GPU tensorflow运行问题

Question

I have installed the GPU tensorflow as the tutorial . The relative environment is:

Ubuntu 14 + cuda8.0 + cudnn5.0 + GeForce Gt 640 (OEM)

The problem is:

E tensorflow/stream_executor/cuda/cuda_driver.cc:491] failed call to cuInit: CUDA_ERROR_NO_DEVICE
I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:153] retrieving CUDA diagnostic information for host: wangyajie
I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:160] hostname: wangyajie
I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:185] libcuda reported version is: 352.63.0
I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:356] driver version file contents: """NVRM version: NVIDIA UNIX x86_64 Kernel Module  340.98  Mon Sep 19 17:31:03 PDT 2016
GCC version:  gcc version 4.8.4 (Ubuntu 4.8.4-2ubuntu1~14.04.3) 
"""
I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:189] kernel reported version is: 340.98.0
E tensorflow/stream_executor/cuda/cuda_diagnostics.cc:296] kernel version 340.98.0 does not match DSO version 352.63.0 -- cannot find working devices in this configuration

What can I do to fix it? Should I update the DSO? If should, what command should be used?

How to check weather I install the cuda and cudnn correctly? What command should be used?

Thanks in advance! ZhQ

Answer 1

No idea how this happened, but your error message states that your NVidia driver kernel module is a different version than CUDA requires. (Normally, if you install CUDA via apt-get, it checks for this.) To fix this, you should update your driver to 352.63.0. Instructions can be found here:

https://help.ubuntu.com/community/BinaryDriverHowto/Nvidia

However, this may not completely fix your problem, since it may prove difficult to get both CUDA and the driver onto the same version if you install them separately. Instead, you should consider to install both from the same repo, thus re-installing CUDA as well.

Generally, if possible for you, I'd recommend updating to Ubuntu 16.04 first, since CUDA is much less painful to install there (in my experience, at least).