, you will still need to get a 16-inch MacBook Pro.
The M1 also integrates RAM in the SoC package, just like the company’s latest A-series processors used in iPhones and iPads. This is a first for processors that aren’t designed for content consumption devices like phones.
Previously, MacBooks featured DDR4 memory soldered on the motherboard. This new approach allows Apple to use a new unified memory architecture. The Apple M1 uses LP-DDR4X memory running at 3733MHz. This allows both the CPU and GPU to access it at high speeds, but this high degree of integration comes at a price.
The downside is that the M1 chip will be available with only 8GB and 16GB of RAM, at least for the time being. Sticking to 8GB or 16GB of RAM on the MacBook Air or Mac Mini shouldn’t be a problem for most users, but many MacBook Pro enthusiasts won’t be thrilled by the prospect of buying a 16GB machine that cannot be upgraded.
Granted, we are used to laptops with soldered RAM, but this is different. With soldered RAM, manufacturers can refresh their product lineup by swapping out the RAM chips with higher capacity ones, e.g. using two 16GB RAM chips instead of two 8GB units. This approach should not require any changes to the motherboard or other components. However, with RAM integrated into the SoC, this would require doubling the memory capacity in the chip package, i.e. a revised M1 chip. Therefore, Apple is unlikely to add a 32GB RAM option in its mid-2021 update, as it will most likely have to wait for a new M-series processor, which could take 12 to 18 months.
In addition to the CPU, GPU, and RAM, the Apple M1 also features a 16-core Neural Engine, a new image signal processor (ISP), Secure Enclave, Rosetta hardware optimization, support for AES encryption hardware, as well as dedicated encode and decode engines for audio and video content. According to early reviews, the latter allows it to outpace x86-based Macs by a significant margin.
The M1 has a standard range of I/O options with a Thunderbolt controller capable of supporting USB 4. However, there’s no support for 10Gbit networking out of the box (if you need it), and the new MacBooks have only two USB Type-C/Thunderbolt ports. Of course, the desktop Mac Mini has a few additional ports compared to MacBooks.
Apple M1 Compatibility and Virtualization Concerns
Since we are discussing new hardware, we cannot make definitive judgments at this early stage, although things are looking quite good. Some applications will need to be optimized for the new processor to ensure support for M1 processors and enable them to utilize its full performance potential. Luckily, the Apple M1 is so fast that some x86 applications running on Rosetta 2 will still run faster than on older x86 chips. Thanks to Rosetta-optimized hardware, the new Macs have enough performance to take the x86-to-ARM performance penalty and come out on top.
It is essential to distinguish between software that is not optimized for the Apple M1 and software that currently cannot run on Rosetta 2. Lack of optimization will result in degraded performance, while lack of compatibility will result in unworkable projects and a lot of frustration.
Virtualization is another source of trouble. We already mentioned issues brought up by the Docker community, and it’s easy to see why x86 images might prove problematic at this early stage. Unfortunately, a lot of information on virtualization support on the M1 processor is still not available. Although Apple isn’t saying much, VMware and Parallels have already announced they are working on M1-optimized updates, though it should be noted that VMware stopped short of revealing a timeline for its rollout.
How serious are these problems? For the time being, some crucial tools either won’t run on new Macs or won’t run properly. These include Docker, Android Studio, and Haskell. The list of tools that will run on Rosetta 2 but aren’t optimized for the M1 is much more extensive and includes Atom, RStudio, PHPStorm, R, Flutter, VSCode, Golang, .NET, and even PHP. They are expected to be optimized for Apple silicon in the coming weeks and months.
You can consult IsAppleSiliconReady.com for additional information and updates. Of course, you can also check the status of each component of your stack on your own.
Designers have less to worry about, as most software suites will work fine, although some still aren’t optimized for the M1. There was a lot of talk about Adobe products and whether or not they will be fully compatible at launch. It seems this won’t be a big problem, as Adobe is expected to roll out updates in early 2021 and make sure everything is working properly. Due to the popularity of Apple hardware among designers, rest assured that Adobe and other software vendors will do their best to optimize software for the new architecture.
Third-party plug-ins for Adobe products are a more significant concern, as it could take a while before they are all updated.
