Evaluating the performance of CPUs with identical cores is relatively straightforward, and they’re easy to compare using single- and multi-core benchmarks. When there are two different types of core, one designed primarily for energy efficiency (E), the other for maximum performance (P), traditional benchmarks can readily mislead. Multi-core results are dominated by the ratio of P to E cores, and variable frequency confounds further. In this series of articles, I set out to disentangle these when comparing core performance between Apple’s original M1 Pro and its third-generation M3 Pro chips.
This first article explains why and how I am investigating this, and shows overall results for performance and power use under a range of loads.
Articles like these will help you make an informed decision about whether or not your workloads can benefit from moving from an M1/M2 to an M3.