Working with C/C++ Code in R

• Recording
• Github repo with notes, code, etc.

Description

In this workshop, we’ll discuss how to call C/C++ code (and hence other languages) from R.
This allows us to reuse existing C/C++ code, or to write our own C/C++ code to improve the speed
of some computations.

We’ll cover the .C(), .Call() interfaces, and using the Rcpp framework,
and discuss some of their trade-offs and pitfalls.

Importantly, we’ll present:

+ a high-level introduction to the C/C++ languages (limited to interfacing to R)
+ the steps involved in “compiling” code (compiling, linking, loading)
+ how to debug compiled code from within R
+ R’s computational model for working in C/C++
+ issues of portable code across operating systems,
+ how to enhance packages you write to customize themselves to a specific machine using
autoconf/cmake
+ a variety of tips and tricks that are often learned the hard, long way!

After this workshop participants should be able to:

+ compile, link and load existing C/C++ code into R
+ call a C routine from R
+ write a simple C routine and call it from R
+ debug C code called from R
+ understand garbage collection and PROTECT()/UNPROTECT() in C
+ understand the R’s C-level Application Programming Interface (API)

While this material is not conceptually complex, it is somewhat technical.
Being familiar with this material makes you a more advanced R user and gives you different
options for efficiently solving computational challenges.
The workshop will get you started quickly and answer questions about problems you might
have already encountered.

Prerequisites & Target Learners

This workshop is aimed at students, postdoctoral scholars, faculty, staff who are
comfortable working with R. You don’t need to know C or C++ to learn something from the workshop.

For those wanting to follow along on their computers, please ensure that you have
installed a compiler such as gcc or clang, and the Rcpp package.
On Windows, install Rtools.