The verdict is in: The Apple M1 chip is great!

Reviews of the new Apple M1-powered Macbook Air and the M1-powered Mac Mini have been overwhelmingly positive. Reviewers talked about blazing speed, low running temperatures, long battery life (for the MacBook Air) and surprisingly affordable pricing — all hallmarks of well designed computers — features that will surely appeal to a large consumer base.

Most of the credit for the hype and interest will have to be attributed to new APPLE M1 chip. But what exactly does this chip bring to the table?

Photo credit: Apple Press Release: Apple unleashes M1

Apple M1 SoC highlights

• High Speed — As NOTEBOOKCHECK.NET states: “In CPU as well as GPU performance, comparable chips from Intel and AMD are surpassed in terms of the theoretical raw performance - and this is the entry-level configuration of the Air.”
• High Power Efficiency — Again from NOTEBOOKCHECK.NET: “The new Apple M1 also has a positive effect on the battery life. With its energy-saving cores, it is able to save a lot of power in our WLAN battery test, and it can even last through a full work day at maximum display brightness.”
• Low Operating temperature under load — NOTEBOOKCHECK.NET: “In regular load situations, the Air even remains particularly cool, which is a sign of the great efficiency of the Apple M1”.
• Onboard Subsystems — the Integrated Subsystems play a significant part in making the CPU/GPU faster, while reducing power requirements
• Supported by ROSETTA 2 — Emulation software to ensure backward compatibility with INTEL-based applications, while the M1-optimized application base is being built up

This innovative System-on-a-Chip (SoC) design, however, was not developed overnight. Let's look at the history of Apple-designed Processors, as summarized by WIKIPEDIA.

• Investments in processor chip technology started with the iPhone A4 chip, manufactured by Samsung, back in 2010. This A-series chip was the first Apple in-house SoC design.
• The Apple A6 was introduced Sep 12 2012, and was manufactured by Samsung using a 32nm process.
• The Apple A8 was introduced Oct 16 2014, and was manufactured by TSMC using a 20nm process — This was the first time TSMC an Apple A-series chip was produced by TSMC.
• The Apple A10X was announced Jun 5 2017, and was manufactured by TSMC using a 10nm process.
• The Apple A12X Bionic chip was announced Oct 10 2018. It was manufactured by TSMC using a 7nm process.
• The Apple M1 SoC chip was announced November 10, 2020. It was manufactured by TSMC, using a 5nm process.

Apple actually has been building its chip design capabilities since 2010, starting with its iPhone chips. With every successful chip design, Apple came closer to the goal of making a chip with a similar performance envelope to those of current INTEL and AMD desktop CPUs of the day.

But Apple also had an ace up its sleeve, the ARM architecture — that the A-series (and subsequently M-series) chips were based on — is highly power efficient, by design. The traditional INTEL and AMD CPUs were primarily built for performance (with power efficiency taking a back seat), also by design. These priorities would set the stage for the CPU battles being waged today.

Apple M1 SoC lowlights

• Not upgradeable (as designed) — As NOTEBOOKCHECK.NET states in the link above: “There is no fan needing maintenance, and all the components such as the CPU, GPU, SSD, RAM, and WLAN card are directly soldered onto the main board”. This means that the user will need to select the appropriate configuration upon purchase (in the same way we currently do when purchasing mobile phones today), as the systems cannot be upgraded thereafter
• Not repairable (as designed) — Since the key subsystems are soldered unto the motherboard, rather than connected to it via interface ports or slots (e.g. PCIe slots / DIMM slots), it is obvious that these subsystems were designed to be maintenance-free, much like a mobile computing device (i.e. iPhone or iPad) rather than a regular computer

Was this “CANNOT UPGRADE / REPAIR” approach a conscious attempt by Apple to force users into buying higher capacity systems at the start, and / or make them purchase another system when they needed the additional capacity — rather than just upgrading their current system when needed?

Due to the SoC design, it can be argued that soldering subsystems to the chip is really the best way to go (thus making it difficult / impossible to upgrade). And in the same way, the lack of repair options can also be traced to this SoC design.

So what was the breakthrough?

The chip technology breakthrough was combining the ARM architecture based chip with an SoC design on the APPLE M1. By putting all systems on the same chip, Apple was able to increase performance significantly — while maintaining the power efficiency advantage of the ARM architecture.

So how did Apple achieve this huge leap in overall performance?

By combining all the key subsystems into one chip, the APPLE M1 design was able to increase CPU (central processing unit) and GPU (graphics processing unit) functional speeds, and minimize latency between subsystems (since they were now all on one chip) — while decreasing power consumption, by being able to reduce the effective clock frequency or through voltage-scaling; and minimizing (power-consuming) external subsystem calls.

The TSMC 5 nanometer (5nm) chip manufacturing process also helped boost CPU/GPU processing power and increase power efficiency.

