Check out Tracy[1]. If you run it as root, it provides a lot of "extra" information, such as when your threads get moved between CPUs. Actually, I saw this post

Check out Tracy[1]. If you run it as root, it provides a lot of "extra" information, such as when your threads get moved between CPUs. Actually, I saw this post and thought "why should I bother when I already have Tracy?" If anyone has an answer to that, I'm curious to know (: Tracy is still a frame-based profiler, though. If you want general system-wide profiling more focused on throughput rather than latency, then I've had a good experience generating flame graphs[2] using plain Linux perf. [1] https://github.com/wolfpld/tracy https://github.com/wolfpld/tracy [2] https://www.brendangregg.com/FlameGraphs/cpuflamegraphs.html https://www.brendangregg.com/FlameGraphs/cpuflamegraphs.html