Articles

• https://pythonfiles.wordpress.com/2017/06/01/hunting-performance-in-python-code-part-3/

Profiling CPU

With cProfile

Collect data:

python -m cProfile -o out.cprof script.py args...

Use snakeviz to visualize collected data:

pip install snakeviz

snakeviz out.cprof

Alternatives to snakeviz:

pip install pyprof2calltree
apt install qcachegrind

pyprof2calltree -k -i out.cprof

With pprofile

https://github.com/vpelletier/pprofile

With line_profiler

https://github.com/rkern/line_profiler

Profiling memory

With memory_profiler

Global memory usage

Note: use the --python python3 option as indicated in this issue.

mprof run --python python3 convert.py source-data json-data
mprof plot

Optionally decorate some functions with @profile, without importing the decorator, as explained in the README. You'll see the memory usage of those functions explicitly on the chart.

Line by line

Decorate the functions you want to profile with @profile and run with:

python -m memory_profiler convert.py source-data json-data

Misc

• https://github.com/nvdv/vprof