Apple Silicon M2: our complete analysis of the new MacBook Air and Pro chip

If we wanted to sum up Apple’s M2 chip in a few words, it is by writing that this is a small evolution of the M1 chip introduced by the company in 2020. So no revolution on the program , but an upgrade in small increments to bring a bit more performance while further constraining the energy footprint. The M2 chip as presented is therefore not intended to completely bury the M1 generation, the Pro, Max and Ultra versions remaining much faster.

If we refer to the “photo” of the chip presented by Apple, the M2 is significantly larger than its predecessor. A point that is not surprising when we look at the number of transistors, up 4 billion to reach a total of 20 billion. Two solutions were then offered to Apple to accommodate this additional transistors: change the fineness of engraving to maintain an equivalent surface (or even reduce it), or enlarge the surface of the chip while maintaining an engraving process identical to that used on the M1. The firm chose the second option by again using a 5 nm process, a little more advanced however (TSMC N5P no doubt), which guarantees a slight optimization of efficiency.

At first glance, the processor part does not seem to move one iota. There is a hybrid configuration with four high-performance cores and four low-power cores. Unfortunately, Apple gives only a few technical details about these cores and is careful not to define their precise nature.

However, by proclaiming an increase in performance of 18% at best, the company gives us some clues. The high-performance cores must therefore be based on the Avalanche architecture of the A15 Bionic processor of the iPhone 13 and iPad mini 2021. No revolution compared to the Firestorm cores of the M1, but some small optimizations and an ability to hold a higher operating frequency high.

The biggest change is undoubtedly to be found on the side of the low-power cores which should be based on the Blizzard architecture of the A15 Bionic. If they still aim for energy efficiency, they are much faster than the low-power Icestorm cores of the M1. What severely boost performance in applications jointly exploiting the eight cores of the chip. These small architectural improvements – because let’s remember, these are clearly more optimizations than radical changes in architecture – are coupled with the classic trick of processor designers, namely the use of a larger cache. The 12 MB of L2 cache are thus abandoned in favor of a 16 MB L2.

These improvements are probably coupled with an increase in operating frequencies, again allowing performance gains. Apple has also given a boost to its memory interface which now peaks at 100 GB / s, against 68 GB / s on the M1. This increase in bandwidth is authorized by a switch to LPDDR5-6400, still of the unified type. Under this term hides the idea that the memory can be used by the system as well as by the graphics circuit. It is also directly associated with the M2 chip by being housed on the same physical envelope.

Apple obviously didn’t just make some changes to the processor part. The graphics circuit has also been reviewed, although we are not aware of the exact nature of the improvements made. Apple speaks for example of a Next Generation architecture, much faster with a raw power of 3.6 TFLOPS, against 2.6 TFLOPS for the GPU of the M1. Although adjustments and optimizations have been developed for the GPU and its operating frequency, this boost in performance is primarily due to the use of 10 cores instead of eight cores, which results in a slight increase in the energy footprint (15 W versus 12 W). A 25% increase in the number of cores which results in a performance increase of 35%. This last “surplus” is not explained, but we note that Apple mentions an enlarged cache, which, combined with a frequency once again undoubtedly revised upwards, leads to the 35% mentioned.

The block dedicated to artificial intelligence is also evolving. The Neural Engine is a 16-core type and probably also taken from the A15 Bionic. Capable of processing up to 15.8 trillion operations per second, it is 40% faster than the M1’s Neural Engine. The Media Engine, a block dedicated to video management, is evolving smoothly. It still allows for faster encoding and decoding of H.264 and HEVC video streams, but is no longer limited to 4K streams and is now capable of processing 8K streams. It is also able to speed up the processing of Apple ProRes streams, but still ignores the AV1 format, which is nevertheless gaining momentum. Finally, Apple has not seen fit to implement support for Thunderbolt 4, thus still limiting management to Thunderbolt 3. USB4 is on the other hand well on the program, with a maximum of two ports managed by the chip.

There remains the question of the behavior of the M2 chip against the competition. If Apple was careful not to comply with the exercise of comparison a few years ago, the firm now delivers some meager clues through some imprecise curves and poor in detail. For processor performance, Apple positions the M2 against the Intel Core i7-1260P, a midrange chip that targets ultrabooks. The processor of the M2 is announced as slower in absolute terms (difference of around 13%), but much less greedy in terms of power consumption, the M2 only requiring about fifteen watts to operate against 35 to 55 W for the chip Intel. Enough to allow a much quieter operation – the MacBook Air M2 is also completely passive – and obtain a much greater autonomy, which was already the case with the MacBook M1. Apple also mentions superior performance to the Core i7-1255U, but does not compare to AMD’s Ryzen solutions.

Regarding graphics performance, Apple obviously did not want to offer too much comparison, contenting itself with evoking performance more than twice as high as on the Intel Core i7-1255U iGPU. Still, we are talking about an integrated graphics circuit with a memory bandwidth of barely 100 GB/s. In all likelihood, and in particular based on the 35% performance increase announced compared to the M1, we should not see a revolution.

We are betting on a solution capable of doing better than Intel integrated graphics circuits and coming quite close to what AMD offers on its recent Ryzen 6000 mobiles based on RDNA2 iGPUs. Not enough to transform the MacBook M2 into a gaming machine, even if the MetalFX Upscaling technology announced at the same time as the M2 will undoubtedly allow pretty things. For our part, we are rather eager to see how the hardware acceleration of the iGPU behaves on compatible applications, because it is much more in this field that Apple has a card to play.

We can’t wait to test the MacBook Air M2 and MacBook Pro M2 in our labs, the first two laptops using Apple’s new chip. The opportunity for us to confront this promising SoC against the new processors from Intel and AMD. However, it will be necessary to be patient: the pre-orders are not yet open and the delivery of the first models will not take place before the current month of July.