Lest we forget, most servers still use x86 chips, although ARM processors made inroads in certain niches of the server market. For years, Macs were the go-to platform for software developers because they allowed them to work on a UNIX-based operating system running on x86 hardware. They would produce code designed to run on servers using the same instruction set and another UNIX-based operating system. With the M1, this will change as Apple developers will be developing software on ARM hardware and then rolling it out on x86 servers.
On the other hand, new MacBooks can run iOS apps natively, as Macs and iPhones now share the same CPU architecture. Users can download and install iOS apps through the App Store, though the UI might not offer a polished user experience we’ve come to expect from Apple.
Implications for Real-world Users
Apple has managed to design a potent mobile processor that will breathe new life into MacBooks and Mac Mini. Although some of the company’s performance figures were vague, reviewers confirm the new processor easily beats previous generation products based on Intel processors. It even outpaces more powerful desktop CPUs from Intel and AMD in some scenarios, such as video, thanks to dedicated hardware encoders.
So, all is well in the MacBook universe? It’s looking good so far, but it ultimately depends on your priorities and your stack.
Improved Battery Life and Efficiency
The M1 excels at many things. Performance in most scenarios is second to none, and due to improved efficiency, your next MacBook could run a few hours longer with no changes to battery capacity. Efficiency doesn’t only improve battery life. It also means the MacBook Air can deliver a lot of performance with passive cooling. The MacBook Pro has a fan, though it probably won’t spin until you put it under a lot of load. Everyone loves silent computers, and the M1 promises a lot of performance without much fan noise or heat.
There is a caveat worth mentioning. ARM processors tend to be more efficient than their x86 counterparts in low-power scenarios, but due to higher leakage and loss of efficiency at high core clocks, this advantage is likely to decrease under heavy load. Battery life improvements will be higher if you spend most of your time browsing, editing documents, or writing code. They probably won’t be as impressive if you compile a lot of code or do 3D rendering daily.
The MacBook Air, typically used for content consumption and web applications, is likely to benefit more than the MacBook Pro, which is mostly used for productivity and high-load applications. In both cases, though, users can expect a lot more battery life.
But will MacBook Pro users gain a lot of performance thanks to superior cooling, which will enable the processor to run at high clock speeds without thermal throttling? It’s not as straightforward as with Intel chips, which greatly benefit from higher clocks. As we noted earlier, ARM chips are different and they lose efficiency and deliver a smaller performance boost at higher clocks.
Connectivity and Expansion
This has been a source of controversy following every MacBook Pro launch in recent years, as Apple tends to remove physical ports with each new generation. With the new MacBooks, users will have even fewer options, and they may require more dongles and USB Type-C hubs than ever before.
Limited connectivity will not be much of an issue for the average MacBook Air user, but MacBook Pro lovers will have something to complain about, again. Let’s not forget the lack of on-board 10Gbit networking, either. This won’t affect most users, but some professionals still rely on fast, wired networks to quickly copy large video files and databases over their local network. Also, some video professionals are reporting compatibility issues with specialized hardware and peripherals.
Once again, the MacBook Air seems to come out on top, as most of its user base won’t miss things like 10Gbit networking. If you’re starting to see a pattern here, you’re not alone.
Upgradeability and Customization
Finally, here is something the Apple M1 does not excel at. We already outlined the problem with integrated RAM as opposed to soldered RAM modules. Integrating RAM on the processor has its advantages as it simplifies power delivery, reduces the footprint of the motherboard, and unlocks more performance. Still, there’s a high price to pay for this approach. It is hard to see how Apple could offer 32GB or 64GB options anytime soon. If you need a lot of RAM, your only option at this time is to choose a MacBook Pro with an Intel processor.
Is this a deal-breaker for many users? Probably not, as we are talking about 13-inch laptops. If you’re in the market for a powerful mobile workstation, you would probably go for the 16-inch version anyway. So what’s the problem? Well, judging by the performance figures, Intel-based Macs may end up slower in many scenarios, so some users may have to sacrifice CPU performance to get a system with more RAM.
However, if you’re looking to replace your 5-year-old MacBook Air, you should be fine with 16GB or maybe even 8GB of memory. Again, the Air customer doesn’t have to compromise like their Pro counterpart.
