Apple’s Latest SoC Expands on the Original
Apple has announced their M2 chip at WWDC 22, and this time it’s a new design – not a fusion of two existing chip designs as we saw with the M1 Ultra back in March, Apple’s fusion of two M1 Max SoCs. One thing that hasn’t changed is the technology node: M2 remains a 5nm product, though Apple is quick to point out that it uses “second-generation” 5nm tech.
“Built using second-generation 5-nanometer technology, M2 takes the industry-leading performance per watt of M1 even further with an 18 percent faster CPU, a 35 percent more powerful GPU, and a 40 percent faster Neural Engine.1 It also delivers 50 percent more memory bandwidth compared to M1, and up to 24GB of fast unified memory. M2 brings all of this — plus new custom technologies and greater efficiency — to the completely redesigned MacBook Air and updated 13-inch MacBook Pro.”
The M2 SoC may not be manufactured with a smaller process node, but it boasts 25% more transistors than the original M1, now up to 20 billion, and offers more (and faster) memory – now up to 24 GB unified memory with up to 100 GB/s of bandwidth.
Building an SoC with more transistors and memory – without a process shrink – requires more space.
Apple is also improving the GPU with the M2 SoC, as it will be available with up to 10 cores (up from a max of 8 cores).
“Combined with a larger cache and higher memory bandwidth, the 10-core GPU delivers a big boost in graphics performance, bringing up to 25 percent higher graphics performance than M1 at the same power level, and up to 35 percent better performance at its max power. Compared with the integrated graphics of the latest PC laptop chip, the GPU in M2 delivers 2.3x faster performance at the same power level, and matches its peak performance using a fifth of the power. The higher performance per watt from M2 enables systems to have exceptional battery life, and run cool and quietly, even when playing graphics-intensive games or editing massive RAW images.”
Finally, as this is an SoC and not simply a processor with graphics, Apple is able to boast about some impressive integrated technologies, with include (quoting Apple here):
- The Neural Engine can process up to 15.8 trillion operations per second — over 40 percent more than M1.
- The media engine includes a higher-bandwidth video decoder, supporting 8K H.264 and HEVC video.
- Apple’s powerful ProRes video engine enables playback of multiple streams of both 4K and 8K video.
- Apple’s latest Secure Enclave provides best-in-class security.
- A new image signal processor (ISP) delivers better image noise reduction.
As to availability, so far M2 versions have been announced for “the completely redesigned MacBook Air and updated 13-inch MacBook Pro”, which should both be available next month, but no word on the Mac mini or iMac just yet.
Apple’s M SOCs are starting to look pretty similar in design and scale to what AMD is doing with Sony and Microsofts custom console APUs and probably quickly approaching similar performance if given similar power budgets. The only catch with Apple hardware is how tightly integrated it is with their software and thats where more than a fair bit of the magic happens.
Still I wonder if Apple would ever do custom designs for third parties, even without the integrated technologies they have one of the best ARM designs on the market. They don’t need to for the money like AMD does (or did) but AMD got to develop a lot of high-performance IP through those deals with Sony and Microsoft. While a lot of that IP is exclusive to those customers not all of it is and you still get to keep all the ‘lessons learned’ at the end as well as the fat check.