R | Calling C++ from R

One way to significantly improve the performance of R code is to replace the most frequently used and/or time-consuming parts of the code (determined via performance profiling [1]) with equivalent C++ code, which is typically much faster.

It is very easy to use C++ code in R, which has a couple of packages to facilitate this: Rcpp [2] and cpp11 [3]. These two largely do the same job (though they can be used side-by-side). The main differences are:

• Rcpp is older and more established, and has more advanced features than cpp11, which is more streamlined

• cpp11 has support for the ALTREP framework [4], which Rcpp does not

More information on how to use these, see [5] and [6] respectively.

Resources:

• [1] R | Code Performance Profiling

• [2] https://cran.r-project.org/package=Rcpp

• [3] https://cran.r-project.org/package=cpp11

• [4] R | ALTREP Framework

• [5] Rcpp usage: https://adv-r.hadley.nz/rcpp.html

• [6] cpp11 usage: https://cran.r-project.org/web/packages/cpp11/vignettes/cpp11.html