What’s our purpose? What factors confound the results of tests, and how to eliminate them? Which tests? What should we believe, and where do we go from here?
benchmarks
Before deciding on internal and external storage, you need to be realistic about the performance it will achieve. Here are the numbers – and a couple of things we tend to forget about.
Select a test, time it, and compare the result with those from other systems. Choose whether to use a synthetic or application benchmark, and don’t forget your confirmation bias.
Which is faster, a MacBook Pro 16-inch with an M1 Pro, or a Mac Studio with an M1 Max? Tests cover P and E cores, Neural Engine, SSD and more.
Just how fast is a Mac Studio with an M1 Ultra? It all depends on the factors discussed here, with details of M1 series CPU performance.
All disks cache data to be written, which makes benchmarking them tricky. It has more serious consequences which macOS tries to allow for in file systems and backups.
From the anatomy of the CPU cores, to the queues of threads in GCD, and assignment to a core cluster, this details how threads are managed for the M1 series chips.
Why can’t the taskpolicy command tool be used to promote software to be able to run on the M1 chip’s Performance cores? Does it change QoS?
An accessible summary of the CPU cores in M1 chips, and how they appear to be managed by macOS to get the best for different classes of process.
Results from running 10-70 identical compute-intensive processes on M1 chips shows the differences in strategy between background and high priority settings.