Long-term Implications for Linux and Windows Users
For the first time in almost two decades, Mac users will be using processors superior to x86 chips powering Windows PCs. Since 2006, Mac and Windows machines used the same processors, but now Apple has its own silicon to back its operating system. This isn’t just a matter of prestige or fodder for online flame wars – this is vertical integration at a level previously unseen in the industry. Apple now controls its OS as well as its CPU design. It relies on third parties only for manufacturing and commoditized components such as storage, displays, touchpads, and so on.
The ARM architecture offers more efficiency and scales better than x86. Intel’s x86 chips no longer provide substantial performance improvements with each new generation, though it should be noted that AMD is doing better with its Ryzen processors. ARM chips are evolving faster and delivering far greater performance boosts from generation to generation. In under ten years, Apple’s A-series smartphone chips have managed to catch up with Intel’s x86 notebook chips, and the M-series has overtaken them. If we continue to see similar performance improvements with future M-series processors, Apple will be in a very strong position for years to come.
Windows remains stuck on x86, not just because Microsoft hasn’t made much progress with Windows on ARM, but because there aren’t that many ARM processors and hardware platforms suitable for Windows desktop applications. In fact, Apple recently stated that it’s “up to Microsoft” to make Windows run on the M1. However, some intrepid developers already managed to run Windows for ARM on Apple silicon, and the results are good, as it outperformed Microsoft’s own Surface Pro X.
Of course, Apple isn’t going to sell its chips to the higher bidder, but let’s not forget other ARM chipmakers such as Qualcomm and Samsung. It is also worth noting that Nvidia is in the process of acquiring ARM and this could shake up the market as well.
The Hackintosh community might end up being the biggest loser of this transition. In the long run, as Apple starts to tie its OS to its silicon, the Hackintosh may become a footnote in computing history. For the time being, you can forget about running Windows via Bootcamp as well, and Linus Torvalds recently expressed doubts that Linux will get ported to the Apple M1. UNIX-based operating systems can run on ARM and x86, so in theory, it shouldn’t be too hard, but there could be problems with drivers, bootloaders, and so on.
Should I Buy an Apple M1 Mac?
Yes, provided you do your homework first, as early adopters could experience some compatibility issues. As companies update and optimize their products for the M1 processor, most of these concerns will fade away.
If you can live with 16GB of RAM and make sure the tools you use will work from day one, there’s no reason to have misgivings about switching to a new architecture. After all, we are talking about an industry heavyweight with a significant market share. Every software vendor will make sure their products work on Apple hardware, although this may take a few weeks or months. No, you won’t feel like a beta tester, and no, Apple will not treat new Macs like the first-generation iPad or the zero-gen Apple Watch.
While Apple’s marketing may have been vague and overly optimistic in some respects, it is obvious that the M1 is a very capable processor. While it might not outpace some Intel and AMD chips in the high-end segment, the fact that Mac users can get this level of performance in a passively cooled MacBook Air or a compact desktop like the Mac Mini is a testament to Apple’s engineering prowess.
Eagle-eyed readers may have noticed I omitted the MacBook Pro, and hardware enthusiasts probably know why. The difference between M1-based MacBook Pros and MacBook Airs simply isn’t that big anymore. The Pro has a Touch Bar, a marginally bigger battery, and better cooling. Unfortunately, we already mentioned that ARM chips don’t excel at high clocks as much as their x86 counterparts, so this advantage is blunted by virtue of the architecture. In other words, the Pro won’t be noticeably faster than the Air in most scenarios. In burst loads, they should be on a par.
Apple has been removing physical ports from the MacBook Pro for years, causing a lot of frustration in professional circles. With this generation, they went a step further, maybe even a step too far. Like the Air, the MacBook Pro lacks connectivity, expansion, and RAM options desired by enthusiasts and professionals, yet it only delivers marginally better performance compared to the Air. This makes it a much harder sell than the MacBook Air.
Bottom line: the Apple M1 is an impressive feat of engineering that will shake up the industry. No, it won’t kill off cheaper Windows laptops or x86 chips. However, at this point, it looks like it could very well end up hurting the 13-inch MacBook Pro unless Apple can surprise us with a more enticing version soon. A 32GB variant could do the trick, but given the integrated RAM, it’s unlikely we will see one until Apple unveils the M2 processor.