What is the purpose of timing matrix multiplication for different matrix sizes, and how is this data organized in the resulting DataFrame?