http://lukewassink.com/tags/optimization/Recent content in Optimization on Luke WassinkHugoen-usThu, 03 Sep 2026 16:58:27 -0500http://lukewassink.com/posts/matrix-multiplication-optimization/Thu, 03 Sep 2026 16:58:27 -0500http://lukewassink.com/posts/matrix-multiplication-optimization/<h1 id="optimizing-matrix-multiplication"> Optimizing Matrix Multiplication <a class="heading-link" href="#optimizing-matrix-multiplication"> <i class="fa-solid fa-link" aria-hidden="true" title="Link to heading"></i> <span class="sr-only">Link to heading</span> </a> </h1> <p>Matrix multiplication is a great test case to understand performance optimization. The underlying algorithm is fairly simple, and it benefits significantly from some classic optimizations like:</p> <ul> <li>reducing array access</li> <li>sequential memory access</li> <li>cache locality</li> <li>parallelization</li> </ul> <p>There are also fancy algorithms that reduce the asymptotic complexity, but they are so complex they usually only show benefits for very large matrices, if at all. In any case, our plan will be to use the traditional algorithm for matrix multiplication and explore the benefits we can get from other changes like: how we store the matrices, what order we do the computations in, and running parts of the computation in parallel.</p>