The Apple M1: The Beginning of a New Era?
This article was originally published by David Howell.
The current iMacs and Macbooks are the last and perhaps the best with Intel silicon at their heart. As Apple is moving to its own custom CPUs, what does this mean for the desktop PC and the masses of applications that have been Intel-compatible for decades?
Last November, Apple announced the M1. Apple had already publicised its intention to move its own silicon to give it ultimate control over its PCs’ software and hardware. Dubbed an SoC (System on a Chip), the Apple M1 chip’s architecture’s unified design should deliver performance boosts the industry has not witnessed for several years.
The Apple Mi uses 5-nanometre process technology delivering 16 billion transistors. Apple claims the M1 delivers up to 3.5x faster CPU performance, up to 6x faster GPU performance and up to 15x faster machine learning, all while enabling battery life up to 2x longer than previous-generation Macs.
“There has never been a chip like M1, our breakthrough SoC for the Mac,” said Johny Srouji, Apple’s senior vice president of Hardware Technologies. “It builds on more than a decade of designing industry-leading chips for iPhone, iPad and Apple Watch, and ushers in a whole new era for the Mac. When it comes to low-power silicon, M1 has the world’s fastest CPU core, the world’s fastest integrated graphics in a personal computer, and the amazing machine learning performance of the Apple Neural Engine. With its unique combination of remarkable performance, powerful features and incredible efficiency, M1 is by far the best chip we’ve ever created.”
Speaking to Silicon UK, Jason Wudi, CTO, Jamf explained: “For the majority of mainstream businesses, technology is an accelerant to their unique value, and if there is one solution that brings forward ‘more, faster, cheaper’ without significant trade-offs, the M1, will absolutely make an impact.
“Whether Apple will lead the enterprise to switch from Intel, or other competitors will take the baton and use the concepts of the Apple M1 to deliver a new and more compelling offering is yet to be determined. It may not be an M1-based approach, but I expect more experiments and offerings in architecture and software to begin to appear following this launch.”
With a new powerful chip to choose for IT investments, it’s early days for any wholesale move to M1. Clearly, for some of Apple’s traditional strongholds such as creative arts, the M1 will be quickly embraced.
For CTOs looking at their hardware roadmaps, the M1 will certainly feature, but a cautious approach will be taken to ensure the applications and services needed have native versions available. Rosetta 2 is workable but does inject a level of risk and anxiety when stability, performance and security are concerned.
Evolutionary or revolutionary?
The initial responses by industry watchers about the performance of the M1 have been favourable. Is the M1 a gamechanger?
“The M1 is a gamechanger in the aspects of computing performance and hardware design. By using an ARM-based processor design, Apple was able to wring out the cost and, more importantly, parts from within the laptops. The more efficient M1 processor has allowed Apple to do away with legacy components like fans, which is a moving part and suspectable to failure – not to mention making the Mac quieter and slimmer. These new Macs adopt a design more representative of an iPad than a Mac laptop.
“Normalised to performance, the Apple M1 design also requires less RAM, which has significance to Bill of Materials costs. The base model of the M1 powered MacBook Air (MBA) dropped $100 relative to the base 2019 MBA model featuring an Intel processor (keeping DRAM and Storage constant). While 2019 MBA started at $1099, a subsequent update to Intel’s MBA design came out in early 2020 for an equal $100 discount. This interim cost reduction is likely due to usual component cost erosion and the “end of life” of the Intel design.
“In terms of performance, Apple is better able to optimise MacOS to run at the deepest silicon levels since they own the entire stack. Macs and PCs have relied on Intel x86 chipsets for decades, which are designed to support multiple operating systems, therefore, OEM is not able to reach very deep into the circuitry of the Intel chipset to optimise for their respective OS’s. For Apple, who likes to control every aspect of their products, the M1 allows for this ownership of the entire hardware and software stack is unique in the industry allowing them to realise a huge jump in performance and cost.”
Apple now controls the hardware and software of their devices. Is this closed approach good for the tech industry and customers?
“The total control of hardware and software benefits Apple, the industry and customers alike by providing a step-function improvement in computing. However, left unchecked, Apple can leverage their position to extract more profit from consumers or capture most of the value [of the supply chain] leaving industry players like Intel with less of the industry value pie. “Typically, economists argue that competition is good for consumers because it drives innovation and keeps costs down. Apple felt it needed to create its own chipset points to the fact that there had been little competition and innovation for computer chipsets. From that perspective, the introduction of the M1 is a good sign for the industry as other competition which will drive Intel to innovate and innovate faster.”
