[Ubuntu] Ubuntu server(20.04.5 LTS) + CUDA 11.6 + cudnn8.4.0 설치

1. ubuntu 20.04.5 설치 (Rufus)

 

https://releases.ubuntu.com/20.04.5/?_ga=2.163840441.788706021.1663029194-1495615447.1663029194 

Ubuntu 20.04.5 LTS (Focal Fossa)

Swap 10000MB, EFI 500MB, else root (/) 

 

2. conda 설치

 

sudo apt-get update
wget https://repo.anaconda.com/archive/Anaconda3-2022.05-Linux-x86_64.sh
bash Anaconda3-2022.05-Linux-x86_64.sh

nano ~/.bashrc
# 추가
export PATH="$HOME/anaconda3/bin:$PATH"
source ~/.bashrc

# 확인
conda -V

 

3. driver 설치 

 

ubuntu-drivers devices

sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt update
sudo apt-get install nvidia-driver-510
# 재부팅
sudo reboot
# 확인
nvidia-smi

 

4. cuda 11.6 설치

 

wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/x86_64/cuda-ubuntu2004.pin
sudo mv cuda-ubuntu2004.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/11.6.0/local_installers/cuda-repo-ubuntu2004-11-6-local_11.6.0-510.39.01-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu2004-11-6-local_11.6.0-510.39.01-1_amd64.deb
sudo apt-key add /var/cuda-repo-ubuntu2004-11-6-local/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda

sudo nano ~/.bashrc
# 추가
export PATH=/usr/local/cuda-11.6/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-11.6/lib64:${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
source ~/.bashrc

# 확인
nvcc -V

 

5.  cudnn 설치

 

cudnn 8.4.0 다운

https://developer.nvidia.com/compute/cudnn/secure/8.4.0/local_installers/11.6/cudnn-linux-x86_64-8.4.0.27_cuda11.6-archive.tar.xz

(로그인 필요)

# 압축풀기
unxz cudnn-linux-x86_64-8.4.0.27_cuda11.6-archive.tar.xz
tar -xvf cudnn-linux-x86_64-8.4.0.27_cuda11.6-archive.tar

# 파일 복붙
sudo cp cudnn-linux-x86_64-8.4.0.27_cuda11.6-archive/include/* /usr/local/cuda-11.6/include
sudo cp -P cudnn-linux-x86_64-8.4.0.27_cuda11.6-archive/lib/* /usr/local/cuda-11.6/lib64
sudo chmod a+r /usr/local/cuda-11.6/lib64/libcudnn*

# 확인
cat /usr/local/cuda/include/cudnn_version.h | grep CUDNN_MAJOR -A 2

 

6. 가상환경 만들기

 

conda create -n pytorch python=3.9

conda activate pytorch

 

7. pytorch 1.12.1  설치 

 

conda install pytorch==1.12.0 torchvision==0.13.0 torchaudio==0.12.0 cudatoolkit=11.6 -c pytorch -c conda-forge

https://pytorch.org/get-started/previous-versions/ 참고