Data Analysis with R, Second Edition - Second Edition

Using parallelization

As we saw in this chapter's introduction, one of the limitations of R (and most other programming languages) was that it was created before commodity personal computers had more than one processor or core. As a result, by default, R runs only one process and thus, makes use of one processor/core at a time.

If you have more than one core on your CPU, it means that when you leave your computer alone for a few hours during a long-running computation, your R task is running on one core while the others are idle. Clearly this is not ideal; if your R task took advantage of all the available processing power, you can get massive speed improvements.

Parallel computation (of the type we'll be using) works by starting multiple ...