我需要检查我的tensorflow版本在计算过程中是否使用了gpu。
我按照以下链接中的信息
https://www.tensorflow.org/programmers_guide/using_gpu
但是，当执行此操作时，
# Creates a graph.
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)
# Creates a session with log_device_placement set to True.
sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
# Runs the op.
print(sess.run(c))
输出应如下
Device mapping:
/job:localhost/replica:0/task:0/device:GPU:0 -> device: 0, name: Tesla K40c, pci bus
id: 0000:05:00.0
b: /job:localhost/replica:0/task:0/device:GPU:0
a: /job:localhost/replica:0/task:0/device:GPU:0
MatMul: /job:localhost/replica:0/task:0/device:GPU:0
[[ 22. 28.]
[ 49. 64.]]
但我只是输出，
[[ 22. 28.]
[ 49. 64.]]
这是否意味着GPU没有使用？
请咨询
答案 0 :(得分：0)
从头开始创建一个新的python会话，当我运行你的代码看起来不错时，我怀疑你是按照预期在GPU上运行的。这是我的输出：
您还可以配置代码以查看CPU和GPU之间的实际移动（chrome profiler页面右上角有一个复选框，可以查看CPU / GPU之间的移动）。设置分析非常简单：
https://towardsdatascience.com/howto-profile-tensorflow-1a49fb18073d
$ python
Python 3.6.3 |Anaconda, Inc.| (default, Oct 13 2017, 12:02:49)
[GCC 7.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
/home/dfparksucscedu/data/anaconda3/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6
return f(*args, **kwds)
>>> # Creates a graph.
... 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)
>>> # Creates a session with log_device_placement set to True.
... sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
2018-01-27 09:25:12.422108: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2018-01-27 09:25:12.730164: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Found device 0 with properties:
name: Tesla M40 24GB major: 5 minor: 2 memoryClockRate(GHz): 1.112
pciBusID: 0000:85:00.0
totalMemory: 22.40GiB freeMemory: 21.12GiB
2018-01-27 09:25:13.016210: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1030] Found device 1 with properties:
name: Tesla M40 24GB major: 5 minor: 2 memoryClockRate(GHz): 1.112
pciBusID: 0000:8d:00.0
totalMemory: 22.40GiB freeMemory: 22.19GiB
2018-01-27 09:25:13.016437: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1045] Device peer to peer matrix
2018-01-27 09:25:13.016666: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1051] DMA: 0 1
2018-01-27 09:25:13.016686: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1061] 0: Y Y
2018-01-27 09:25:13.016694: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1061] 1: Y Y
2018-01-27 09:25:13.016821: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:0) -> (device: 0, name: Tesla M40 24GB, pci bus id: 0000:85:00.0, compute capability: 5.2)
2018-01-27 09:25:13.016831: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1120] Creating TensorFlow device (/device:GPU:1) -> (device: 1, name: Tesla M40 24GB, pci bus id: 0000:8d:00.0, compute capability: 5.2)
Device mapping:
/job:localhost/replica:0/task:0/device:GPU:0 -> device: 0, name: Tesla M40 24GB, pci bus id: 0000:85:00.0, compute capability: 5.2
/job:localhost/replica:0/task:0/device:GPU:1 -> device: 1, name: Tesla M40 24GB, pci bus id: 0000:8d:00.0, compute capability: 5.2
2018-01-27 09:25:13.897699: I tensorflow/core/common_runtime/direct_session.cc:299] Device mapping:
/job:localhost/replica:0/task:0/device:GPU:0 -> device: 0, name: Tesla M40 24GB, pci bus id: 0000:85:00.0, compute capability: 5.2
/job:localhost/replica:0/task:0/device:GPU:1 -> device: 1, name: Tesla M40 24GB, pci bus id: 0000:8d:00.0, compute capability: 5.2
>>> # Runs the op.
... print(sess.run(c))
MatMul: (MatMul): /job:localhost/replica:0/task:0/device:GPU:0
2018-01-27 09:25:13.904011: I tensorflow/core/common_runtime/placer.cc:874] MatMul: (MatMul)/job:localhost/replica:0/task:0/device:GPU:0
b: (Const): /job:localhost/replica:0/task:0/device:GPU:0
2018-01-27 09:25:13.904063: I tensorflow/core/common_runtime/placer.cc:874] b: (Const)/job:localhost/replica:0/task:0/device:GPU:0
a: (Const): /job:localhost/replica:0/task:0/device:GPU:0
2018-01-27 09:25:13.904077: I tensorflow/core/common_runtime/placer.cc:874] a: (Const)/job:localhost/replica:0/task:0/device:GPU:0
[[ 22. 28.]
[ 49. 64.]]
>>>
以下是我在另一个系统上运行的代码示例，其中GPU遇到了一些问题（幸运的是我碰巧遇到了这个问题）。创建会话时请注意CUDA_ERROR_UNKNOWN消息：
python
Python 3.6.3 |Anaconda, Inc.| (default, Oct 13 2017, 12:02:49)
[GCC 7.2.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import tensorflow as tf
/home/davidparks21/opt/anaconda3/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6
return f(*args, **kwds)
>>> # Creates a graph.
... 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)
>>> # Creates a session with log_device_placement set to True.
... sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
2018-01-27 09:24:51.640349: I tensorflow/core/platform/cpu_feature_guard.cc:137] Your CPU supports instructions that this TensorFlow binary was not compiled to use: SSE4.1 SSE4.2 AVX AVX2 FMA
2018-01-27 09:24:51.655178: E tensorflow/stream_executor/cuda/cuda_driver.cc:406] failed call to cuInit: CUDA_ERROR_UNKNOWN
2018-01-27 09:24:51.655250: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:158] retrieving CUDA diagnostic information for host: ghostmint
2018-01-27 09:24:51.655261: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:165] hostname: ghostmint
2018-01-27 09:24:51.655317: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:189] libcuda reported version is: 375.66.0
2018-01-27 09:24:51.655351: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:369] driver version file contents: """NVRM version: NVIDIA UNIX x86_64 Kernel Module 375.66 Mon May 1 15:29:16 PDT 2017
GCC version: gcc version 5.4.1 20160904 (Ubuntu 5.4.1-2ubuntu1~16.04)
"""
2018-01-27 09:24:51.655399: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:193] kernel reported version is: 375.66.0
2018-01-27 09:24:51.655411: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:300] kernel version seems to match DSO: 375.66.0
Device mapping: no known devices.
2018-01-27 09:24:51.656138: I tensorflow/core/common_runtime/direct_session.cc:299] Device mapping:
>>> # Runs the op.
... print(sess.run(c))
MatMul: (MatMul): /job:localhost/replica:0/task:0/device:CPU:0
2018-01-27 09:24:52.385483: I tensorflow/core/common_runtime/placer.cc:874] MatMul: (MatMul)/job:localhost/replica:0/task:0/device:CPU:0
b: (Const): /job:localhost/replica:0/task:0/device:CPU:0
2018-01-27 09:24:52.385507: I tensorflow/core/common_runtime/placer.cc:874] b: (Const)/job:localhost/replica:0/task:0/device:CPU:0
a: (Const): /job:localhost/replica:0/task:0/device:CPU:0
2018-01-27 09:24:52.385517: I tensorflow/core/common_runtime/placer.cc:874] a: (Const)/job:localhost/replica:0/task:0/device:CPU:0
[[ 22. 28.]
[ 49. 64.]]
>>> quit()