jazelevator.com

Insights on Apple Silicon M2: Unveiling the Hidden Details

Written on

Understanding the M2 Architecture

The WWDC 2022 event was undoubtedly impressive, yet many intricate details may have gone unnoticed. During the presentation, I found myself so absorbed in the information that I had to ask my friend Andrew Gribben to hold his comments until the end. Nevertheless, even with focused attention, we still missed several key points. Thus, it became essential for me to explore the M2 architecture in greater detail and share my insights.

While many—including myself—view the M2 as merely an incremental upgrade, it’s important to recognize the extensive groundwork that Apple has laid over the past decade. We shouldn’t expect radical technological breakthroughs every few years; this is not only unrealistic from Apple’s perspective but also from a user standpoint. It’s time for tech enthusiasts like us to temper our expectations, remembering that advancements in technology require time and are influenced by various external factors.

The M2 may be an iteration, but it builds upon a foundation that has already transformed the PC landscape. Improving upon existing excellence is no trivial achievement!

Performance Metrics and Industry Standards

Numerous reviewers, including some prominent names, have expressed indifference towards the 18% performance enhancement. Here’s a little secret: this increase is very much in line with industry standards, and actually leans towards the favorable end. For example, if a process previously took five hours and now takes a little over four, that’s a substantial improvement. Anyone who has dealt with processing times can appreciate this.

Interestingly, many are more excited about the GPU performance. A 35% increase may seem more impressive compared to the CPU's 18%, but it is likely to benefit a smaller segment of users. Nonetheless, a 35% boost is certainly commendable.

However, the full picture emerges when we consider the system-on-chip (SoC) as a whole. The CPU, GPU, a neural engine that is 40% faster, and 50% additional memory bandwidth collectively paint a much more compelling narrative. In practical terms, the M2 can offer an average of 25% faster performance without any increase in operational costs compared to the M1. This translates to greater cost efficiency for users.

If your primary concern is energy costs—something that is increasingly relevant in today’s economy—the M2 presents an attractive option. While it may not appeal to those who already own M1 machines, users still on Intel should consider making the switch to the M2.

Delving into Technology Advancements

The aforementioned 25% cost efficiency aligns perfectly with the M2 being built on second-generation 5-nm technology. This distinction is crucial. Many may gloss over the significance of the 5-nm process, assuming it to be a mere reiteration of the previous generation. However, the second-generation aspect is what makes the M2 stand out. Delaying advancements for a 3-nm process would have left us with fewer topics to discuss, leaving enthusiasts feeling frustrated.

It's important to note that the so-called "small print" often serves to downplay potential drawbacks. One such limitation is the continued support for only one external monitor. While I believe this won’t be a concern for the majority of users, it’s a factor to keep in mind.

Evaluating Performance Through RAM

A closer examination of Apple’s own press release regarding the M2 reveals an interesting detail about the RAM used during testing. This information comes with both advantages and disadvantages. It’s great for those purchasing an M2 with 16 GB of memory, as they will have clear expectations regarding performance. Those opting for the 24 GB variant will certainly enjoy enhanced performance, although I’ve previously cautioned against fully spec’ing out an M2 machine.

Conversely, those opting for the base model with just 8 GB of RAM may find themselves at a disadvantage. Apple has been somewhat reticent about this, but while the performance gap may not be drastic, it will be noticeable, especially under sustained loads on the M2 Air, which operates without a fan. Users may encounter slower performance due to increased reliance on SWAP memory, which, despite being on a fast SSD, can still hinder overall speed. While the M2 in its base form remains a solid machine, the average gains in performance and efficiency may not apply as strongly in this configuration.

Optimistic Outlook on Single-Core Performance

Let’s end on a high note—single-core performance. Many may not have considered this aspect, but the M2 is expected to outperform not only the M1 but also the M1 Pro and M1 Max in single-core tasks. The architecture learned from the M1 series allowed for similar scores across the board, with differences primarily evident in multi-core performance. This makes the initial 18% CPU enhancement even more appealing, particularly since many everyday applications still operate on a single core.

The M2 provides a glimpse into the future of Apple Silicon iterations. Once we accept that significant breakthroughs come infrequently, we can start appreciating the M2 for what it truly is—a more advanced SoC compared to last year’s model.

Continuing the Conversation on Apple Innovations

Attila Vago—Software Engineer committed to enhancing the world through code. A lifelong tech enthusiast, advocate for web accessibility, and fan of Lego and vinyl records. Enjoys exploring craft beers!