Mac 安装LightGBM的GPU版

使用AppleClang版本安装

 

export CC=clang
export CXX=clang++
brew install boost  #boost 不可少 否则会报错
brew install libomp # 不可少
mkdir build;cd build
cmake -DUSE_GPU=1 ..
make -j4  # 4 cpu核心 一般默认是4

安装

cd ../python-package

python setup.py install  --gpu

代码测试 参考

https://github.com/Microsoft/LightGBM/tree/master/examples/python-guide

测试代码

import lightgbm as lgb
import time

dtrain = lgb.Dataset('/LightGBM/tests/data/categorical.data')
params = {'max_bin': 63,
          'num_leaves': 255,
          'learning_rate': 0.1,
          'tree_learner': 'serial',
          'task': 'train',
          'is_training_metric': 'false',
          'min_data_in_leaf': 1,
          'min_sum_hessian_in_leaf': 100,
          'ndcg_eval_at': [1, 3, 5, 10],
          'sparse_threshold': 1.0,
          'device': 'gpu',
          'gpu_platform_id': 1,
          'gpu_device_id': 0
          }
t0 = time.time()
gbm = lgb.train(params, train_set=dtrain, num_boost_round=10,
                valid_sets=None, valid_names=None,
                fobj=None, feval=None, init_model=None,
                feature_name='auto', categorical_feature='auto',
                early_stopping_rounds=None, evals_result=None,
                verbose_eval=True,
                keep_training_booster=False, callbacks=None)
t1 = time.time()

print('gpu version elapse time: {}'.format(t1-t0))

gpu版本运行结果

[LightGBM] [Info] This is the GPU trainer!!
[LightGBM] [Info] Total Bins 325
[LightGBM] [Info] Number of data: 7000, number of used features: 8

[LightGBM] [Info] Using GPU Device: GeForce GTX 1070, Vendor: NVIDIA
[LightGBM] [Info] Compiling OpenCL Kernel with 64 bins...
[LightGBM] [Info] GPU programs have been built
[LightGBM] [Info] Size of histogram bin entry: 12
[LightGBM] [Info] 8 dense feature groups (0.05 MB) transferred to GPU in 0.261116 secs. 0 sparse feature groups
[LightGBM] [Info] Start training from score 0.181286
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
gpu version elapse time: 8.268070697784424

cpu运行版本

params = {'max_bin': 63,
          'num_leaves': 255,
          'learning_rate': 0.1,
          'tree_learner': 'serial',
          'task': 'train',
          'is_training_metric': 'false',
          'min_data_in_leaf': 1,
          'min_sum_hessian_in_leaf': 100,
          'ndcg_eval_at': [1, 3, 5, 10],
          'sparse_threshold': 1.0,
          'device': 'cpu'
          }
print("*****************************")
t0 = time.time()
gbm = lgb.train(params, train_set=dtrain, num_boost_round=10,
                valid_sets=None, valid_names=None,
                fobj=None, feval=None, init_model=None,
                feature_name='auto', categorical_feature='auto',
                early_stopping_rounds=None, evals_result=None,
                verbose_eval=True,
                keep_training_booster=False, callbacks=None)
t1 = time.time()

print('cpu version elapse time: {}'.format(t1-t0))

cpu运行结果

[LightGBM] [Info] Total Bins 325
[LightGBM] [Info] Number of data: 7000, number of used features: 8
[LightGBM] [Info] Start training from score 0.181286
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
[LightGBM] [Warning] No further splits with positive gain, best gain: -inf
cpu version elapse time: 0.0941917896270752

gpu反而比cpu运行时间多 .....

可能是数据量太小导致的 

 

 

缺少boost库会提示
​
-- Found OpenCL: /System/Library/Frameworks/OpenCL.framework (found version "1.2")
-- OpenCL include directory:/System/Library/Frameworks/OpenCL.framework
CMake Error at /usr/local/Cellar/cmake/3.14.1/share/cmake/Modules/FindBoost.cmake:2132 (message):
  Unable to find the requested Boost libraries.
​
  Unable to find the Boost header files.  Please set BOOST_ROOT to the root
  directory containing Boost or BOOST_INCLUDEDIR to the directory containing
  Boost's headers.
Call Stack (most recent call first):
  CMakeLists.txt:100 (find_package)
​
​
CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
Boost_INCLUDE_DIR (ADVANCED)
   used as include directory in directory /Users/yh/Downloads/LightGBM
   used as include directory in directory /Users/yh/Downloads/LightGBM
   used as include directory in directory /Users/yh/Downloads/LightGBM
   used as include directory in directory /Users/yh/Downloads/LightGBM
   used as include directory in directory /Users/yh/Downloads/LightGBM
   used as include directory in directory /Users/yh/Downloads/LightGBM
   used as include directory in directory /Users/yh/Downloads/LightGBM
   used as include directory in directory /Users/yh/Downloads/LightGBM
​
-- Configuring incomplete, errors occurred!
See also "/Users/yh/Downloads/LightGBM/build/CMakeFiles/CMakeOutput.log".
See also "/Users/yh/Downloads/LightGBM/build/CMakeFiles/CMakeError.log".
​

部分参考链接

https://www.jianshu.com/p/30555fd2bd50