[Ubuntu 18.04 Deep learning Local 환경설정] NVIDIA cuDNN Installation 설치 확인 Tensorflow Pytorch

1. 설치 상태 확인

nvidia-smi

nvidia 드라이버 상태, 디바이스 상태 등을 확인

실행결과

2. cuDNN 다운로드

cuDNN 사이트에서 로그인 후, 위의 CUDA 버전에 맞는 cuDNN을 다운로드 받는다.
https://developer.nvidia.com/cudnn

CUDA Deep Neural Network

Ubuntu 18.04

3. 파일 설치

터미널에서 다운로드 된 경로로 이동(cd Downloads/) 후 압축 해제

tar -xvf cudnn-linux-x86_64-8.9.7.29_cuda11-archive.tar.xz

/usr/local/cuda 디렉토리로 복사

cd cudnn-linux-x86_64-8.9.7.29_cuda11-archive
sudo cp include/cudnn* /usr/local/cuda/include
sudo cp lib/libcudnn* /usr/local/cuda/lib64
sudo chmod a+r /usr/local/cuda/include/cudnn.h /usr/local/cuda/lib64/libcudnn*

cuda 디렉토리와 실제 설치된 cuda-11.7 디렉토리를 심볼릭 링크 설정

sudo ln -sf /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_adv_train.so.8.9.7 /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_adv_train.so.8
sudo ln -sf /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8.9.7  /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_ops_infer.so.8
sudo ln -sf /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8.9.7  /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_cnn_train.so.8
sudo ln -sf /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8.9.7  /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_adv_infer.so.8
sudo ln -sf /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_ops_train.so.8.9.7  /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_ops_train.so.8
sudo ln -sf /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8.9.7 /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn_cnn_infer.so.8
sudo ln -sf /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn.so.8.9.7 /usr/local/cuda-11/targets/x86_64-linux/lib/libcudnn.so.8

4. 설치 확인

sudo ldconfig
ldconfig -N -v $(sed 's/:/ /' <<< $LD_LIBRARY_PATH) 2>/dev/null | grep libcudnn

실행결과

4-1. Tensorflow에서 확인

import tensorflow as tf
gpus = tf.config.experimental.list_physical_devices('GPU')
for gpu in gpus: 
    tf.config.experimental.set_memory_growth(gpu, True)
print(len(gpus))

실행결과 본인 컴퓨터의 gpu 갯수가 나오면 성공 !

4-2. Pytorch로 확인

import torch

# Check if GPU is available
if torch.cuda.is_available():
    # Get the number of available GPUs
    num_gpus = torch.cuda.device_count()
    print(f"Number of available GPUs: {num_gpus}")

    # Get the name of each GPU
    for i in range(num_gpus):
        print(f"GPU {i + 1}: {torch.cuda.get_device_name(i)}")
else:
    print("No GPU available. Using CPU.")

실행 결과

Number of available GPUs: 1
GPU 1: NVIDIA GeForce RTX 3080