Ubuntu20.04 安裝 cuda 11

3 min readMay 10, 2021

Photo by Caspar Camille Rubin on Unsplash

筆者上一次配置訓練用gpu 還是ubuntu 16.04 + cuda 9 + tensorflow 1.10，那時後也是搞了兩天多才完成，最近有一些範例訓練code 都是使用tensorflow 2來跑，以前都是用colab 先撐著，剛好公司配發新的電腦，趁著裝系統和環境的機會順便升級一下自己的訓練環境，在安裝過程中著實踩了不少坑，這裡就順便紀錄一下…

首先介紹一下筆者的電腦:

CPU : I7-11700

Board : msi mpg-z490

GPU: GTX 1660 super

SSD :512G SSD *2

HDD : 1TB

先說說筆者裝系統踩到的坑，第一個就是不知道是什麼問題，ubuntu 16.04 和18.04 通通裝不上去，只能用20.04，然後就遇到第一個坑了，這主機板安裝20.04 竟然沒辦法識別有線網路,上網查詢之後發現原來ubuntu 正式板驅動還未更新到這張板子,所以必須先用開發者版本的kernal , 安裝make和build套件,再降到正式kernal安裝驅動,才有網路可以使用.....

-----以上純屬廢話----

-----以下正文開始----

刪除所有已安裝的nvidia驅動、cuda、cudnn

sudo apt remove nvidia-cuda-toolkit
sudo apt-get remove *nvidia*
sudo apt-get purge *nvidia*
sudo apt-get remove libnvidia*
sudo apt-get purge libnvidia*
sudo apt-get autoremove
sudo apt-get autoclean
sudo rm -rf /usr/local/cuda*

禁用nouveau 驅動

sudo gedit /etc/modprobe.d/blacklist.conf

輸入以下內容保存

blacklist nouveau
options nouveau modeset=0

更新系統和重開機

sudo update-initramfs -u
sudo reboot

查看是否禁用成功(無輸出代表禁用成功)

lsmod | grep nouveau

安裝NVIDIA驅動(450版本)

下載nvidia 驅動(注意須下載450版本,460不支援cuda11.0)

Linux x64 (AMD64/EM64T) Display Driver

Download drivers for NVIDIA products including GeForce graphics cards, nForce motherboards, Quadro workstations, and…

www.nvidia.com

安裝驅動並檢查

sudo ./NVIDIA-Linux-x86_64–450.119.03.run
nvidia-smi

安裝CUDA

下載CUDA 11.0

wget http://developer.download.nvidia.com/compute/cuda/11.0.2/local_installers/cuda_11.0.2_450.51.05_linux.run
sudo ./cuda_11.0.2_450.51.06_linux.run

成功之後會看到以下內容

driver: Not Selected
Toolkit: Installed in /usr/local/cuda-11.0/
Samples: Installed in /home/itts-ai/
Please make sure that
— PATH includes /usr/local/cuda-11.0/bin
— LD_LIBRARY_PATH includes /usr/local/cuda-11.0/lib64, or, add /usr/local/cuda-11.0/lib64 to /etc/ld.so.conf and run ldconfig as root
To uninstall the CUDA Toolkit, run cuda-uninstaller in /usr/local/cuda-11.0/bin
***WARNING: Incomplete installation! This installation did not install the CUDA Driver. A driver of version at least .00 is required for CUDA 11.0 functionality to work.
To install the driver using this installer, run the following command, replacing <CudaInstaller> with the name of this run file:
sudo <CudaInstaller>.run — silent — driver
Logfile is /var/log/cuda-installer.log

添加環境變數

nano ~/.bashrc

在最下面新增以下內容

export PATH=”/usr/local/cuda-11.0/bin:$PATH”
export LD_LIBRARY_PATH=”/usr/local/cuda-11.0/lib64:$LD_LIBRARY_PATH”

更新

source ~/.bashrc

檢查cuda 安裝

nvcc -V

安裝CUDNN

下載cudnn 8.0.5.39 ,此處需要新建/登入NVIDIA會員

https://developer.nvidia.com/feedback/ws-gpu-2021-04

下載請選擇cudnn_Library_for_linux[x86_64]

解壓並將相關檔案移至cuda 資料夾

tar -xzvf cudnn-11.0-linux-x64-v8.0.5.39.tgz
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 /usr/local/cuda/lib64/libcudnn*

大功告成