포스트

Ubuntu22.04 딥러닝 학습 서버 환경 구축 (cuda, cudnn 설치)

  • gpu : nvidia GeForce RTX 3060
  • Ubuntu 22.04

cuda driver 510.108.03, 528.79, 531.41, 536

python 3.10.11

cuda 툴킷 11.8, 11.5, 11.7

cudnn 8.9.6, 8.5.1, 8.8.1

compute capability 8.6인 경우 CUDA 11.1부터 최신 버전인 12.4까지 사용할 수 있음

  1. driver 설치 https://www.nvidia.com/Download/index.aspx

  2. turm window feature

virtual machine platform window subsystem for linux 체크 -> reboot

  1. setting defualt wsl version to wsl2
    • command prompt ``` wsl –set-default-version 2

wsl –list –online

wsl –install -d Ubuntu

1
2
* ubuntu

username password

cd /mnt/c

dir

exit

1
2
3
3. how to install nvidia cuda for wsl2
* command prompt

wsl –list –verbose

wsl -t Ubuntu

wsl –list –verbose

wsl –distribution Ubuntu

exit

1
2
* ubuntu

gcc –version

sudo apt install gcc –fix-missing

1
2
* cuda toolkit, linux ->  x86_64 -> wsl-ubuntu -> 2.0 -> deb(local)

wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600 wget https://developer.download.nvidia.com/compute/cuda/11.7.0/local_installers/cuda-repo-wsl-ubuntu-11-7-local_11.7.0-1_amd64.deb sudo dpkg -i cuda-repo-wsl-ubuntu-11-7-local_11.7.0-1_amd64.deb sudo cp /var/cuda-repo-wsl-ubuntu-11-7-local/cuda-*-keyring.gpg /usr/share/keyrings/ sudo apt-get update sudo apt-get -y install cuda

1
2
4. how to set path variable for nvidia cuda on linux

cd ~

nano .bashrc

export PATH=/usr/local/cuda-11.8/bin${PATH:+:${PATH}}

source ~/.bashrc

echo $PATH

sudo apt install nvidia-cuda-toolkit

nvcc -V

nvidia-smi

1

sudo apt-get install python3-pip

1
2
3
* pytorch : stable -> linux -> pip -> python -> cuda11.8
https://pytorch.org/get-started/locally/

pip3 install torch torchvision torchaudio –index-url https://download.pytorch.org/whl/cu118 python3 import torch torch.cuda.is_available() torch.version

1
2
5. compliling cuda program to see if cuda works

cd ~ nano test.cu nvcc test.cu -o test ./test nvprof ./test

1
2
3
4
5
6
7
6. fix nvidia profiler error
* nvidia control panel
desktop -> enable developer settings
developer -> manage gpu -> allow access 

* ubuntu 

nvprof ./test

```


이 기사는 저작권자의 CC BY 4.0 라이센스를 따릅니다.