LightGBM（Light Gradient Boosting Machine）是一个基于决策树算法的快速的、分布式的、高性能 gradient boosting（GBDT、GBRT、GBM 或 MART）框架，可被用于排行、分类以及其他许多机器学习任务中。
开源项目地址： https://github.com/Microsoft/LightGBM

开源|LightGBM：三天内收获GitHub 1000+ Star，表现超越已有boosting工具。

LightGBM

LightGBM 是一个梯度 boosting 框架，使用基于学习算法的决策树。它可以说是分布式的，高效的，它有以下优势：
更快的训练效率
低内存使用
更好的准确率
支持并行学习
可处理大规模数据

环境描述

操作系统：mac os Sierra 10.12

iMac (21.5-inch, Late 2015)

编译过程

lightgbm 依赖 OpenMP 进行编译，苹果的 Clang 不支持。请使用 GCC / G++ 代替。参考官方给出的安装指南：

brew install cmake
brew install gcc --without-multilib
git clone --recursive https://github.com/Microsoft/LightGBM ; cd LightGBM
mkdir build ; cd build
cmake -DCMAKE_CXX_COMPILER=g++-6 -DCMAKE_C_COMPILER=gcc-6 .. 
make -j 

以下为我逐条执行的输出：1、使用 brew 安装 cmake，CentOS下使用 yum install cmake,Ubuntu下使用apt-get install cmake

$ brew install cmake
Updating Homebrew...
==> Downloading https://homebrew.bintray.com/bottles/cmake-3.7.2.sierra.bottle.tar.gz
######################################################################## 100.0%
==> Pouring cmake-3.7.2.sierra.bottle.tar.gz
==> Caveats
Emacs Lisp files have been installed to:
  /usr/local/share/emacs/site-lisp/cmake
==> Summary
  /usr/local/Cellar/cmake/3.7.2: 2,143 files, 29MB
avenMac:conf aven$ cmake
Usage


  cmake [options] <path-to-source>
  cmake [options] <path-to-existing-build>


Specify a source directory to (re-)generate a build system for it in the
current working directory.  Specify an existing build directory to
re-generate its build system.


Run 'cmake --help' for more information.

2、安装gcc，CentOS下使用 yum install gcc,Ubuntu下使用apt-get install gcc，可能还需要安装g++

$ brew install gcc --without-multilib
Updating Homebrew...
==> Auto-updated Homebrew!
Updated 1 tap (homebrew/core).
==> Updated Formulae
aws-sdk-cpp              docker                   go                       pari                     qjson                    tarsnap-gui
awscli                   dspdfviewer              godep                    pdf2htmlex               qscintilla2              tomcat@7
certbot                  flow                     knot-resolver            percona-server           qtads                    transcrypt
chapel                   gammaray                 kobalt                   pngquant                 quazip                   urh
cockatrice               gearboy                  liblastfm                pushpin                  qwt                      yarn
codequery                gearsystem               logtalk                  qbs                      qwtpolar                 zurl
color-code               git-cola                 mono                     qca                      qxmpp
dbus                     gitlab-ci-multi-runner   nano                     qcachegrind              rabbitmq
djview4                  gnupg                    paket                    qjackctl                 riemann
==> Renamed Formulae
pyqt5 -> pyqt                                                               qt5 -> qt


==> Installing dependencies for gcc: gmp, mpfr, libmpc, isl
==> Installing gcc dependency: gmp
==> Downloading https://homebrew.bintray.com/bottles/gmp-6.1.2.sierra.bottle.1.tar.gz
######################################################################## 100.0%
==> Pouring gmp-6.1.2.sierra.bottle.1.tar.gz
  /usr/local/Cellar/gmp/6.1.2: 18 files, 3.1MB
==> Installing gcc dependency: mpfr
==> Downloading https://homebrew.bintray.com/bottles/mpfr-3.1.5.sierra.bottle.tar.gz
######################################################################## 100.0%
==> Pouring mpfr-3.1.5.sierra.bottle.tar.gz
  /usr/local/Cellar/mpfr/3.1.5: 25 files, 3.6MB
==> Installing gcc dependency: libmpc
==> Downloading https://homebrew.bintray.com/bottles/libmpc-1.0.3.sierra.bottle.tar.gz
######################################################################## 100.0%
==> Pouring libmpc-1.0.3.sierra.bottle.tar.gz
  /usr/local/Cellar/libmpc/1.0.3: 11 files, 345.6KB
==> Installing gcc dependency: isl
==> Downloading https://homebrew.bintray.com/bottles/isl-0.18.sierra.bottle.tar.gz
######################################################################## 100.0%
==> Pouring isl-0.18.sierra.bottle.tar.gz
  /usr/local/Cellar/isl/0.18: 80 files, 3.8MB
==> Installing gcc --without-multilib
==> Using the sandbox
==> Downloading https://ftpmirror.gnu.org/gcc/gcc-6.3.0/gcc-6.3.0.tar.bz2
==> Downloading from http://mirrors.ustc.edu.cn/gnu/gcc/gcc-6.3.0/gcc-6.3.0.tar.bz2
######################################################################## 100.0%
==> Downloading https://raw.githubusercontent.com/Homebrew/formula-patches/e9e0ee09389a54cc4c8fe1c24ebca3cd765ed0ba/gcc/6.1.0-jit.patch
######################################################################## 100.0%
==> Patching
==> Applying 6.1.0-jit.patch
patching file gcc/jit/Make-lang.in
==> ../configure --build=x86_64-apple-darwin16.4.0 --prefix=/usr/local/Cellar/gcc/6.3.0_1 --libdir=/usr/local/Cellar/gcc/6.3.0_1/lib/gcc/6 --enable-la
==> make bootstrap
==> make install
  /usr/local/Cellar/gcc/6.3.0_1: 1,358 files, 237.8MB, built in 63 minutes 14 seconds