The move from the INTEL-based platform to the M1-based platform was not a spur of the moment decision. It took a lot of planning and effort on Apple's part to come up with the M1 Macbook Air and M1 Mac Mini — The new chips and circuit boards had to be designed and tested (with TSMC, its chip supplier), key MacOS systems updated to run on the M1 chips natively, and Software Development Kits (used by developers to create M1-optimized applications) updated and released to developers, so that transitioning efforts can being as soon as possible.

A breakthrough high-perfornance chip with no software is a dud. Apple knew this, so it had to come up with a way for users to keep using their INTEL-based software, while M1-optimized software is being developed. AND, the mere ability to run the INTEL-based software would not be enough — the M1 systems will either have to run the old software at a decent speed or at a higher speed.

To ensure that current INTEL-based Mac software will be able to run on the M1 chip, Apple came up with ROSETTA 2 — a nod to the original ROSETTA emulation software used when Apple transitioned from its PowerPC architecture to that of INTEL. Apple basically had to make sure ROSETTA 2 was running well enough (on launch date) to minimize performance issues on Apple M1 systems, while users waited for M1-optimized software to be released.

But the real breakthrough actually came from Apple making sure that the new APPLE M1 chip worked with existing software — namely the Mac OS operating system and all the MacOS software applications. The M1 SoC (and attached hardware) make up one half of the ground-breaking system. The second half is made up of the MacOs Operating System and M1-optimized application software.

Apple has the unique position (in the market) of being able to dictate what goes into its hardware AND software. This unique position allows it to define and design synergies into its hardware and software — which will be difficult to emulate and achieve in the other computing platforms (e.g. Windows / Android / Linux) in use today.

Another by-product of the transition from INTEL-based to APPLE M1-based MacOS systems is that the entire MacOS application code base will also begin transitioning to the new platform. Several popular applications (Firefox, MS Office) have already been ported and optimized for the APPLE M1 platform — and surely, more application ports will be announced in the next few months.

As more and more applications move to the APPLE M1 platform — with the corresponding increase in performance due to M1-specific optimizations — and as INTEL-based version releases slow down, the incentives for users to move to the APPLE M1 ecosystem will increase.

What’s next?

We are already at the point in history wherein personal computers are small enough to lug around everywhere we go (i.e. our mobile phones) — one recent high end MOBILE PHONE CPU (e.g. Apple A13 Bionic, announced 2019) is roughly 2.3x more powerful than one high end DESKTOP CPU of just a little over eight years back (e.g. Intel Core i7-3960x, from back in 2011). The APPLE M1 chip — using the same comparative method — is roughly 2.95x more powerful than the Intel Core i7-3960x.

Although direct comparisons between chip performance cannot be made (due to the lack of pertinent benchmarks), the quoted performance estimates above can be inferred from the values in the table below:

Comparative Benchmarks of select CPUs

• Intel Core i7-3960X (from 2011) — 176,160 MIPS
• AMD Ryzen 7 1800X (from 2017) — 304,510 MIPS / Geekbench 5.0 64 bit Single-Core median score of 1,006.5 / Geekbench 5.3 64 bit Single-Core median score of 1,023
• Apple A13 Bionic (from 2019) — Geekbench 5.0 64 bit Single-Core median score of 1,343 / Geekbench 5.3 64 bit Single-Core median score of 1,336
• Apple M1 (from 2020) — Geekbench 5.3 64 bit Single-Core median score of 1,745

Sources: NOTEBOOKCHECK.NET / WIKIPEDIA.ORG

The fact that a low power MOBILE CPU can perform at effectively triple the speed of an 8 year old cutting edge DESKTOP CPU just shows the significance of the improvements made over that time.

The introduction of the APPLE M1 chip has opened the doors for Apple to new markets — as its competitors struggle to catch up in the technology game. Although the M1 chip can be classified as an incremental innovation (when compared to its previous INTEL iteration), subsequent innovations from Apple in the chip arena may be more disruptive to the market.

With Apple's innovative SoC design, we can expect a raft of new developments in CPU and GPU chips (from Apple and its competitors — Intel, AMD, Qualcomm and others) in the years to come.

The gradual and continuous push to smaller and more power-efficient chips and subsystems is slowly transforming our lives, as companies find more ways to automate and enhance parts of our daily life through personal computing devices — much like the way our lives were (and are being) transformed through functionalities we currently access through our mobile phones — e.g. video chat & email, GPS navigation, information search, online transactions and health-fitness (e.g. heart-rate monitoring) applications.

With every increase in computing power and improvement in power efficiency, the list of useful applications and functionalities will just get longer and better — hopefully providing us with more convenience, and more ways to make the most of our lives.

Imagine the day when you could lug around a small deck-of-cards sized computer (either in your pocket or your bag), that is hooked up to some virtual reality (VR) goggles, with even more computing capability than today's desktop systems — the possibilities seem endless. With the relentless drive for innovation, that day may come sooner rather than later.

Related links on INNOVATION

Can industry leaders truly innovate? — Innovations have the potential to disrupt markets, and provide opportunities to leapfrog competitors. We look at how market leaders push (or don't push) innovations in order to stay on top.
Gamechangers: On Radical Innovation — We talk about some examples of radical innovation — innovations that changed the game, so to speak.