Power use in two in-core performance tests, by number of threads run, leading to estimates of total energy used by P and E cores running the same code, at high frequencies. How efficient are the CPU cores in the M4?
ARM
From the first 8 MHz Motorola 68000, through PowerPCs reaching 2.5 GHz and more in up to 4 cores, and Intel x86 with up to 28 cores, to Apple’s M4 Max with 12 P cores at 4.5 GHz.
In-core performance compared across P and E cores in M1, M3 and M4 chips shows substantial performance improvements, particularly in vector and matrix computation.
Less glamorous than the P cores, E cores are used to run background threads. Details of their architecture, how threads are managed on them and their efficiency.
macOS virtual machines are preferentially run on P cores. Details on their performance, core allocation, frequencies and power use/
Details of their frequency, ISA, power use, and how macOS allocates threads to P cores and relocates them. Supported by data from an M4 Pro.
When running on M3 hosts, macOS VMs lack support for some of the instruction set, and Accelerate commands may be much slower. Why?
Virtual CPU cores are of one type, and QoS has no effect in virtualised macOS. This has consequences for both the host and guest macOS.
Many apps could benefit users of Apple silicon Macs by giving them controls over core use by their threads. Here’s how that can be done simply and effectively.
How you can use the taskpolicy command to confine all the threads of a process to the E cores, as a brake, but there’s no accelerator in macOS.
The Eclectic Light Company 