A look at some of the factors determining the performance of new M3 chips compared to the M1 and M2 versions, and how not to compare them.
cores
From the 8 cores of the original M1 chip to the M1 Ultra’s 20, this is how macOS manages threads from apps, services and other code.
There are two situations when M1 chips confine code to just one of their Performance cores: during startup, before the other cores are running, and when preparing an update.
In Activity Monitor, % CPU isn’t on a scale of 0-100. In M1 Macs, it also makes no distinction between E and P cores, nor does it allow for their changing frequency.
Apple’s first quad-core chip, the A10 launched 5 years ago, had 2 P and 2 E cores, but could only run one type of core at a time. We’ve come a long way since then.
We commonly suffer failures of the psychology of Mac performance, when system background tasks overwhelm the processor and bring the interface to a grinding halt
What you can learn about the processes running on your Mac, its processor cores, even the files which an app has open.
By segregating macOS background tasks on Efficiency cores, M1 Macs can run user apps unfettered on their Performance cores. And that feels really fast.
How the M1’s asymmetric cores can run background tasks more efficiently, or deliver high performance, according to Quality of Service.
Trying to get a more coherent overview of the different paths to concurrent, asynchronous, and background task management in macOS. Draft for discussion.
The Eclectic Light Company 