Honors thesis - Automating the Search for Faster Matrix Multiplication Algorithms

For my senior project I have had the good fortune to work with Virginia Vassilevska Williams and Ryan Williams on research in matrix multiplication algorithms, leading to a thesis for CS departmental honors. I implemented the search program described in Virginia Williams' paper "Multiplying Matrices Faster Than Coppersmith-Winograd" (STOC 2012) to look for better theoretical bounds on the exponent of matrix multiplication, and I was able to confirm bounds nearly matching Williams' current result of 2.3727.

Source code is available below in static archive form; to clone the git repository, see the instructions page.