# Failed call to cuInit: CUDA_ERROR_UNKNOWN

## post by sabalaba on Nov 3, 2017

[sabalaba](https://deeptalk.lambda.ai/u/sabalaba)

The following python code

```java
import tensorflow as tf
s = tf.Session()
```

Causes the following error to appear:

```auto
2017-11-03 12:34:59.694469: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations.
2017-11-03 12:34:59.694511: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations.
2017-11-03 12:34:59.694522: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations.
2017-11-03 12:34:59.694531: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations.
2017-11-03 12:34:59.694542: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.
2017-11-03 12:34:59.708269: E tensorflow/stream_executor/cuda/cuda_driver.cc:406] failed call to cuInit: CUDA_ERROR_UNKNOWN
2017-11-03 12:34:59.708334: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:158] retrieving CUDA diagnostic information for host: hostname
2017-11-03 12:34:59.708349: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:165] hostname: hostname
2017-11-03 12:34:59.708408: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:189] libcuda reported version is: 384.90.0
2017-11-03 12:34:59.709302: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:369] driver version file contents: '''NVRM version: NVIDIA UNIX x86_64 Kernel Module  384.90  Tue Sep 19 19:17:35 PDT 2017 GCC version:  gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.4)'''
2017-11-03 12:34:59.709341: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:193] kernel reported version is: 384.90.0
2017-11-03 12:34:59.709353: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:300] kernel version seems to match DSO: 384.90.0
```

The main issue being this Error:

```bash
tensorflow/stream_executor/cuda/cuda_driver.cc:406] failed call to cuInit: CUDA_ERROR_UNKNOWN
```

After you verify you have a working CUDA toolkit installation, installing nvidia-modprobe package fixes this issue.

```bash
sudo apt-get install nvidia-modprobe
```