First, install the Scalene extension from the VS Code Marketplace or by searching for it within VS Code by typing Command-Shift-X (Mac) or Ctrl-Shift-X (Windows). Once that's installed, click Command-Shift-P or Ctrl-Shift-P to open the Command Palette. Then select "Scalene: AI-powered profiling..." (you can start typing Scalene and it will pop up if it's installed). Run that and, assuming your code runs for at least a second, a Scalene profile will appear in a webview.

Scalene uses a verb-based command structure with two main commands: run (to profile) and view (to display results).

# Profile a program (saves to scalene-profile.json)
scalene run your_prog.py
python3 -m scalene run your_prog.py              # equivalent alternative

# View a profile
scalene view                                     # open profile in browser
scalene view --cli                               # view in terminal
scalene view --html                              # save to scalene-profile.html
scalene view --standalone                        # save as self-contained HTML

# Common profiling options
scalene run --cpu-only your_prog.py              # only profile CPU (faster)
scalene run -o results.json your_prog.py         # custom output filename
scalene run -c config.yaml your_prog.py          # load options from config file

# Pass arguments to your program (use --- separator)
scalene run your_prog.py --- --arg1 --arg2

# Get help
scalene --help                                   # main help
scalene run --help                               # profiling options
scalene run --help-advanced                      # advanced profiling options
scalene view --help                              # viewing options

You can store Scalene options in a YAML configuration file and load them with -c or --config:

scalene run -c scalene.yaml your_prog.py

Example scalene.yaml:

# Output options
outfile: my-profile.json

# Profiling mode (use only one)
cpu-only: true              # CPU profiling only (faster)
# gpu: true                 # Include GPU profiling
# memory: true              # Include memory profiling

# Filter what gets profiled
profile-only: "mypackage,mymodule"    # Only profile these paths
profile-exclude: "tests,venv"          # Exclude these paths
profile-all: false                     # Profile all code, not just target

# Performance tuning
cpu-percent-threshold: 1     # Min CPU% to report (default: 1)
cpu-sampling-rate: 0.01      # Sampling interval in seconds
malloc-threshold: 100        # Min allocations to report

# Other options
use-virtual-time: false      # Measure CPU time only (not I/O)
stacks: false                # Collect stack traces
memory-leak-detector: true   # Detect likely memory leaks

Command-line arguments override config file settings.

Invoke using scalene as above and then:

from scalene import scalene_profiler

# Turn profiling on
scalene_profiler.start()

# your code

# Turn profiling off
scalene_profiler.stop()

from scalene.scalene_profiler import enable_profiling

with enable_profiling():
    # do something

Just preface any functions you want to profile with the @profile decorator and run it with Scalene:

# do not import profile!

@profile
def slow_function():
    import time
    time.sleep(3)

% scalene --help
Scalene: a high-precision CPU and memory profiler, version 1.5.51 (2025.01.29)
https://github.com/plasma-umass/scalene

commands:
  run     Profile a Python program (saves to scalene-profile.json)
  view    View an existing profile in browser or terminal

examples:
  % scalene run your_program.py              # profile, save to scalene-profile.json
  % scalene view                             # view scalene-profile.json in browser
  % scalene view --cli                       # view profile in terminal

in Jupyter, line mode:
  %scrun [options] statement

in Jupyter, cell mode:
  %%scalene [options]
   your code here

% scalene run --help
Profile a Python program with Scalene.

examples:
  % scalene run prog.py                 # profile, save to scalene-profile.json
  % scalene run -o my.json prog.py      # save to custom file
  % scalene run --cpu-only prog.py      # profile CPU only (faster)
  % scalene run -c scalene.yaml prog.py # load options from config file
  % scalene run prog.py --- --arg       # pass args to program
  % scalene run --help-advanced         # show advanced options

options:
  -h, --help            show this help message and exit
  -o, --outfile OUTFILE output file (default: scalene-profile.json)
  --cpu-only            only profile CPU time (no memory/GPU)
  -c, --config FILE     load options from YAML config file
  --help-advanced       show advanced options

