add nccl (#920)

doc/fluid/beginners_guide/install/install_CentOS.md

@@ -23,7 +23,18 @@
     * *CUDA 工具包8.0配合cuDNN v5*
     * *GPU运算能力超过1.0的硬件设备*
 
-您可参考NVIDIA官方文档了解CUDA和CUDNN的安装流程和配置方法，请见[CUDA](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/)，[cuDNN](https://docs.nvidia.com/deeplearning/sdk/cudnn-install/)
+	您可参考NVIDIA官方文档了解CUDA和CUDNN的安装流程和配置方法，请见[CUDA](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/)，[cuDNN](https://docs.nvidia.com/deeplearning/sdk/cudnn-install/)
+
+
+* 请确保您已经正确安装nccl2，或者按照以下指令安装nccl2（这里提供的是CentOS 7，CUDA9，cuDNN7下nccl2的安装指令），更多版本的安装信息请参考NVIDIA[官方网站](https://developer.nvidia.com/nccl/nccl-download):
+
+ 
+  		wget http://developer.download.nvidia.com/compute/machine-learning/repos/rhel7/x86_64/nvidia-machine-learning-repo-rhel7-1.0.0-1.x86_64.rpm
+ 		rpm -i nvidia-machine-learning-repo-rhel7-1.0.0-1.x86_64.rpm
+ 		sudo apt-get install -y libnccl2=2.3.7-1+cuda9.0 libnccl-dev=2.3.7-1+cuda9.0
+ 		yum update -y
+ 		yum install -y libnccl-2.3.7-2+cuda9.0 libnccl-devel-2.3.7-2+cuda9.0 libnccl-static-2.3.7-2+cuda9.0
+
 
 ## 安装方式
 

doc/fluid/beginners_guide/install/install_Ubuntu.md

@@ -23,7 +23,15 @@
     * *CUDA 工具包8.0配合cuDNN v5*
     * *GPU运算能力超过1.0的硬件设备*
 
-您可参考NVIDIA官方文档了解CUDA和CUDNN的安装流程和配置方法，请见[CUDA](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/)，[cuDNN](https://docs.nvidia.com/deeplearning/sdk/cudnn-install/)
+	您可参考NVIDIA官方文档了解CUDA和CUDNN的安装流程和配置方法，请见[CUDA](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/)，[cuDNN](https://docs.nvidia.com/deeplearning/sdk/cudnn-install/)
+
+* 请确保您已经正确安装nccl2，或者按照以下指令安装nccl2（这里提供的是ubuntu 16.04，CUDA9，cuDNN7下nccl2的安装指令），更多版本的安装信息请参考NVIDIA[官方网站](https://developer.nvidia.com/nccl/nccl-download):
+
+
+		wget https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu1604/x86_64/nvidia-machine-learning-repo-ubuntu1604_1.0.0-1_amd64.deb
+		dpkg -i nvidia-machine-learning-repo-ubuntu1604_1.0.0-1_amd64.deb`
+		sudo apt-get install -y libnccl2=2.3.7-1+cuda9.0 libnccl-dev=2.3.7-1+cuda9.0
+
 
 ## 安装方式