【Linux】在一台机器上同时安装多个版本的CUDA（切换CUDA版本）

TangPlusHPC

已于 2022-06-11 20:47:45 修改

阅读量1.9w

点赞数 83

分类专栏： Linux 环境配置文章标签： linux tensorflow keras

于 2022-06-11 09:11:08 首次发布

本文链接：https://blog.csdn.net/tangjiahao10/article/details/125228143

版权

一、前言

正如题目所言，最近笔者要跑一个TensorFlow搭建的模型，等我按照要求将对应版本的TensorFlow和Keras安装好之后，发现训练模型巨慢，GPU显存只用了一点点而且利用率一直是零，而且提示找不到一些库，提示如下。

2022-06-10 13:06:14.299058: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcudart.so.10.0'; dlerror: libcudart.so.10.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/cuda-10.0/lib64:/usr/local/cuda-10.0/extras/CUPTI/lib64
2022-06-10 13:06:14.299110: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcublas.so.10.0'; dlerror: libcublas.so.10.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/cuda-10.0/lib64:/usr/local/cuda-10.0/extras/CUPTI/lib64
2022-06-10 13:06:14.299155: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcufft.so.10.0'; dlerror: libcufft.so.10.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/cuda-10.0/lib64:/usr/local/cuda-10.0/extras/CUPTI/lib64
2022-06-10 13:06:14.299198: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcurand.so.10.0'; dlerror: libcurand.so.10.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/cuda-10.0/lib64:/usr/local/cuda-10.0/extras/CUPTI/lib64
2022-06-10 13:06:14.299239: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcusolver.so.10.0'; dlerror: libcusolver.so.10.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/cuda-10.0/lib64:/usr/local/cuda-10.0/extras/CUPTI/lib64
2022-06-10 13:06:14.299281: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcusparse.so.10.0'; dlerror: libcusparse.so.10.0: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/cuda-10.0/lib64:/usr/local/cuda-10.0/extras/CUPTI/lib64
2022-06-10 13:06:14.299326: I tensorflow/stream_executor/platform/default/dso_loader.cc:53] Could not dlopen library 'libcudnn.so.7'; dlerror: libcudnn.so.7: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/local/cuda-10.0/lib64:/usr/local/cuda-10.0/extras/CUPTI/lib64
2022-06-10 13:06:14.299336: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1663] Cannot dlopen some GPU libraries. Skipping registering GPU devices...
2022-06-10 13:06:14.299421: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:

观察提示信息和一些现象，笔者得出结论，是CUDA和cuDNN版本没有装合适，因为该程序会去/usr/local/cuda-10.0/lib64文件夹下找库，但是我就没有装CUDA 10.0。去网上找了一番资料后，笔者发现果然是CUDA和cuDNN的版本问题，TensorFlow版本与CUDA