A cursory glance might make us think that the M1 is just a beefed-up mobile chip, as it is quite similar to the 'Apple A14 Bionic' in the current iPhone 12 lineup, with 4 high-performance and 4 high-efficiency cores. And it has I/O controllers and built-in memory that makes it more of a mobile system-on-a-chip (SoC) than a plain CPU.

Apple claims that it has the world’s fastest CPU core in low-power silicon. And early reviews show that this claim isn’t far from reality, with many reviewers gushing over its performance. But how did Apple pull this off without fundamentally changing the design from what it’s using in its smartphones?

Part of the answer is that Apple used a lot more transistors in the M1 than the A14, over 4 billion more according to the company. But while this may have boosted performance, the real secret sauce lies in how Apple’s design, both in the A14 and the M1 fundamentally differs from x86, which is the architecture that Intel has been using for decades.

Apple has an Architecture License agreement with ARM Ltd. (owned by Nvidia), the broadest license that ARM sells. ARM was founded in 1990 as a joint venture between Apple and Acorn Computers. Apple's license is also perpetual that is the reason why Apple isn't interested in acquiring ARM. Apple does not use standard ARM designs, they use custom cores based on ARM architecture and the ISA, that is the reason why we see a big performance difference between Apple-designed SoCs vs. the Qualcomm or Samsung designs based on standard ARM Cortex-Ax cores. In 2013 Apple went a year ahead and delivered the first 64-bit SoC with iPhone 5S before even ARMs 64-bit design was finalized. 

Apple’s philosophy was to make the chips much more parallel than conventional CPU designs. Both the M1’s decoder which translates incoming instructions, and the execution units that process them, are wider, meaning they can accept more instructions at once. Additionally, the M1 can go significantly deeper with out-of-order execution, meaning that it can read ahead on the page and anticipate which instructions the program will need to have processed ahead of time to a greater extent than that of x86.

Then you have the fact that the M1 features a lot more L1 cache than x86 processors, which is the fastest cache memory available to a CPU core. And because it’s based on a chip originally meant for mobiles, it does all this while drawing significantly less power. But hold on a second, couldn’t Intel just implement some of the changes themselves and catch up with Apple? Well, it turns out that might be quite a challenge as the x86 architecture has some inherent limitations.

This all made M1 chip much faster than the Core i7 in the previous generation Mac Mini, that’s especially impressive when we consider that this is a CPU with no hyper-threading and with fewer cores. In single-threaded performance, even the lowly MacBook Air, which is the slowest of the three new M1 machines, surpassing the performance of a Core i9 iMac. This shows that if Apple had included 8 performance cores in the M1 instead of just 4, the M1 could’ve met or beaten the Intel Core i9’s 20 threads in multi-threaded performance.

This is exactly the reason why Rosetta 2 doesn’t face the huge performance penalty like other x86 emulators that we’ve seen in the past, like Microsoft’s for example. This is mostly down to hardware tweaks made in silicon to accelerate common x86 load/store instructions. Essentially, Apple Silicon can take x86-like instructions and process them directly rather than having to fully translate to ARM. As of right now, the only ARM CPU that can run x86 programs like this is the M1.

Apple is betting that because developers who have large user bases on the Mac platform would like to it that way, those developers will adapt and come out with versions specifically written for the Apple Silicon sooner than later. This approach is more or less in line with Apple’s playbook for other new products. They bring a new platform to market and use their brand power to force developers to catch up, which is something that they hope will again happen with the M1.

Ultimately, it’s not hard to imagine offering a more tightly controlled App Store experience for Macs, with programs being specifically vetted for full compatibility with Apple Silicon, especially as we expect to see more of Apple’s chips in desktop Macs down the line after seeing how powerful their first crack at a laptop processor has been.