Comparing the speed of various operations

• To calculate a sum of squares, sum(x^2) or sum(x*x) is much faster than t(x)%*%x.
• rowSums() and colSums() are fast ways to sum over rows or columns.
• Try to use the standard arithmetic operations, +,-,*,/ whenever possible to do matrix-vector operations. For example, if D is a diagonal matrix, then $D^{-1}C$ can be done in R as C/diagD where diagD is a vector of the diagonal elements of D. This is much faster than doing matrix multiplication using %*%.

Last modified 12/28/07