ubuntu18.04安装tensorRT和重装cuda
1. 环境
查看cuda版本：
$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Thu_Jun_11_22:26:38_PDT_2020
Cuda compilation tools, release 11.0, V11.0.194
Build cuda_11.0_bu.TC445_37.28540450_0
查看cudnn版本：
$ find / -name cudnn_version.h 2>&1 | grep -v "Permission denied"
/home/zxq/Downloads/cudnn-11.0-linux-x64-v8.0.5/cuda/include/cudnn_version.h
# 由路径可知，我的是v8.0.5，与cuda11.0是匹配的就行。cuda11.0+ cudnn8.0.5
2.安装TensorRT
严格按照官网操作，我下载的是nv-tensorrt-repo-ubuntu1804-cuda11.0-trt7.2.3.4-ga-20210226_1-1_amd64.deb，这需要注意，我选择的安装是deb安装方式，这里的前提是cuda必须也是deb安装的，不然就会报一堆的依赖错误，我的问题已经解决了，贴个示例。cuda卸载重装请看第3步。
The following packages have unmet dependencies.
tensorrt : Depends: libnvinfer6 (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvinfer-plugin6 (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvparsers6 (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvonnxparsers6 (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvinfer-bin (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvinfer-dev (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvinfer-plugin-dev (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvparsers-dev (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvonnxparsers-dev (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvinfer-samples (= 6.0.1-1+cuda10.1) but it is not going to be installed
Depends: libnvinfer-doc (= 6.0.1-1+cuda10.1) but it is not going to be installed
E: Unable to correct problems, you have held broken packages.是这个人提醒的我：
C++接口安装：
os="ubuntu1x04"
tag="cudax.x-trt7.x.x.x-ga-yyyymmdd"
sudo dpkg -i nv-tensorrt-repo-${os}-${tag}_1-1_amd64.deb
sudo apt-key add /var/nv-tensorrt-repo-${tag}/7fa2af80.pub
sudo apt-get update
sudo apt-get install tensorrt验证C++接口安装成功：
dpkg -l | grep TensorRT
如果要安装python接口：
If using Python 2.7:
sudo apt-get install python-libnvinfer-dev
The following additional packages will be installed:
python-libnvinfer
If using Python 3.x:
sudo apt-get install python3-libnvinfer-dev
The following additional packages will be installed:
python3-libnvinfer如果想要卸载tensorRT，可参考官方
3 cuda卸载重装。
各个版本的卸载方式可能不同，卸载官方文档的位置：/usr/local/cuda-10.0/doc/pdf/CUDA_Installation_Guide_Linux.pdf
sudo /usr/local/cuda-11.0/bin/cuda-uninstaller # 卸载cuda
# 卸载显卡驱动，后面还要再做一边
# （1）卸载现有的显卡驱动
sudo apt-get purge nvidia*
# （2）如果是离线安装的
卸载驱动和安装一样，首先ctrl+Alt+F2进入命令行状态，然后停止lightdm
sudo service lightdm stop
或者
sudo /etc/init.d/lightdm stop
# 卸载
sudo /usr/bin/nvidia-uninstall
也可以
sudo apt-get install autoremove --purge nvidia*deb方式按照cuda，按照官网来即可。
Installation Instructions:
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin
sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget http://developer.download.nvidia.com/compute/cuda/11.0.2/local_installers/cuda-repo-ubuntu1804-11-0-local_11.0.2-450.51.05-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1804-11-0-local_11.0.2-450.51.05-1_amd64.deb
sudo apt-key add /var/cuda-repo-ubuntu1804-11-0-local/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda但是官网默认会再次安装显卡驱动，官网的安装的显卡驱动有问题，所以需要重新卸载显卡驱动，重新安装显卡驱动。
sudo apt --purge remove nvidia*
# 安装
$ ubuntu-drivers devices
== /sys/devices/pci0000:00/0000:00:01.0/0000:01:00.0 ==
modalias : pci:v000010DEd00002204sv00007377sd0000140Abc03sc00i00
vendor : NVIDIA Corporation
driver : nvidia-driver-460 - third-party free
driver : nvidia-driver-460-server - distro non-free recommended
driver : xserver-xorg-video-nouveau - distro free builtin
$ sudo apt-get install nvidia-driver-460 # 推荐是server，都可以。离线安装驱动，重启会黑屏。建议采用上面的方式安装驱动。
cudnn去官网下载和cuda匹配的版本。然后将里面解压的文件复制到cuda目录中，命令如下。
sudo cp cuda/include/* /usr/local/cuda/include
sudo cp cuda/lib64/* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*
环境配置。
sudo gedit ~/.bashrc打开的文档末尾输入（cuda版本对应修改下）：
export PATH=/usr/local/cuda-11.0/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-11.0/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}验证显卡驱动是否安装成功。
$ nvidia-smi
Fri Apr 9 14:01:10 2021
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 460.67 Driver Version: 460.67 CUDA Version: 11.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 GeForce RTX 3090 Off | 00000000:01:00.0 On | N/A |
| 0% 42C P8 31W / 350W | 206MiB / 24265MiB | 18% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 1233 G /usr/lib/xorg/Xorg 78MiB |
| 0 N/A N/A 1787 G /usr/bin/gnome-shell 42MiB |
| 0 N/A N/A 2172 G ...AAAAAAAAA= --shared-files 82MiB |
+-----------------------------------------------------------------------------+
验证cuda是否安装成功。
$ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Thu_Jun_11_22:26:38_PDT_2020
Cuda compilation tools, release 11.0, V11.0.194
Build cuda_11.0_bu.TC445_37.28540450_0