CUDA 10.2 & cuDNN 7.6 install for Ubuntu 설치

CUDA 10.2 & cuDNN install for Ubuntu 설치

CUDA install 설치

CPU : i3-4150

GPA : GTX 750

OS : Ubuntu 18.04

First, move to
https://developer.nvidia.com/cuda-downloads

select OS, Arc, Version, Type. I chose deb(network)
원하는 버전과 설치 방식을 고른 후 설치 한다. 나는 deb(network)를 선택하였다.

For me:

$ wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/cuda-ubuntu1804.pin$ sudo mv cuda-ubuntu1804.pin /etc/apt/preferences.d/cuda-repository-pin-600$ sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/7fa2af80.pub$ sudo add-apt-repository "deb http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1804/x86_64/ /"$ sudo apt-get update$ sudo apt-get -y install cuda$ sudo reboot

To check installation, follow these steps:
설치를 확인하려면 아래를 따라하도록 한다.

$ vim .bashrc   export PATH=$PATH:/usr/local/cuda/bin   export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda/lib64$ source .profile$ nvcc --versionnvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Wed_Oct_23_19:24:38_PDT_2019
Cuda compilation tools, release 10.2, V10.2.89$ nvidia-smiMon Mar 30 02:21:26 2020+-----------------------------------------------------------------------------+| NVIDIA-SMI 440.64.00    Driver Version: 440.64.00    CUDA Version: 10.2     ||-------------------------------+----------------------+----------------------+| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC || Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. ||===============================+======================+======================||   0  GeForce GTX 750     On   | 00000000:01:00.0 Off |                  N/A || 33%   31C    P8     1W /  38W |      1MiB /   980MiB |      0%      Default |+-------------------------------+----------------------+----------------------++-----------------------------------------------------------------------------+| Processes:                                                       GPU Memory ||  GPU       PID   Type   Process name                             Usage      ||=============================================================================||  No running processes found                                                 |+-----------------------------------------------------------------------------+

$ cd /usr/local/cuda/samples/$ sudo make -k$ cd bin/x86_64/linux/release/$ ./deviceQuery



$ ./bandwidthTest

cuDNN install 설치

Nvidia login is required for cuDNN installation

cuDNN 을 다운 받기 위해서는 Nvidia 계정이 필요 하다.

Move to
https://developer.nvidia.com/rdp/cudnn-download
download files to local and put the files using sftp on the server
파일을 로컬에 다운받고 sftp를 이용하여 서버에 전송한다.

Installation, check 설치,확인

$ sudo dpkg -i libcudnn7_7.6.5.32-1+cuda10.2_amd64.deb$ sudo dpkg -i libcudnn7-dev_7.6.5.32-1+cuda10.2_amd64.deb$ sudo dpkg -i libcudnn7-doc_7.6.5.32-1+cuda10.2_amd64.deb$ cat /usr/include/cudnn.h | grep CUDNN_MAJOR -A 2#define CUDNN_MAJOR 7#define CUDNN_MINOR 6#define CUDNN_PATCHLEVEL 5--#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)
#include "driver_types.h"