This document presents a framework for practically implementing parallel fast matrix multiplication algorithms. It compares the performance of several Strassen-like algorithms to the Intel MKL library on sequential and parallel systems with up to 24 cores. Certain Strassen-like algorithms like <4,2,4> and <3,2,3> generally outperform MKL for sequential and small parallel problems, while MKL performs best for large parallel problems. Different parallelization strategies like depth-first, breadth-first, and hybrid are explored, with hybrid providing better load balancing.