1.下载cudnn-8.0-linux-x64-v6.0

地址：https://developer.nvidia.com/compute/machine-learning/cudnn/secure/v6/prod/8.0_20170307/cudnn-8.0-linux-x64-v6.0-tgz

2.将下载的源文件放在/usr/local/目录下

$ sudo cp cudnn-8.0-linux-x64-v6.0.tgz /usr/local
$ cd /usr/local

3.解压

$ sudo tar -xvzf cudnn-8.0-linux-x64-v6.0.tgz

4.修改文件权限

$ sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*

5.安装 3 个库文件

下载 runtime library,developer library,cuDNN Library

$ sudo dpkg -i libcudnn6_6.0.21-1+cuda8.0_amd64.deb
$ sudo dpkg -i libcudnn6-dev_6.0.21-1+cuda8.0_amd64.deb
$ sudo dpkg -i libcudnn6-doc_6.0.21-1+cuda8.0_amd64.deb

6.测试 cuDNN

6.1复制 cuDNN sample 到一个目录下,这里复制到 HOME 下

$ cp -r /usr/src/cudnn_samples_v6 /$HOME

6.2进入 HOME 目录

$ cd $HOME/cudnn_samples_v6/mnistCUDNN/

6.3 编译 mnistCUDNN sample

$ make clean && make –j8

6.4 运行 mnistCUDNN sample

$ ./mnistCUDNN

如果出现 Test passed!表明 cuDNN 已成功安装

./mnistCUDNN 
cudnnGetVersion() : 6021 , CUDNN_VERSION from cudnn.h : 6021 (6.0.21)
Host compiler version : GCC 5.4.0
There are 1 CUDA capable devices on your machine :
device 0 : sms  2  Capabilities 2.1, SmClock 950.0 Mhz, MemSize (Mb) 1985, MemClock 900.0 Mhz, Ecc=0, boardGroupID=0
Using device 0

Testing single precision
CUDNN failure
Error: CUDNN_STATUS_ARCH_MISMATCH
mnistCUDNN.cpp:394
Aborting...

查看Compute Capability，官方的cudnn加速是不支持3.0以下的版本的(http://blog.csdn.net/songyu0120/article/details/43233907)

https://developer.nvidia.com/cuda-gpus