Python

VizTracer

Styles: Timeline

VizTracer is a simple Python package for timeline profiling. It’s profiling output can be visualised in a web browser with it’s sub-package VizViewer.

cProfile

Styles: Function-Level

cProfile is a function-level profiler that is part of Python’s standard library. It is a command-line profiler, that can also be called directly from Python. 3rd party packages such as snakeviz can be used to visualise it’s output in a web browser or Jupyter Notebook.

line_profiler

Styles: Line-Level

line_profiler is a Python package for line-level profiling. It is a command-line profiler, that can also be used within Jupyter Notebooks. Functions marked with the @profile decorator are profiled when line_profiler is enabled during execution.

Intel VTune

Styles: Function-Level, Line-Level, Memory, Hardware-Metrics

Intel VTune Profiler is a fully-featured profiler that supports a broad range of languages. Users wishing to perform function or line level profiling can use the default “Hotspots” configuration, that will output profiling results in an interactive GUI with several visualisation options to aide interpretation of CPU time within the profiled application.

There are also many more advanced profiling configurations available, covering memory and hardware metrics, these are unlikely to be useful for the typical programmer.

Scalene

Styles: Function-Level, Line-Level, Memory

Scalene is a modern Python profiler that supports a wide range of profiling styles. It provides both a command line interface and a web-based GUI for visualising results. Unlike many other Python profilers it uses a sampling approach to minimise the profiling overhead, making it suitable for profiling expensive use-cases. It has support for AI powered optimisation suggestions, however this requires running a local LLM or having an API key for a cloud LLM platform.