[DL] Tensorflow와 Cuda Version dependency

앞전에 Deep Learning Server 구현할 때 발생했던 문제점이 몇 가지 있었습니다.

 

그중 Tensorflow와 Cuda의 버전이 맞지 않아 발생하는 문제가 있었네요.

 

Tensorflow 공식 홈에서의 내용을 보면 Tensor GPU 2.0 = CUDA 10.0 버전이 필요합니다.

 

 

필자는 Tensorflow GPU 2.0 버전을 사용하려고 하는 이유였고 ( 쓰고 있는 랩탑 Mac에서는 Tensor 2.0까지 )

 

최신 nvidia 드라이버를 설치하니 Cuda 10.2 버전이 자연스래 따라와서 Jupyter Notebook 서버를 사용할 때의 문제점이 발생했었습니다.

 

2020-04-06 12:01:23.253018: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic 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-10.1/lib64
2020-04-06 12:01:23.253149: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic 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-10.1/lib64
2020-04-06 12:01:23.253261: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic 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-10.1/lib64
2020-04-06 12:01:23.253373: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic 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-10.1/lib64
2020-04-06 12:01:23.253481: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic 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-10.1/lib64
2020-04-06 12:01:23.253587: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic 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-10.1/lib64
2020-04-06 12:01:23.259616: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-04-06 12:01:23.259663: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1641] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.
Skipping registering GPU devices…
2020-04-06 12:01:23.259702: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1159] Device interconnect StreamExecutor with strength 1 edge matrix:
2020-04-06 12:01:23.259721: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1165]      0
2020-04-06 12:01:23.259737: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1178] 0:   N

 

총 6개의 파일이 10.0 버전이 없다는 것인데 설치 되어있던 Cuda 관련 File 들을 삭제하고 Cuda 10.0 버전으로 재 설치를 했습니다.

libcudart.so.10.0
libcublas.so.10.0
libcufft.so.10.0
libcurand.so.10.0
libcusolver.so.10.0
libcusparse.so.10.0

 

Cuda 10.2 -> 10.0 버전으로 Down 후에는 위와 같은 에러는 발생하지 않았네요.

 

 

이전 포스팅을 보시며 도움이 되셨으면 합니다.

hiwony.tistory.com/3

 

[DL] TensorFlow 2.0 Ubuntu 18.04 Server Setting