In the run-up to the launch of Apple’s M1 chip, the tech giant emphasized the scalability of its mobile System-on-Chip (SoC) architecture, hinting at the potential to power everything from MacBooks to the Mac Pro. The M1’s resemblance to the A14 Bionic was evident, and the same pattern continued with the M2 series, based on the A15 Bionic. However, when it came to the M3 chip, Apple’s silence prompted enthusiasts to dig deeper.
The M3 chip was introduced after the A16 Bionic powered the iPhone 14 Pro models and the more recent A17 Pro empowered the iPhone 15 Pro models. Of these, only the latter was fabricated using TSMC’s new 3nm node (N3B), just like the new M3 series. While the M3 chip shared a similar GPU architecture with the A17 Pro, which Apple touted as the “biggest GPU redesign in Apple’s history,” there were still lingering questions.
The NPU (Neural Processing Unit) in the M3 was an intriguing piece of the puzzle. Despite its 16 cores, it delivered 18 TOPS (trillion operations per second), which was notably slower than the latest iPhone chips. This hinted at the possibility that the M3’s NPU was closely related to the A16 Bionic’s 17 TOPS, with the minor difference potentially attributed to the shift to the 3nm node or a clock boost.
As for the CPU architecture, it remained a bit more enigmatic. However, the shared NPUs between the M3 and the A16 Bionic hinted at commonality. Notably, the next-generation GPU architecture introduced in the A17 Pro and M3 was initially slated for the A16 Bionic. Reports indicated that Apple had to pivot away from this GPU due to overheating issues and power consumption, suggesting a deviation from the original plan.
This alludes to the possibility that the M3 series was initially intended to be based on the A16 Bionic, complete with its proposed next-gen GPU. While the A16 Bionic had improved graphics memory bandwidth over the A15 Bionic, the implementation of the new GPU architecture was delayed until the A17 Pro. Apple likely had the technical documentation and schematics for the M3 with the A16 architecture, including the new GPU architecture, ready. This allowed them to make sure the GPU was refined and ready to roll for the A17 Pro, while the M3 series retained the intended CPU and NPU architecture from the A16 Bionic.
In conclusion, the Apple M3 series appears to be a fortuitous blend of the A16 Bionic and A17 Pro, showcasing the unexpected convergence of these two powerful architectures. This accidental hybrid offers a fascinating insight into the complex world of chip development, where even the most ambitious plans can take unexpected turns.