Getting Started - nolanlab/Rclusterpp GitHub Wiki

Table of Contents Standard Installation Installation using MacPorts and g++4.8 (With OpenMP Support) Required Tools Compiling and Installing

Standard Installation

Simply start R and install directly from CRAN.

 R> install.packages("Rclusterpp")

All done! A this time, the binary version of Rclusterpp for Mac OS X hosted on CRAN does not support multithreaded operation. To compile with multithreaded support, see the next section.

Installation using MacPorts and g++4.8 (With OpenMP Support)

Required Tools

Version 2.15 or later of R. The 64-bit version of the program is recommended.
If installing from source (which you're probably doing if you're reading this page), you'll need a compiler installed.
- Linux users will most likely already have a complier installed.
- Windows users will need to install the Rtools package.
- For Mac OS X, XCode should work although there are some incompatibilities between the newly-released Xcode 5 and the pre-compiled version of R supplied by CRAN. In order to use the compiler supplied by any version of XCode, the command line compile tools must also be installed . To install them after installing XCode:
  - Launch XCode.
  - Choose "Preferences..." under the "XCode" Menu.
  - Select the "Downloads" section and then click on the "Components" bar.
  - Click the "Install" button for the Command Line Tools option.
  - NOTE: If Command Line Tools are not available in XCode, try running the following command in the terminal:

 xcode-select --install

Compiling and Installing

To enable Rclusterpp to take advantage of OpenMP multithreaded support, we must first install a compiler that supports OpenMP. To do this:

Download and install MacPorts

Open a terminal and from the command line, install gcc version 4.8:

 shell> sudo port install gcc48

Confirm gcc48 is active:

 shell> g++-mp-4.8 -v

Now tell R to use our newly installed version of g++. Open a terminal and use a text editor to create the ~/.R/Makevars file:

 shell> mkdir ~/.R
 shell> nano ~/.R/Makevars

Add the following lines to the text file:

 CXX=g++-mp-4.8 -arch x86_64
 CC=gcc-mp-4.8
 SHLIB_OPENMP_CXXFLAGS= -fopenmp

Write changes and exit the editor. If you're using nano, type control-o, hit "enter", and then control-x.

Start R and recompile the Rclusterpp and its dependancies using the newly installed compiler:

 R> install.packages("Matrix")
 R> install.packages(c("Rcpp","RcppEigen","Rclusterpp"),type="source")

Multithreaded support should now be enabled.

You may tell R to ignore the new compiler in the future by commenting out the added lines in the Makevars file.