Demonstration of matrix multiplication being 10x faster when you make sure to access the array row-wise so that cpu cache hits are maximized. - View it on GitHub