Link to article

Project to analyze execution time via different ways in python

• Normal python file execution
• Building python package and running by importing it
• Compiling it with cython and running by importing it

Project package at https://test.pypi.org/project/py-performance-poc/

pip install -i https://test.pypi.org/simple/ py-performance-poc==0.0.3

python3 -m build
python3 -m twine upload --repository testpypi dist/*

• Create a folder, and paste your whole code in file_name.pyx
• create setup.py and copy the contents similar from setup.py in the project
• Building cython file

python3 setup.py build_ext --inplace

• Execute it by importing via your console

• Create virtual environment which will be needed for installing python Follow this if you are new
• Activate the virtual environment
• Create a folder <your_folder_name>
• create am empty init.py file
• create a <your_file_name>.py file
• Paste sample content into it or create your own conent for computations
• create a <test_your_file_name>.py file.
• Import your python class in the test file and call the method for computations.
• Run above multiple times and record your execution times.

python path_to_your_file/test_your_file_name.py

• Now lets start cython programming
• Create <your_cython_file_name>.pyx and paste the contents of the <your_file_name>.py file
• Create setup.py and add below code in it.

from setuptools import setup, Extension
from Cython.Build import cythonize

extensions = [Extension("<your_cython_file_name>", ["<your_cython_file_name>.pyx"])]

setup(
    ext_modules = cythonize(extensions)
)

• Now build your cython file

python3 setup.py build_ext --inplace

• After successful build, you will see one new build folder, one new <your_cython_file_name>.c and one new <your_cython_file_name.cypthon...>.so file
• Now create <test_your_cython_file_name>.py
• Import your cython class in the test file and call the method for computations.
• Run above multiple times and record your execution times. Run in same machine and in similar conditions.

python path_to_your_file/test_your_cython_file_name.py

• file_name.pyc
• compiled_file_name.so
• You only need your .so and .pyx file, copy that in your project and start using by importing it.

source <path_to_project>/.venv/bin/activate
python <path_to_project>/test_performance.py
python <path_to_project>/test_cython_performance.py

from performance import Performance
my_test_performance = Performance()
my_test_performance.some_random_computations()

from cython_performance import Performance

my_test_performance = Performance()
my_test_performance.some_random_computations()

• Converting same piece of code into python package, did not made a difference in execution time
• Comiling with cython same piece of code, executed faster than normal python compiler
• Execution in cython seems faster, needs more methods to test it out

• test_performance.py took on average of 35.367467 seconds
• test_cython_performance.py took on average of 26.993737 seconds
• Cython code executed 23.67636% faster on average.

• Please add your methods and raise PR to compare execution time via different ways.