likwid

Supports a wide range of Intel, AMD, ARM, and POWER CPUs, plus Nvidia and AMD GPUs, as detailed in the extensive architecture list in the README.

Offers multiple command-line tools for topology discovery, performance counters, power monitoring, thread pinning, and MPI support, covering diverse performance analysis needs.

Emphasizes simplicity with easy installation and command-line tools, making deep hardware analysis approachable without complex dependencies, as per the project philosophy.

Provides in-depth hardware performance counters, power consumption data, and micro-benchmarking results, enabling precise code optimization, shown in usage examples like likwid-perfctr.