Apple’s M2 chip debuts
At the WWDC 22 developer conference, in addition to introducing iOS 16, iPadOS 16, watchOS 9, and macOS Ventura, Apple also released a new MacBook Air equipped with a new generation of M2 chips. As the successor of the first self-developed chip M1 in 2020, Apple has made many improvements on the M2 chip to shoulder the heavy responsibility of launching the M2 series of chips.
The news about the M2 specification has already been circulated, but the actual product is still slightly different from the past rumors. For example, the 4nm process was not adopted, and Apple subsequently introduced the M2 chip in more detail.
The M2 chip is manufactured using the second-generation 5nm process (N5P) and consists of 20 billion transistors, 25% more than the M1 chip. Under the 8-core design consisting of 4 performance cores and 4 efficiency cores, the CPU performance is improved by 18%; the number of GPU cores is increased to 10, and the GPU performance is increased by 35%; the number of neural engine cores is 16, and the speed of the neural engine is increased by 40%. It supports LPDDR5 and provides a unified memory bandwidth of 100 GB/s, which is 50% more than the M1. With a larger unified memory capacity of 24GB, the M2 chip can handle more cumbersome and complex workloads.
Compared to N5, N5P mainly improves performance characteristics and does not increase density. While Apple doesn’t provide the size of the chip, the schematic shows that the M2 is slightly larger than the M1. The performance core of the M2 chip has a larger shared L2 cache, which is increased from 12MB of the M1 chip to 16MB. The M2 chip is also equipped with a new version of the video codec module, which adds support for ProRes and ProRes RAW codecs, and officially supports 8K video decoding.
