cuInit CUDA_ERROR_NOT_INITIALIZED调用失败（设备映射：无已知设备）Tensorflow-gpu 1.2.1

Question

我有2个GPU，但无法通过Tensorflow-GPU使用，有人可以帮助解决问题。

版本：

• Tensorflow GPU 1.2.1
• Cudnn 6.0.21

• Cuda 10.0（工具包）

  import tensorflow as tf a = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[2, 3], name='a') b = tf.constant([1.0, 2.0, 3.0, 4.0, 5.0, 6.0], shape=[3, 2], name='b') c = tf.matmul(a, b) with tf.Session(config=tf.ConfigProto(log_device_placement=True)) as sess: print (sess.run(c))

警告/错误消息：

2018-11-09 12:06:42.827220: E tensorflow/stream_executor/cuda/cuda_driver.cc:406] **failed call to cuInit: CUDA_ERROR_NOT_INITIALIZED**
2018-11-09 12:06:42.827292: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:158] retrieving CUDA diagnostic information for host: ca47dppaia001
2018-11-09 12:06:42.827309: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:165] hostname: ca47dppaia001
2018-11-09 12:06:42.827370: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:189] libcuda reported version is: 410.72.0
2018-11-09 12:06:42.827419: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:369] driver version file contents: """NVRM version: NVIDIA UNIX ppc64le Kernel Module  410.72  Wed Oct 17 20:19:50 CDT 2018
GCC version:  gcc version 4.8.5 20150623 (Red Hat 4.8.5-36) (GCC)
"""
2018-11-09 12:06:42.827457: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:193] kernel reported version is: 410.72.0
2018-11-09 12:06:42.827471: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:300] kernel version seems to match DSO: 410.72.0
**Device mapping: no known devices.**
2018-11-09 12:06:42.871454: I tensorflow/core/common_runtime/direct_session.cc:265] Device mapping:

---