3、克隆项目代码

$ git clone --recursive https://github.com/Microsoft/LightGBM ; cd LightGBM
Cloning into 'LightGBM'...
remote: Counting objects: 5315, done.
remote: Compressing objects: 100% (31/31), done.
remote: Total 5315 (delta 5), reused 2 (delta 2), pack-reused 5282
Receiving objects: 100% (5315/5315), 4.06 MiB | 21.00 KiB/s, done.
Resolving deltas: 100% (3644/3644), done.
Submodule 'include/boost/compute' (https://github.com/boostorg/compute) registered for path 'compute'
Cloning into '/Users/aven/software/LightGBM/compute'...
Submodule path 'compute': checked out '1380a04582080bbe2364352b336270bc4bfa3025'
avenMac:LightGBM aven$ pwd
/Users/aven/software/LightGBM

4、创建构建目录

mkdir build ; cd build

5、执行 cmake 

$ cmake -DCMAKE_CXX_COMPILER=g++-6 -DCMAKE_C_COMPILER=gcc-6 .. 
-- The C compiler identification is GNU 6.3.0
-- The CXX compiler identification is GNU 6.3.0
-- Checking whether C compiler has -isysroot
-- Checking whether C compiler has -isysroot - yes
-- Checking whether C compiler supports OSX deployment target flag
-- Checking whether C compiler supports OSX deployment target flag - yes
-- Check for working C compiler: /usr/local/bin/gcc-6
-- Check for working C compiler: /usr/local/bin/gcc-6 -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Checking whether CXX compiler has -isysroot
-- Checking whether CXX compiler has -isysroot - yes
-- Checking whether CXX compiler supports OSX deployment target flag
-- Checking whether CXX compiler supports OSX deployment target flag - yes
-- Check for working CXX compiler: /usr/local/bin/g++-6
-- Check for working CXX compiler: /usr/local/bin/g++-6 -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Try OpenMP C flag = [-fopenmp]
-- Performing Test OpenMP_FLAG_DETECTED
-- Performing Test OpenMP_FLAG_DETECTED - Success
-- Try OpenMP CXX flag = [-fopenmp]
-- Performing Test OpenMP_FLAG_DETECTED
-- Performing Test OpenMP_FLAG_DETECTED - Success
-- Found OpenMP: -fopenmp  
-- Configuring done
CMake Warning (dev):
  Policy CMP0042 is not set: MACOSX_RPATH is enabled by default.  Run "cmake
  --help-policy CMP0042" for policy details.  Use the cmake_policy command to
  set the policy and suppress this warning.


  MACOSX_RPATH is not specified for the following targets:


   _lightgbm


This warning is for project developers.  Use -Wno-dev to suppress it.


-- Generating done
-- Build files have been written to: /Users/aven/software/LightGBM/build

