nvidia cuda9.0 cudnn安装

https://www.jianshu.com/p/fc5edbd6f480
https://blog.csdn.net/wanzhen4330/article/details/81699769?from=singlemessage
pcl:配置PC以将Nvidia GPU与PCL配合使用
http://pointclouds.org/documentation/tutorials/gpu_install.php#gpu-install
CUDA Tutorial
https://cuda-tutorial.readthedocs.io/en/latest/
1.查看电脑是32位还是64位
sudo uname --s 显示内核名字s
sudo uname --r 显示内核版本
sudo uname --n 显示网络主机名
sudo uname --p 显示cpu
x86_64是64位的。
查询更全uname -m && cat /etc/*release，得到如下信息：

x86_64
DISTRIB_ID=Ubuntu
DISTRIB_RELEASE=16.04
DISTRIB_CODENAME=xenial
DISTRIB_DESCRIPTION=“Ubuntu 16.04.5 LTS”
NAME=“Ubuntu”
VERSION=“16.04.5 LTS (Xenial Xerus)”
ID=ubuntu
ID_LIKE=debian
PRETTY_NAME=“Ubuntu 16.04.5 LTS”
VERSION_ID=“16.04”
HOME_URL=“http://www.ubuntu.com/”
SUPPORT_URL=“http://help.ubuntu.com/”
BUG_REPORT_URL=“http://bugs.launchpad.net/ubuntu/”
VERSION_CODENAME=xenial
UBUNTU_CODENAME=xenial

2. 查看nvidia 驱动是否安装
   nvidia-smi

GeForce GTX 1080

3. 查看gcc是否安装
gcc -v

gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.11)

4. 查看适用的CUDA版本
   （查看本机 NVIDIA 显卡类型是否支持 cuda）
   https://developer.nvidia.com/cuda-gpus#compute
   (对 nvidia GPU 驱动也是有最低版本要求的)
   https://docs.nvidia.com/deploy/cuda-compatibility/index.html#abstract
   GeForce GTX 1080 算力 6.1
   所以我们的nvidia显卡可以安装的cuda是<9.1

5. 官网cuda安装链接
   https://developer.nvidia.com/cuda-downloads
   含有早期版本列表： https://developer.nvidia.com/cuda-toolkit-archive
   安装教程：https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html
   cuda9.0: https://docs.nvidia.com/cuda/archive/9.0/cuda-installation-guide-linux/index.html

6. 下载安装cuda9.0
   
   有4个补丁。
   这些补丁要求首先安装基本安装程序。

Do you accept the previously read EULA?
accept/decline/quit: accept
Install NVIDIA Accelerated Graphics Driver for Linux-x86_64 384.81?
(y)es/(n)o/(q)uit: y
Install the CUDA 9.0 Toolkit?
(y)es/(n)o/(q)uit: y
Enter Toolkit Location
[ default is /usr/local/cuda-9.0 ]: 点回车
Do you want to install a symbolic link at /usr/local/cuda?
(y)es/(n)o/(q)uit: y
Install the CUDA 9.0 Samples?
(y)es/(n)o/(q)uit: y
Enter CUDA Samples Location
[ default is /home/gnn ]: 回车

接下，开始安装

Installing the CUDA Toolkit in /usr/local/cuda-9.0 …
…
Driver: Not Selected
Toolkit: Installed in /usr/local/cuda-9.0
Samples: Installed in /home/gnn

下载补丁

7. 检查是否安装成功
   1）定位到/usr/local/cuda/bin
   ./nvcc -V得到如下显示

nvcc: NVIDIA ® Cuda compiler driver
Copyright © 2005-2017 NVIDIA Corporation
Built on Fri_Sep__1_21:08:03_CDT_2017
Cuda compilation tools, release 9.0, V9.0.176)

2）编译自带的samples
测试设备
cd /usr/local/cuda-9.0/samples/1_Utilities/deviceQuery
sudo make
./deviceQuery
显示PASS说明安装成功。

deviceQuery, CUDA Driver = CUDART, CUDA Driver Version = 9.0, CUDA Runtime Version = 9.0, NumDevs = 1
Result = PASS

测试宽带
cd /usr/local/cuda-9.0/samples/1_Utilities/bandwidthTest
sudo make
./bandwidthTest
显示：

Result = PASS

8. 安装cudnn
   官网安装教程：https://docs.nvidia.com/deeplearning/sdk/cudnn-install/index.html
   计算能力为3.0或更高的GPU
   cuDNN支持列表：https://docs.nvidia.com/deeplearning/sdk/cudnn-support-matrix/index.html
   
9. 下载cuDNN for Linux
   https://developer.nvidia.com/rdp/cudnn-download

cuDNN Library for Linux

下载解压，
sudo cp cuda/include/cudnn.h /usr/local/cuda/include/
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64/
sudo chmod a+r /usr/local/cuda/include/cudnn.h
sudo chmod a+r /usr/local/cuda/lib64/libcudnn*
10. 验证是否安装成功
cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2

#define CUDNN_MAJOR 7
#define CUDNN_MINOR 6
#define CUDNN_PATCHLEVEL 3
–
#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)
#include “driver_types.h”