By the end of 2020, we could see the first Macs powered by Apple Silicon. Specifically, a trio of computers — the MacBook Air, 13″ MacBook Pro and Mac mini — immediately garnered a fair amount of attention. Apple was pleasantly surprised by the breathtaking performance and low power consumption. Upcoming models have followed this trend.
Apple Silicon brings a clear dominance in the performance/consumption ratio, which crushes all competition. But when it comes to breaking bread in terms of raw performance, several far better alternatives on the market are ahead in terms of performance.
Apple’s response to this is quite clear — it’s not focused on performance but on performance per watt, the power-to-power above ratio. But there is one point where it may pay the price.
Is low power always an advantage?
In principle, we have to ask ourselves a very fundamental question. Although this strategy seems perfect at first glance — laptops have extreme battery life and offer full performance in virtually every situation — is low power always an advantage?
Doug Brooks, a member of Apple’s marketing team, weighs in. In his words, the new systems perfectly combine top-notch performance with low endurance, which puts Apple computers in a fundamentally advantageous position.
They outperform virtually all competitors in this respect. But looking at the whole situation from a slightly different perspective, it looks completely different. As we mentioned above, for example, in the case of MacBooks, the new systems play an extremely important role in favour of the MacBooks.
But the same cannot be applied to the so-called high-end models. Let’s pour ourselves a glass of wine. Probably no one who buys a high-end computer and needs maximum performance has more regard for its consumption.
It’s more or less already associated with it, and no one cares about raw performance. So although Apple boasts of lower consumption, it may fall slightly in the target group because of this.
Apple M2 Pro: Core i7-12700K performance at ~3× lower TDP
The twelve-core Apple M2 Pro in the version designed for the desktop Mac mini delivers very nice performance. Compared to Apple’s M1 Pro, it sped up by about half. Apple didn’t overdo it in its presentation, even holding its ground a bit.
The first result of the M2 Pro in GeekBench shows a very nice competitive position. However, keep in mind that this was the desktop version. In laptops, Apple’s SoCs were a bit slower in the last generation.
The 15k point performance in total load is numerically closest to the Core i7–12700K. However, that’s a processor with a 125-watt TDP (power consumption limits aren’t even worth addressing in that situation), so it’s better to compare it to something closer in power terms.
A candidate would be the results of the 45W Phoenix APU, but we don’t have those yet as their availability won’t be until next month. Plus, it’s an octa-core, while the M2 Pro is a twelve-core.
The closest product and configuration I could develop is the Ryzen 9 7900, limited to 45W TDP. Both are twelve-core configurations. Both are 45W products, and both are made with a 5nm process, although with Ryzen, some of the silicon is 6nm.
However, I’ve only found the result of the Ryzen 9 7900X limited to 45W TDP. This is a slight disadvantage because the X series uses higher leakage silicon handling higher clock speeds (but more or less at the expense of power). In contrast, the 65W series (without X) goes to lower leakage silicon (lower maximum clock speeds and power consumption).
For an indicative comparison of how ARM is currently doing compared to x86, this will have to do. At 14160 for x86 vs 15013 for ARM, ARM has a 6% lead, which is not much when considering the manufacturing process.
We are in a significantly different situation than when Apple came out with the M1. No product in the x86 world could offer a comparable combination of performance and power.
Single-core performance is not a big surprise. The current x86 high-end offers roughly 10–15% more. At the same time, this is not a major difference, on the other hand when Apple released the M1 in November 2020.
It was outperforming AMD’s desktop offering, and only Intel’s high-end Rocket Lake models stood above it, whose goals were to achieve the highest single-core performance at the expense of multi-core performance and operating characteristics.
So, on the one hand, Apple has managed to bring a significant intergenerational shift. On the other hand, staying on the 5nm process has shown that the top of the ARM world and the top of the x86 world are pretty close in their capabilities.
The problem with the Mac Pro name
It’s clear that this more or less moves us on to arguably the most anticipated Mac of our time. Apple fans eagerly await the moment when the Mac Pro with the Apple Silicon chipset will be shown to the world.
When Apple revealed its plans to move away from Intel, it said it would complete the entire process within two years. But he missed that deadline, and so far, the wait for Apple’s most powerful computer is more or less still in sight.
There are a lot of question marks hanging over it — what it will look like, what will be running in its guts and how it will perform in practice. Indeed, it’s quite possible that, given the zero modularity of Macs, the Cupertino giant will clash with Apple Silicon, especially in the case of these high-end desktops.