BGMP / MatrixMultiplier

Matrix multiplication program written in Python, with asymptotic visualization in R

MatrixMultiplier

This project aims to compare three different algorithms for matrix multiplication by running them on different sizes of matrices and comparing their execution times. The three algorithms are:

• Naive (Traditional).
• Divide and conquer.
• Strassen's algorithm.

The project is written in Python and uses the numpy library for matrix operations.

Running

At the top of main.py, you will find the following:

MATRIX_SIZE = [32, 64, 128, 256, 512, 1024] # n

You may modify the n values to change the size of the matrices to be multiplied. The program will generate two random matrices of size n x n for each of the values you provide, and then multiply them using the three algorithms mentioned above.

To run the program, first install the dependencies:

pip install -r requirements.txt

Then run the main script:

python main.py

The results are saved to a CSV file in the data/ directory, which are then used to generate graphs with R. If you want to generate the graphs, simply open MatrixMultiplier.Rproj with RStudio and run the script. The generated graphs will look as follows:

Dependencies

Python

• Python 3.8 or above
• Python modules:
  • numpy
• I suggest using venv & pip for development & dependency management. (In case of Apple Silicon use conda)

R

• R 4.0 or above

Notes

This project was written as part of the Masters in Computer Science course "Advanced Data Structures and Algorithms" at the University of the Bío-Bío.

Some of the TeX/PDF files are in Spanish, as the course was taught in Spanish, but I've made an effort to translate the relevant parts to english.