Apple finally lifted the veil from the M2 Ultra at WWDC 2023, and it is now the company’s fastest custom silicon that is not only found in the Mac Pro but the Mac Studio too. The latest SoC directly succeeds the M1 Ultra, but just how big of a difference did the company bring in a single generation? Here, we compare the specifications and the performance differences of both chipsets.
For some reason, Apple chose TSMC’s 5nm process for the M2 Ultra, which has not changed with the M1 Ultra. Likely, A16 Bionic’s 4nm node was not chosen due to higher costs, but this means that none of the company’s custom chips featured in its Mac family have gone beyond the 5nm threshold. Despite this limitation, Apple has managed to improve the M2 Ultra, as it features 134 billion transistors, which are 20 billion more than the M1 Ultra’s 114 billion transistors.
Another similarity shared between the M2 Ultra, and M1 Ultra is that they were both developed using a process called UltraFusion, which combines two M2 Max and M1 Max to create a more powerful version of the Apple Silicon. This process may have improved because the new M2 Ultra can accommodate more CPU and GPU cores this time, not to mention support for higher unified RAM and at the same 800GB/s memory bandwidth as the M1 Ultra. Previously, the maximum supported unified RAM support was 128GB, and it has jumped to 192GB with the latest release.
Also, users can configure their latest Mac Pro or the Mac Studio with the M2 Ultra with up to a 24-core CPU and 76-core GPU configuration, whereas earlier, the M1 Ultra was limited to a 20-core CPU and 64-core GPU. The 32-core Neural Engine remains the same across both chipsets, but here is the kicker, the unit in the M2 Ultra is 40 percent
Read more on wccftech.com