cuda.md

실습실 CUDA 설치

설치 사양

• Ubuntu 16.04 LTS
• CUDA 8.0
• cuDNN 5.1
• Tensorflow on python3

CUDA Toolkit 8 설치

wget http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/cuda-repo-ubuntu1604_8.0.44-1_amd64.deb
sudo dpkg -i cuda-repo-ubuntu1604_8.0.44-1_amd64.deb
sudo apt update
sudo apt install --no-install-recommends cuda

환경변수 설정

/etc/environment에 다음과 같이 추가

CUDA_HOME="/usr/local/cuda"
PATH="/usr/local/cuda/bin:(기존 PATH 내용)"

/etc/ld.so.conf.d/에 새 설정파일을 만들고 (e.g. cuda.conf) 다음과 같이 추가

/usr/local/cuda/lib64

(LD_LIBRARY_PATH 역할)

적용되려면 재부팅 필요

cuDNN 설치

cuDNN 파일은 sherry에 보관 (변경 시 추가바람)

tar -zxvf cudnn.tgz
cd cuda
sudo cp include/cudnn.h $CUDA_HOME/include/
sudo cp lib64.* $CUDA_HOME/lib64/

테스트

$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2016 NVIDIA Corporation
Built on Tue_Jan_10_13:22:03_CST_2017
Cuda compilation tools, release 8.0, V8.0.61

$ cd $CUDA_HOME/samples/1_Utilities/bandwidthTest/
$ sudo make
$ ./bandwidthTest
[CUDA Bandwidth Test] - Starting...
Running on...

 Device 0: GeForce GT 720
 Quick Mode

 Host to Device Bandwidth, 1 Device(s)
 PINNED Memory Transfers
   Transfer Size (Bytes)	Bandwidth(MB/s)
   33554432			3204.9

 Device to Host Bandwidth, 1 Device(s)
 PINNED Memory Transfers
   Transfer Size (Bytes)	Bandwidth(MB/s)
   33554432			3250.3

 Device to Device Bandwidth, 1 Device(s)
 PINNED Memory Transfers
   Transfer Size (Bytes)	Bandwidth(MB/s)
   33554432			12410.7

Result = PASS

NOTE: The CUDA Samples are not meant for performance measurements. Results may vary when GPU Boost is enabled.