% scalene run --help-advanced
Advanced options for scalene run:

background profiling:
  Use --off to start with profiling disabled, then control it from another terminal:
    % scalene run --off prog.py          # start with profiling off
    % python3 -m scalene.profile --on  --pid <PID>   # resume profiling
    % python3 -m scalene.profile --off --pid <PID>   # suspend profiling

options:
  --profile-all         profile all code, not just the target program
  --profile-only PATH   only profile files containing these strings (comma-separated)
  --profile-exclude PATH exclude files containing these strings (comma-separated)
  --profile-system-libraries  profile Python stdlib and installed packages (default: skip)
  --gpu                 profile GPU time and memory
  --memory              profile memory usage
  --stacks              collect stack traces
  --profile-interval N  output profiles every N seconds (default: inf)
  --use-virtual-time    measure only CPU time, not I/O or blocking
  --cpu-percent-threshold N  only report lines with at least N% CPU (default: 1%)
  --cpu-sampling-rate N CPU sampling rate in seconds (default: 0.01)
  --allocation-sampling-window N  allocation sampling window in bytes
  --malloc-threshold N  only report lines with at least N allocations (default: 100)
  --program-path PATH   directory containing code to profile
  --memory-leak-detector  EXPERIMENTAL: report likely memory leaks
  --on                  start with profiling on (default)
  --off                 start with profiling off

% scalene view --help
View an existing Scalene profile.

examples:
  % scalene view                    # open in browser
  % scalene view --cli              # view in terminal
  % scalene view --html             # save to scalene-profile.html
  % scalene view --standalone       # save as self-contained HTML
  % scalene view myprofile.json     # open specific profile in browser

options:
  -h, --help     show this help message and exit
  --cli          display profile in the terminal
  --html         save to scalene-profile.html (no browser)
  --standalone   save as self-contained HTML with all assets embedded
  -r, --reduced  only show lines with activity (--cli mode)

This notebook illustrates the use of Scalene in Jupyter.

Installation:

!pip install scalene
%load_ext scalene

Line mode:

%scrun [options] statement

Cell mode:

%%scalene [options]
code...
code...

Scalene is distributed as a pip package and works on Mac OS X, Linux (including Ubuntu in Windows WSL2) and Windows platforms.

Note for Windows users

Starting with Scalene 2.0, Windows supports full memory profiling. If you encounter issues, ensure you have the Visual C++ Redistributable installed. If building from source, you will need Visual C++ Build Tools and CMake.

You can install it as follows:

  % pip install -U scalene

or

  % python3 -m pip install -U scalene

You may need to install some packages first.

See https://stackoverflow.com/a/19344978/4954434 for full instructions for all Linux flavors.

For Ubuntu/Debian:

  % sudo apt install git python3-all-dev

  % conda install -c conda-forge scalene

Scalene is distributed as a conda package and works on Mac OS X, Linux (including Ubuntu in Windows WSL2) and Windows platforms.

Note for Windows users

Starting with Scalene 2.0, Windows supports full memory profiling. If you encounter issues, ensure you have the Visual C++ Redistributable installed.

You can install Scalene on Arch Linux via the AUR package. Use your favorite AUR helper, or manually download the PKGBUILD and run makepkg -cirs to build. Note that this will place libscalene.so in /usr/lib; modify the below usage instructions accordingly.

A: Yes! You can run it as follows (for example):

scalene run -m pytest your_test.py

or

python3 -m scalene run -m pytest your_test.py

A: Yes! There are several options:

1. Use --reduced-profile to include only lines and files with memory/CPU/GPU activity.
2. Use --profile-only to include only filenames containing specific strings (as in, --profile-only foo,bar,baz).
3. Decorate functions of interest with @profile to have Scalene report only those functions.
4. Turn profiling on and off programmatically by importing Scalene profiler (from scalene import scalene_profiler) and then turning profiling on and off via scalene_profiler.start() and scalene_profiler.stop(). By default, Scalene runs with profiling on, so to delay profiling until desired, use the --off command-line option (scalene run --off yourprogram.py).

A: In PyCharm, you can run Scalene at the command line by opening the terminal at the bottom of the IDE and running a Scalene command (e.g., scalene run <your program>). Then use scalene view --html to generate an HTML file (scalene-profile.html) that you can view in the IDE.

A: Pass in the --noreload option (see https://github.com/plasma-umass/scalene/issues/178).

A: Yes! Put the following code in the beginning of your program, or modify the call to monkey.patch_all as below:

from gevent import monkey
monkey.patch_all(thread=False)

A: Scalene works with PyTorch version 1.5.1 on Mac OS X. There's a bug in newer versions of PyTorch (https://github.com/pytorch/pytorch/issues/57185) that interferes with Scalene (discussion here: https://github.com/plasma-umass/scalene/issues/110), but only on Macs.

@inproceedings{288540,
author = {Emery D. Berger and Sam Stern and Juan Altmayer Pizzorno},
title = {Triangulating Python Performance Issues with {S}calene},
booktitle = {{17th USENIX Symposium on Operating Systems Design and Implementation (OSDI 23)}},
year = {2023},
isbn = {978-1-939133-34-2},
address = {Boston, MA},
pages = {51--64},
url = {https://www.usenix.org/conference/osdi23/presentation/berger},
publisher = {USENIX Association},
month = jul
}