docker中使用gpu需要安装相关的插件环境。
系统环境:
操作系统:Ubuntu 16.04 64位
gpu: nvidia p100
需要的软件:
cuda driver
nvidia-docker
安装步骤:
cuda driver
cuda driver包括了cuda工具包和nvidia driver
安装方式
1.Package安装
2.Runfile安装
https://developer.nvidia.com/cuda-downloads下载安装,安装方式参看官网
cuda driver验证:
cat /proc/driver/nvidia/version
安装成功显示:
NVRM version: NVIDIA UNIX x86_64 Kernel Module 410.66 Wed Oct 10 12:01:53 CDT 2018
GCC version: gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.11)
nvidia-docker
nvidia-docker是一个可以使用gpu的docker,是在docker上做了一层封装,实现了container和host之间cuda连接
安装方式
# If you have nvidia-docker 1.0 installed: we need to remove it and all existing GPU containers
docker volume ls -q -f driver=nvidia-docker | xargs -r -I{} -n1 docker ps -q -a -f volume={} | xargs -r docker rm -f
sudo apt-get purge -y nvidia-docker
# Add the package repositories
curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | \
sudo apt-key add -
distribution=$(. /etc/os-release;echo $ID$VERSION_ID)
curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | \
sudo tee /etc/apt/sources.list.d/nvidia-docker.list
sudo apt-get update
# Install nvidia-docker2 and reload the Docker daemon configuration
sudo apt-get install -y nvidia-docker2
sudo pkill -SIGHUP dockerd
参考:https://github.com/NVIDIA/nvidia-docker
验证:
# Test nvidia-smi with the latest official CUDA image
docker run --runtime=nvidia --rm nvidia/cuda:9.0-base nvidia-smi
runtime也可以配置默认
cat /etc/docker/daemon.json
{
"default-runtime": "nvidia",
"runtimes": {
"nvidia": {
"path": "nvidia-container-runtime",
"runtimeArgs": []
}
}
}
docker container运行:
项目中通过ffmpeg调用gpu的编解码功能。需要基于包含有nvidia cuda的Ubuntu系统镜像(nvidai/cuda tensorflow/tensorflow)加入编解码程序编译镜像,然后运行container
运行tensorflow模型
docker run --runtime=nvidia --rm --name tensorflow -ti tensorflow/tensorflow:r0.9-devel-gpu
Kubernetes nvidia gpu使用
部署docker容器gpu环境的基础上,Kubernetes 集群中使用gpu需要插件 NVIDIA/k8s-device-plugin,插件实现gpu数量统计,状态监测以及负载均衡等。参考https://github.com/NVIDIA/k8s-device-plugin
参考文章:
https://blog.csdn.net/a632189007/article/details/78801166/
https://segmentfault.com/a/1190000016805400?utm_source=tag-newest
来源:CSDN
作者:galen6
链接:https://blog.csdn.net/zhengbin6072/article/details/89468441