Will Intel be able to compete as the M1 rolls out to the rest of the Mac range of PCs – especially the Mac Pros?
“Intel can still compete in the higher-powered Macs. If we look at Apple’s offering for the MacBook Pro, the more expensive models still use Intel processors. And for the professional equipment like the Mac Pro, those designs are exclusively Intel. For high-performance computing (HPC), Intel is still the go-to processor. What we’re seeing with this move by Apple is that for mobile computing, ARM processors are moving up in capability – rather than Intel using HPC designs to address the mobile designs. We really need to make that distinction of HPC processors and mobile processors; ARM has essentially won the mobile processor war.”
Do you see any limitations to Apple’s approach to its processor design?
“Moore’s Law perhaps, Apple’s M and A processors are only as good as the fabrication technology leveraged by foundries such as TSMC. The next technology node will be 3nm which is using exotic technologies such as EUV (extreme ultra-violet) lithography. The concern now is that Moore’s law will soon run out of steam and the rate of scaling (making smaller and smaller transistors) will start to plateau.”
For business owners, is the M1 something that will persuade them to switch from their Intel PCs and later servers?
“Unlikely as it is the choice of OS’s which is the driving force for businesses to choose between ARM-based and x86 based PCs. In fact, Qualcomm has attempted to bring ARM-based computing to laptops for some years now, but the performance is mixed. While these Qualcomm powered ACPC (always connected pcs) improve dramatically on battery life and cellular connectivity, the overall computing performance, however, is limited by Microsoft Windows.”
A calls to ARMs
Joining the M1 is Apple’s latest desktop OS, Big Sur. This iteration is significant as the OS has been optimised for the M1 chip. The ability to run iOS and iPadOS apps on a Mac should not be underestimated, as this level of integration could be significant, especially for developers. And Rosetta 2 allows Intel applications to be translated on the fly, while native versions of these applications are developed. All of the significant application suites from Microsoft and Adobe have already been through the optimisation process.
Apple’s move to further leverage ARM isn’t surprising. Their iPhones have been astonishing successful using this technology. It was only a matter of time before that power would be extended to their desktop and notebook PCs. And ARM is clearly a trend across the tech sector’s leading players: Amazon have their Graviton processor for AWS
[https://aws.amazon.com/ec2/graviton/], which the company claims is 20% faster than the ubiquitous Xeon and 20% cheaper. Microsoft also saw these changes coming and ported Windows to ARM.
And let’s not right off Intel. The Tiger Lake processor that focuses on laptop users and later Alder Lake, which should improve battery life, are coming. How they compare in the real world against the Apple M1 is yet to be seen. And Nvidia Arm-compatible CPU they call Carmel, could give Intel PC manufacturers at least some performance gains against M1. However, with improvements coming to Apple’s new chip coming this year, it remains to be seen whether Intel and Nvidia can deliver anything near M1 performance as it stands today let alone what the M1 could offer by the Autumn.
Jamf’s Jason Wudi concludes: “The one thing that Apple has done better than any of its competitors is to focus on the outcomes, not the components. It isn’t about clock cycles, or the pixels, or any of the technical specifications. It’s about what the user in front of the device is trying to do, and I believe the M1 represents a next step in what Apple can do in terms of bringing the full capability of the hardware, software, and services together to achieve the outcome the user truly wants. That is what is special to me – the focus Apple has on experience that allows them to do just that.”
For CTOs looking at their next hardware refresh cycle, is the Apple M1 even on their shortlist? That will depend on their mission-critical desktop and hosted applications they have. As M1 becomes more widespread – especially as the rumoured iMac redesign becomes a reality, and the Mac Pro adopts M1, all current Intel applications will see M1 native versions. The timeframe for that is, of course, unclear.
For now, Apple has drawn a line in the sand. Whether businesses will be prepared to embrace this new era will quickly come into focus as Apple release more of their range equipped with M1. All eyes are on the Pro range and how these will perform, but your business could be waiting until sometime in 2022 to find out if your enterprise needs the M1.