Overhead cost of a function call

October 1, 2011
By

(This article was first published on R HEAD, and kindly contributed to R-bloggers)

Recently, I would like to apply unit testing method to my R program. The first thing i need to chop every few lines of the code into functions so that I can test each of them.

A Question comes up to my mind: What is the overhead cost of a function call? To answer this question, i wrote the following :

library(rbenchmark)
library(compiler)

f<-function(x,y){
    x+y
}
g<-function(x,y){
   f(x,y)
}

cmpf<-cmpfun(f)
cmpg<-cmpfun(g)

benchmark(1+2,f(1,2),g(1,2),cmpf(1,2),cmpg(1,2),cmpg2(1,2), replications =1000000, columns = c("test", "replications", "elapsed", "relative"),order='relative')

          test replications elapsed relative
1       1 + 2      1000000    4.00    1.000
4  cmpf(1, 2)      1000000    4.34    1.085
2     f(1, 2)      1000000    4.82    1.205
5  cmpg(1, 2)      1000000    5.44    1.360
3     g(1, 2)      1000000    5.68    1.420

The result suggests several things

  1. The overhead cost is about 0.82 second for 1,000,000 times function call.
  2. If we compile the function, the overhead cost is about 0.34 second for 1,000,000 times function call.

I don’t know whether it is a huge cost, but I believe the benefit of cleaner writing code with unit testing must worth more than that!


To leave a comment for the author, please follow the link and comment on his blog: R HEAD.

R-bloggers.com offers daily e-mail updates about R news and tutorials on topics such as: visualization (ggplot2, Boxplots, maps, animation), programming (RStudio, Sweave, LaTeX, SQL, Eclipse, git, hadoop, Web Scraping) statistics (regression, PCA, time series, trading) and more...



If you got this far, why not subscribe for updates from the site? Choose your flavor: e-mail, twitter, RSS, or facebook...

Comments are closed.