by Guangming Lang
1 min read


  • r

XGBoost is probably the most cutting-edge and optimized implementation of the gradient boosting algorithms. If you’re a R user, you probably used GBM in the past. XGBoost is generally over 10 times faster than GBM. Unfortunately, its official installation document for R users on a Mac OS isn’t clear. In this post, I’m going to walk you through how to install the multi-thread version of the XGBoost R package on a Mac OS X El Capitan. You want the multi-thread version because you want to have the option of using more than 1 core of your machine simultaneously to run the algorithms in parallel.

Step 1. Open up your terminal and run the following cmd.

  1. install gcc-6.x.x with openmp. This can take a while (~ 30 minutes).

brew install gcc --without-multilib

  1. clone the repository.

git clone --recursive

  1. build xgboost.

cd xgboost; cp make/ ./; make -j4

  1. make an empty Makevars file and open it.
mkdir ~/.R
touch ~/.R/Makevars
open -a TextEdit ~/.R/Makevars
  1. paste the following into the Makevars file
CC = gcc-6
CXX = g++-6
CXX1X = g++-6

  1. save and close the file.

Step 2. Open Rstudio and install the xgboost package.


Done! Now you can start with some tutorials.