[우분투/리눅스] 우분투 20.04 CUDA 11.3, CuDNN 8.2.1 설치

1. 기존 설치된 CUDA 확인 및 제거

1-1. CUDA 확인

$ ls -l /usr/local | grep cuda
$ nvidia-smi # 추천 cuda 버전 확인 및 드라이버 설치 확인 
# nvcc -V # cuda 버전 확인 

2. 기존 NVIDA, CUDA 삭제

$ sudo apt-get purge nivida*
$ sudo apt-get autoremove
$ sudo apt-get autoclean
$ sudo rm -rf /usr/local/cuda*

1-3. 설치 가능한 드라이버 확인
$ ubuntu-drivers devices

1-4. 드라이버 설치
$ sudo apt install nvidia-driver-470

1-5. nvidia kernel module을 도와주는 modprobe 패키지 설치

$ sudo apt-get install dkms nvidia-modprobe
$ sudo apt update
$ sudo apt upgrade
$ sudo reboot # reboot
$ nvidia-smi # 설치 확인

2. CUDA 11.3.1 설치

2-1. 다음 링크에서 11.3 runfile 명령어 실행
https://developer.nvidia.com/cuda-toolkit-archive

$ wget https://developer.download.nvidia.com/compute/cuda/11.3.1/local_installers/cuda_11.3.1_465.19.01_linux.run
$ sudo sh cuda_11.3.1_465.19.01_linux.run
# continue
# accept
# driver 해제 후 install 

2-2. CUDA Toolkit 관련 설정을 환경변수에 추가

$ sudo sh -c "echo 'export PATH=$PATH:/usr/local/cuda-11.3/bin'>> /etc/profile"
$ sudo sh -c "echo 'export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-11.3/lib64'>> /etc/profile"
$ sudo sh -c "echo 'export CUDARDIR=/usr/local/cuda-11.3'>> /etc/profile"
$ source /etc/profile  

2-3. CUDA 설치 확인
$nvcc --version

3. cuDNN 8.2.1 설치

3-1. 다음 링크를 통해 회원가입 & 로그인 & 다운로드 실행
https://developer.nvidia.com/cudnn

# download cudnn 
# archive cudnn release
# download cudnn v8.2.1 for cuda 11.x
# scp server

3-2. 압축 풀기 & 복사

$ tar xvzf .tar

$ sudo cp cuda/include/cudnn* /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*

$ sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_adv_train.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_adv_train.so.8
$ sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8.2.1  /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8
$ sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8.2.1  /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8
$ sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8.2.1  /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8
$ sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_ops_train.so.8.2.1  /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_ops_train.so.8
$ sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8
$ sudo ln -sf /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn.so.8.2.1 /usr/local/cuda-11.3/targets/x86_64-linux/lib/libcudnn.so.8

3-3.

# 새로 추가된 라이브러리를 시스템에서 찾을 수 있도록
$ sudo ldconfig 
# 설정 잘 되었는지 확인 
$ ldconfig -N -v $(sed 's/:/ /' <<< $LD_LIBRARY_PATH) 2>/dev/null | grep libcudnn
$ ls -l /usr/local | grep cuda