6、构建

$ make -j 
Scanning dependencies of target _lightgbm
Scanning dependencies of target lightgbm
[  2%] Building CXX object CMakeFiles/_lightgbm.dir/src/c_api.cpp.o
[  6%] Building CXX object CMakeFiles/_lightgbm.dir/src/application/application.cpp.o
[  8%] Building CXX object CMakeFiles/lightgbm.dir/src/main.cpp.o
[  8%] Building CXX object CMakeFiles/_lightgbm.dir/src/boosting/boosting.cpp.o
[ 10%] Building CXX object CMakeFiles/lightgbm.dir/src/boosting/boosting.cpp.o
[ 14%] Building CXX object CMakeFiles/lightgbm.dir/src/application/application.cpp.o
[ 16%] Building CXX object CMakeFiles/_lightgbm.dir/src/boosting/gbdt.cpp.o
[ 18%] Building CXX object CMakeFiles/lightgbm.dir/src/boosting/gbdt.cpp.o
[ 20%] Building CXX object CMakeFiles/lightgbm.dir/src/io/dataset.cpp.o
[ 20%] Building CXX object CMakeFiles/lightgbm.dir/src/io/bin.cpp.o
[ 22%] Building CXX object CMakeFiles/_lightgbm.dir/src/io/bin.cpp.o
[ 24%] Building CXX object CMakeFiles/_lightgbm.dir/src/io/config.cpp.o
[ 30%] Building CXX object CMakeFiles/_lightgbm.dir/src/io/dataset.cpp.o
[ 30%] Building CXX object CMakeFiles/lightgbm.dir/src/io/config.cpp.o
[ 30%] Building CXX object CMakeFiles/lightgbm.dir/src/io/dataset_loader.cpp.o
[ 32%] Building CXX object CMakeFiles/_lightgbm.dir/src/io/dataset_loader.cpp.o
[ 38%] Building CXX object CMakeFiles/_lightgbm.dir/src/io/metadata.cpp.o
[ 40%] Building CXX object CMakeFiles/_lightgbm.dir/src/metric/dcg_calculator.cpp.o
[ 40%] Building CXX object CMakeFiles/_lightgbm.dir/src/io/parser.cpp.o
[ 40%] Building CXX object CMakeFiles/_lightgbm.dir/src/io/tree.cpp.o
[ 44%] Building CXX object CMakeFiles/lightgbm.dir/src/io/tree.cpp.o
[ 46%] Building CXX object CMakeFiles/_lightgbm.dir/src/network/linkers_mpi.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm.dir/src/metric/metric.cpp.o
[ 58%] Building CXX object CMakeFiles/_lightgbm.dir/src/network/linkers_socket.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm.dir/src/network/network.cpp.o
[ 62%] Building CXX object CMakeFiles/lightgbm.dir/src/io/parser.cpp.o
[ 72%] Building CXX object CMakeFiles/_lightgbm.dir/src/metric/metric.cpp.o
[ 72%] Building CXX object CMakeFiles/_lightgbm.dir/src/network/network.cpp.o
[ 72%] Building CXX object CMakeFiles/_lightgbm.dir/src/objective/objective_function.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm.dir/src/network/linkers_mpi.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm.dir/src/objective/objective_function.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm.dir/src/metric/dcg_calculator.cpp.o
[ 72%] Building CXX object CMakeFiles/_lightgbm.dir/src/network/linker_topo.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm.dir/src/network/linker_topo.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm.dir/src/io/metadata.cpp.o
[ 72%] Building CXX object CMakeFiles/lightgbm.dir/src/network/linkers_socket.cpp.o
[ 74%] Building CXX object CMakeFiles/_lightgbm.dir/src/treelearner/data_parallel_tree_learner.cpp.o
[ 76%] Building CXX object CMakeFiles/_lightgbm.dir/src/treelearner/serial_tree_learner.cpp.o
[ 78%] Building CXX object CMakeFiles/_lightgbm.dir/src/treelearner/feature_parallel_tree_learner.cpp.o
[ 80%] Building CXX object CMakeFiles/_lightgbm.dir/src/treelearner/gpu_tree_learner.cpp.o
[ 82%] Building CXX object CMakeFiles/lightgbm.dir/src/treelearner/voting_parallel_tree_learner.cpp.o
[ 84%] Building CXX object CMakeFiles/_lightgbm.dir/src/treelearner/tree_learner.cpp.o
[ 86%] Building CXX object CMakeFiles/lightgbm.dir/src/treelearner/serial_tree_learner.cpp.o
[ 88%] Building CXX object CMakeFiles/lightgbm.dir/src/treelearner/tree_learner.cpp.o
[ 94%] Building CXX object CMakeFiles/lightgbm.dir/src/treelearner/gpu_tree_learner.cpp.o
[ 94%] Building CXX object CMakeFiles/lightgbm.dir/src/treelearner/data_parallel_tree_learner.cpp.o
[ 94%] Building CXX object CMakeFiles/lightgbm.dir/src/treelearner/feature_parallel_tree_learner.cpp.o
[ 96%] Building CXX object CMakeFiles/_lightgbm.dir/src/treelearner/voting_parallel_tree_learner.cpp.o
[100%] Linking CXX shared library ../lib_lightgbm.so
[100%] Linking CXX executable ../lightgbm
[100%] Built target lightgbm
[100%] Built target _lightgbm

执行到 96% 的时候会卡一会儿，如果不希望执行“make -j”命令时对电脑影响太大的话，执行“make -j1”，使用CPU的一核，不会占满CPU。

构建成功后会在上层目录生成两个文件，lib_lightgbm.so  lightgbm，一个库文件，一个可执行文件。

相关阅读：

微软开源分布式高性能GB框架LightGBM安装使用

微软开源分布式高性能GB框架LightGBM安装使用——Python

====================文档信息=======================

版权声明：非商用自由转载-保持署名-注明出处

署名(BY) ：testcs_dn(微wx笑)

文章出处：[无知人生，记录点滴](http://blog.csdn.NET/testcs_dn)

==============欢迎关注我的个人微信订阅号（微wx笑）==========