Redhat Configuring the GPU computing environment

First step: Install Nvidia graphics driver

installation file Download URL: HTTP://WWW.NVIDIA.CN/DOWNLOAD/INDEX.ASPX?LANG=CN

Select the appropriate video card model: down

$ wget-r-o xxxx.rpm  "Download_url"

The instruction set is as follows:

Once you accept the download follow the steps listed below 
i) ' Rpm-i nvidia-diag-driver-local-repo-rhel7-384.5 9-1.0-1.x86_64.rpm ' 
II) ' Yum clean all '
III ' yum install cuda-drivers ' 
IV) ' reboot '

Step Two: Install cuda-8.0 environment download Address: https://developer.nvidia.com/cuda-downloads

The instruction set is as follows (Law 1, the directory/uer/local directory is not installed on my server):

Installation instructions:
' sudo rpm-i cuda-repo-rhel7-8-0-local-ga2-8.0.61-1.x86_64.rpm ' #注意对应相对的操作系统
' sudo yum clean all '
sudo yum install Cuda '

The instruction set is as follows (Act 2):

$ chmod 755 cuda_8.0.61_375.26_linux.run
$./cuda_8.0.61_375.26_linux.run

To configure cuda-8.0 environment variables:

$ vim/etc/profile

Export cuda_home=/usr/local/cuda-8.0  
export path= $CUDA _home/bin: $PATH    
export ld_library_path= $CUDA _home/ LIB64: $LD _library_path  

$ source/etc/profile

Step three: Install cudnn-v5+ to cuda-8.0 to download the CUDNN installation package: HTTPS://DEVELOPER.NVIDIA.COM/CUDNN Unzip and copy the contents to/usr/local/ In the Cuda-8.0/include and lib64 directories:

sudo tar xvzf cudnn-8.0-linux-x64-v5.1.tgz  
sudo cp cuda/include/cudnn.h/usr/local/cuda-8.0/include  
sudo cp cuda/lib64/libcudnn*/usr/local/cuda-8.0/lib64  
sudo chmod a+r/usr/local/cuda-8.0/include/cudnn.h/usr/local/ cuda-8.0/lib64/libcudnn*   #分配包的权限