cudnn Segmentation fault

Hello, I use Ubuntu 16.04.1 LTS and have cuda-9.1, I try to install cudnn using cudnn-9.1-linux-x64-v7.tgz.After I cp and chmod the file mentioned in https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html and verifying the cudnn :

The error information showd below :

cudnnGetVersion() : 7005 , CUDNN_VERSION from cudnn.h : 7005 (7.0.5)
Host compiler version : GCC 5.4.0
There are 3 CUDA capable devices on your machine :
device 0 : sms 15 Capabilities 3.5, SmClock 745.0 Mhz, MemSize (Mb) 11441, MemClock 3004.0 Mhz, Ecc=1, boardGroupID=0
device 1 : sms 2 Capabilities 3.0, SmClock 954.0 Mhz, MemSize (Mb) 1999, MemClock 2000.0 Mhz, Ecc=0, boardGroupID=1
device 2 : sms 15 Capabilities 3.5, SmClock 745.0 Mhz, MemSize (Mb) 11441, MemClock 3004.0 Mhz, Ecc=1, boardGroupID=2
Using device 0

Testing single precision
Segmentation fault (core dumped)

What should I do? Thank you !