Topics / #performance-analysis

Performance-analysis projects on GitHub

Profilers and analyzers measuring application speed and resource usage.

As of , the most-starred Performance-analysis projects on GitHub are lighthouse (30.9K stars), hyperfine (29.1K stars) and tracy (17K stars). GitHub.Locker tracks 7 Performance-analysis repositories with 132.4K stars combined. The fastest-growing on October 17 was hyperfine, with +10 stars in a day.

7Projects
132.4KStars in total
+28Stars in a day

Hot right now

Most stars gained in a day · Oct 17

The all-time top

7 projects · by stars

  1. #1 · GoogleChrome/lighthouse

    lighthouse

    Lighthouse is an open-source web auditing tool from Google: it runs a page through automated checks, collects modern performance metrics and best-practice insights, and turns them into a scored report.

    30.9K
  2. #2 · sharkdp/hyperfine

    hyperfine

    hyperfine is a command-line benchmarking tool: it times shell commands over many runs and gives you solid statistics — mean, min, max and relative speed — instead of a single noisy number.

    29.1K
  3. #3 · wolfpld/tracy

    tracy

    Tracy is a real-time, nanosecond-resolution frame and sampling profiler for games and other applications, built on remote telemetry so you can profile a running program from a separate viewer.

    17K
  1. 4

    py-spy

    15.6Kbenfred/py-spy · Rust

    py-spy is a sampling profiler for Python programs: it shows what a running Python process is spending its time on, without restarting it or modifying any code.

  2. 5

    memray

    15.4Kbloomberg/memray · Python

    Memray is a memory profiler for Python. It tracks memory allocations in Python code, native extension modules and the interpreter itself, then builds reports like flame graphs from the captured data.

  3. 6

    scalene

    13.5Kplasma-umass/scalene · Python

    Scalene is a high-performance CPU, GPU, and memory profiler for Python. It is also the first profiler to add AI-powered optimization proposals, via providers like OpenAI, Azure, Bedrock, or local Ollama models.

  4. 7

    go-perfbook

    10.9Kdgryski/go-perfbook

    It is a free, work-in-progress book documenting best practices for writing high-performance Go code, covering both general optimization techniques and Go-specific ones.

Often tagged together