For people who are familiar with the matter, you can skip this article as you may think it is silly to even ask this question. But I am sure there are many out there who, like my colleagues, are asking themselves the question like the headline: Why does not the Apple M2 Ultra have twice the performance of an M2 Max chip, even though it uses two of them?
First of all, the Apple M2 Ultra is much more complicated than two M2 Max chips glued together. It uses a special connection called Apple’s UltraFusion architecture to connect two M2 Max chips together. This architecture allows the two chips to communicate with each other at very high speeds, which can lead to significant performance gains, but it also comes with some overhead.
The M2 Ultra has to share this memory between two chips, which can lead to a slight performance loss. That being said, the M2 Ultra can store more data with a unified memory of up to 192 GB, which can lead to faster performance compared to the 128 GB of the M2 Max.
In addition, the M2 Ultra is clocked slightly lower than the M2 Max. The M2 Ultra has a base clock of 2.3 GHz, while the M2 Max has a base clock of 2.5 GHz. This means that the M2 Ultra is slightly slower than the M2 Max when running at the same power.
Not to mention that the M2 Ultra does not always run at full speed. The M2 Ultra can run at different speeds depending on the workload. For example, it can run at a slower speed when it is not heavily loaded to extend battery life and reduce heat dissipation.
Recent benchmarks have shown that the M2 Ultra is not quite as fast as Apple claims. In fact, it only offers about 18% more performance than an M2 Max chip. Still, the M2 Ultra is the largest and most powerful chip Apple has developed to date. It will be built into the new Mac Studio and Mac Pro, which have entry-level prices of $1,999 and $6,999, respectively.
