cudaGetDevice() 失败。 状态：CUDA 驱动程序版本对于 CUDA 运行时版本不足（在 GCP - 在 VNC GUI 中运行的 Ubuntu 18 LTS）

Question

I have followed this link to get all my versions in order:

I need to use Tensorflow_GPU_1.14.0 for legacy code purposes.

Therefore:

tensorflow_gpu-1.14.0
Python: 2.7, 3.3-3.7
cuDnn: 7.4
Cuda: 10.0

import tensorflow as tf; print(tf.__version__)

returns 1.14.0

My nvcc Version is:

nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Sat_Aug_25_21:08:01_CDT_2018
Cuda compilation tools, release 10.0, V10.0.130

My nvidia-smi is:

nvidia-smi
Wed Jan 22 16:47:10 2020       
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 410.104      Driver Version: 410.104      CUDA Version: 10.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|===============================+======================+======================|
|   0  Tesla K80           Off  | 00000000:00:04.0 Off |                    0 |
| N/A   47C    P8    31W / 149W |     27MiB / 11441MiB |      0%      Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes:                                                       GPU Memory |
|  GPU       PID   Type   Process name                             Usage      |
|=============================================================================|
|    0      2524      G   /usr/lib/xorg/Xorg                             9MiB |
|    0      2574      G   /usr/bin/gnome-shell                           6MiB |
+-----------------------------------------------------------------------------+

And my Cudnn version is:

cat /usr/include/cudnn.h | grep CUDNN_MAJOR -A 2
#define CUDNN_MAJOR 7
#define CUDNN_MINOR 4
#define CUDNN_PATCHLEVEL 2
--
#define CUDNN_VERSION (CUDNN_MAJOR * 1000 + CUDNN_MINOR * 100 + CUDNN_PATCHLEVEL)
#include "driver_types.h"

My python version:

import sys

print(sys.version)
3.6.10 |Anaconda, Inc.| (default, Jan  7 2020, 21:14:29) 
[GCC 7.3.0]

I am using Ubuntu 18.04 LTS

Answer 1

To expand on Robert answer in the comment section below, I first uninstalled the existing nvidia by using

sudo apt-get purge nvidia-*

Then installed the newest version using

sudo add-apt-repository ppa:graphics-drivers/ppa
sudo apt update
sudo apt upgrade
ubuntu-drivers list

Then from the list I choose

sudo apt install nvidia-driver-VERSION_NUMBER_HERE

Then I sudo reboot my instance, and executed the following code in Jupyter notebook.

import tensorflow as tf
print(('Is your GPU available for use?\n{0}').format(
    'Yes, your GPU is available: True' if tf.test.is_gpu_available() == True else 'No, your GPU is NOT available: False'
))

print(('\nYour devices that are available:\n{0}').format(
    [device.name for device in tf.config.experimental.list_physical_devices()]
))

And it worked