显卡环境

### 环境相关整理

~~ ```python

~~ ================================================

~~ # 显卡版本

~~ GeForce GTX 1080 Ti

~~ # 查看方法

~~ lspci |grep VGA

~~ ====================================================

~~ # 编译版本（最低4.8）

~~ gcc version 5.4.0

~~ # 安装指令

~~ sudo apt-get install gcc-4.8

~~ # 查看方法

~~ gcc -v

~~ # 查看全部gcc版本

~~ ls /usr/bin/gcc*

~~ # 详细对应版本

~~ https://www.tensorflow.org/install/source?hl=zh-cn

~~ =====================================================

~~ # python版本（3.6.7）

~~ sudo add-apt-repository ppa:jonathonf/python-3.6.7

~~ sudo apt-get update

~~ sudo apt-get install python3.6.7

~~ # python路径

~~ /usr/bin/python3.6

~~ ===========================

~~ # pip版本

~~ curl https://bootstrap.pypa.io/get-pip.py | sudo python3.6

~~ # pip 路径

~~ sudo /home/bigdata/.local/bin/pip install

~~ # pip 源配置

~~ vim ~/.pip/pip.conf

~~

~~ [global]

~~ index-url = https://pypi.tuna.tsinghua.edu.cn/simple

~~ [install]

~~ trusted-host=pypi.tuna.tsinghua.edu.cn

~~ ========================================

~~ # 库版本

~~ sudo /home/bigdata/.local/bin/pip install bert-serving-server

~~ sudo /home/bigdata/.local/bin/pip install bert-serving-client

~~ sudo /home/bigdata/.local/bin/pip install tensorflow-gpu==1.12

~~ =============================================

~~ # nvidia-docker 环境

~~ # 环境测试

~~ docker run --runtime=nvidia --rm nvidia/cuda:9.0-base nvidia-smi

~~ ============================================

~~ # nvidia driver 版本

~~ NVIDIA UNIX x86_64 Kernel Module  418.39

~~ # 安装指令

~~ https://blog.csdn.net/breeze5428/article/details/80013753

~~ # 查看方法

~~ cat /proc/driver/nvidia/version

~~ =============================================

~~ # cuda 版本

~~ CUDA Version: 10.1

~~ # 下载安装cuda 9.0

~~ curl -LO https://developer.nvidia.com/compute/cuda/9.0/Prod/local_installers/cuda-repo-ubuntu1604-9-0-local_9.0.176-1_amd64-deb

~~ sudo dpkg -i cuda-repo-ubuntu1604-9-0-local_9.0.176-1_amd64.deb

~~ sudo apt-key add /var/cuda-repo-9-0-local/7fa2af80.pub

~~ sudo apt-get update

~~ sudo apt-get install cuda

~~ # 查看方法

~~ cat /usr/local/cuda/version.txt

~~ ==============================================

~~ # cudnn 版本

~~ CUDNN_MAJOR 7

~~ # cudnn 安装

~~

~~ # 查看方法

~~ cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2

~~ ==============================================

~~ ```

### 错误信息整理

~~ ```

~~ # 错误

~~ ImportError: libcublas.so.9.0: cannot open shared object file: No such file or directory

~~ # 检查

~~ ls /usr/local/cuda/lib64/

~~

~~ ```