How to generate random numbers faster in R?

Question

My project requires generating a substantial amount (in the order of 10^7, for example) of random numbers that is normally distributed, and this step takes a significant amount of computational resources.

Is there a more efficient method than rnorm for generating such data?

Answer 1

Yes, you can use the Ziggurat method via RcppZiggurat . This is one of the faster/fastest algorithm for standard normally distributed random numbers.

The pdf vignette has timing comparison charts; there is also one on package page on my server .