系统配置
系统版本： Centos7.6
语言: Python3.5(anaconda3 4.2)
框架: Tensorflow
安装依赖
sudo yum install openjdk-8-jdk git python-dev python3-dev python-numpy python3-numpy build-essential python-pip python3-pip python-virtualenv swig python-wheel libcurl3-dev curl
安装 NVIDIA 驱动 和 CUDA
curl -O http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/cuda-repo-ubuntu1604_9.0.176-1_amd64.deb
sudo apt-key adv --fetch-keys http://developer.download.nvidia.com/compute/cuda/repos/ubuntu1604/x86_64/7fa2af80.pub
sudo dpkg -i ./cuda-repo-ubuntu1604_9.0.176-1_amd64.deb
sudo apt-get update
sudo apt-get install cuda-9-0
# 这一步有可能报错，出现
# trying to replace " /usr/lib/x86_64-linux-gnu/libGLX_indirect.so.0 ",which
# belong to the package libglx-mesa0:amd64 18.0.0~rc5-1ubuntu1
# errors have been encountered during the execution of :
# /var/cuda-repo-9-2-local/./nvidia-396_396.26-0ubuntu1_amd64.deb
# 或者类似，请尝试
# dpkg -i --force-overwrite /var/cache/apt/archives/nvidia-xxx
# 参见 https://askubuntu.com/questions/1037982/nividia-396-installation-blocked-by-libglx-on-18-04
# 重启
sudo reboot
# 检查驱动安装
nvidia-smi
如下图 (双 GPU)
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 396.44 Driver Version: 396.44 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Tesla P4 Off | 00000000:00:07.0 Off | 0 |
| N/A 37C P0 22W / 75W | 0MiB / 7611MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
| 1 Tesla P4 Off | 00000000:00:08.0 Off | 0 |
| N/A 37C P0 23W / 75W | 6722MiB / 7611MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
Install cudnn
tar -zxvf cudnn-9.0-linux-x64-v7.1.solitairetheme8
sudo cp cuda/include/cudnn.h /usr/local/cuda/include/
sudo cp cuda/lib64/libcudnn* /usr/local/cuda/lib64/ -d
sudo chmod a+r /usr/local/cuda/include/cudnn.h
sudo chmod a+r /usr/local/cuda/lib64/libcudnn*
配置环境变量 ~/.bashrc:
export LD_LIBRARY_PATH=/usr/local/cuda-9.0/lib64:$LD_LIBRARY_PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH
export CUDA_HOME=/usr/local/cuda
export PATH="$PATH:/usr/local/cuda/bin"
# 刷新
source ~/.bashrc
安装 Python 环境 (miniconda)
wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash Miniconda3-latest-Linux-x86_64.sh
# press s to skip terms
# Do you approve the license terms? [yes|no]
# yes
# Miniconda3 will now be installed into this location:
# accept the location
# 如果错误信息：bunzip2: command not found
yum install -y bzip2
# Do you wish the installer to prepend the Miniconda3 install location
# to PATH in your /home/ghost/.bashrc ? [yes|no]
# yes
source ~/.bashrc
Create conda env to install tf
conda create -n tensorflow
# press y a few times
Activate env
source activate tensorflow
安装带 gpu 的 tensorflow 版本
pip install tensorflow-gpu -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
测试安装结果
# start python shell
python
# run test script
import tensorflow as tf
hello = tf.constant('Hello, TensorFlow!')
sess = tf.Session()
print(sess.run(hello))
# 正常会在 nvidia-smi 看到(双 gpu)
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 396.44 Driver Version: 396.44 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 Tesla P4 Off | 00000000:00:07.0 Off | 0 |
| N/A 38C P0 23W / 75W | 7241MiB / 7611MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
| 1 Tesla P4 Off | 00000000:00:08.0 Off | 0 |
| N/A 37C P0 23W / 75W | 7387MiB / 7611MiB | 0% Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 54501 C python 7231MiB |