blocks|key|4408413|text|不，我认为“开放CUDA库”不足以说明问题，因为图形的不同节点可能在不同的设备上。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408414|使用tensorflow2时：|4408415|print("Num+GPUs+Available:+",+len(tf.config.list_physical_devices('GPU')))|code-block|syntax|javascript|4408416|对于tensorflow1，要找出使用了哪个设备，您可以像这样启用日志设备放置：|4408417|sess+=+tf.Session(config=tf.ConfigProto(log_device_placement=True))|4408418|检查您的控制台是否有此类型的输出。|4408419|​|4408420|📷|atomic|offset|length|4408421|4408422|entityMap|0|IMAGE|mutability|IMMUTABLE|imageUrl|https://ask.qcloudimg.com/http-save/yehe-900000/14acff2ab711a8f60ae2082f052b8538.png|imageAlt^0|0|0|0|0|0|0|0|0|1|0|0|0^^$0|@$1|2|3|4|5|6|7|15|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|16|8|@]|9|@]|A|$]]|$1|D|3|E|5|F|7|17|8|@]|9|@]|A|$G|H]]|$1|I|3|J|5|6|7|18|8|@]|9|@]|A|$]]|$1|K|3|L|5|F|7|19|8|@]|9|@]|A|$G|H]]|$1|M|3|N|5|6|7|1A|8|@]|9|@]|A|$]]|$1|O|3|P|5|6|7|1B|8|@]|9|@]|A|$]]|$1|Q|3|R|5|S|7|1C|8|@]|9|@$T|1D|U|1E|1|1F]]|A|$]]|$1|V|3|P|5|6|7|1G|8|@]|9|@]|A|$]]|$1|W|3|-4|5|6|7|1H|8|@]|9|@]|A|$]]]|X|$Y|$5|Z|10|11|A|$12|13|14|-4]]]]

No, I don't think &quot;open CUDA library&quot; is enough to tell, because different nodes of the graph may be on different devices.
When using tensorflow2:
<pre><code>print(&quot;Num GPUs Available: &quot;, len(tf.config.list_physical_devices('GPU')))
</code></pre>
For tensorflow1, to find out which device is used, you can enable log device placement like this:
<pre><code>sess = tf.Session(config=tf.ConfigProto(log_device_placement=True))
</code></pre>
Check your console for this type of output.
<a href="https://i.stack.imgur.com/RtRiB.png" rel="noreferrer"><img src="https://i.stack.imgur.com/RtRiB.png" alt="" /></a>

blocks|key|4408279|text|我更喜欢使用nvidia-smi来监控GPU的使用情况。如果在启动程序时它显著上升，这是tensorflow正在使用GPU的强烈信号。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408280|entityMap^0|0^^$0|@$1|2|3|4|5|6|7|D|8|@]|9|@]|A|$]]|$1|B|3|-4|5|6|7|E|8|@]|9|@]|A|$]]]|C|$]]

I prefer to use nvidia-smi to monitor GPU usage. if it goes up significantly when you start you program, it's a strong sign that your tensorflow is using GPU.

blocks|key|4408345|text|这应该会给出Tensorflow可用的设备列表(在Py-3.6下)：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408346|tf+=+tf.Session(config=tf.ConfigProto(log_device_placement=True))
tf.list_devices()
#+_DeviceAttributes(/job:localhost/replica:0/task:0/device:CPU:0,+CPU,+268435456)|code-block|syntax|javascript|4408347|entityMap^0|0|0^^$0|@$1|2|3|4|5|6|7|I|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|J|8|@]|9|@]|A|$E|F]]|$1|G|3|-4|5|6|7|K|8|@]|9|@]|A|$]]]|H|$]]

This should give the list of devices available for Tensorflow (under Py-3.6):

<pre><code>tf = tf.Session(config=tf.ConfigProto(log_device_placement=True))
tf.list_devices()
# _DeviceAttributes(/job:localhost/replica:0/task:0/device:CPU:0, CPU, 268435456)
</code></pre>

blocks|key|4408545|text|我发现从命令行查询gpu是最简单的：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408546|nvidia-smi

%2B-----------------------------------------------------------------------------%2B
%7C+NVIDIA-SMI+384.98+++++++++++++++++Driver+Version:+384.98++++++++++++++++++++%7C
%7C-------------------------------%2B----------------------%2B----------------------%2B
%7C+GPU++Name++++++++Persistence-M%7C+Bus-Id++++++++Disp.A+%7C+Volatile+Uncorr.+ECC+%7C
%7C+Fan++Temp++Perf++Pwr:Usage/Cap%7C+++++++++Memory-Usage+%7C+GPU-Util++Compute+M.+%7C
%7C===============================%2B======================%2B======================%7C
%7C+++0++GeForce+GTX+980+Ti++Off++%7C+00000000:02:00.0++On+%7C++++++++++++++++++N/A+%7C
%7C+22%25+++33C++++P8++++13W+/+250W+%7C+++5817MiB+/++6075MiB+%7C++++++0%25++++++Default+%7C
%2B-------------------------------%2B----------------------%2B----------------------%2B

%2B-----------------------------------------------------------------------------%2B
%7C+Processes:+++++++++++++++++++++++++++++++++++++++++++++++++++++++GPU+Memory+%7C
%7C++GPU+++++++PID+++Type+++Process+name+++++++++++++++++++++++++++++Usage++++++%7C
%7C=============================================================================%7C
%7C++++0++++++1060++++++G+++/usr/lib/xorg/Xorg++++++++++++++++++++++++++++53MiB+%7C
%7C++++0+++++25177++++++C+++python++++++++++++++++++++++++++++++++++++++5751MiB+%7C
%2B-----------------------------------------------------------------------------%2B|code-block|syntax|javascript|4408547|如果您的学习是后台进程，则来自jobs+-p的pid应与来自nvidia-smi的pid匹配|offset|length|style|CODE|4408548|entityMap^0|0|0|F|7|U|A|0^^$0|@$1|2|3|4|5|6|7|O|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|P|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|Q|8|@$I|R|J|S|K|L]|$I|T|J|U|K|L]]|9|@]|A|$]]|$1|M|3|-4|5|6|7|V|8|@]|9|@]|A|$]]]|N|$]]

I find just querying the gpu from the command line is easiest:

<pre><code>nvidia-smi

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 384.98 Driver Version: 384.98 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 980 Ti Off | 00000000:02:00.0 On | N/A |
| 22% 33C P8 13W / 250W | 5817MiB / 6075MiB | 0% Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1060 G /usr/lib/xorg/Xorg 53MiB |
| 0 25177 C python 5751MiB |
+-----------------------------------------------------------------------------+
</code></pre>

if your learning is a background process the pid from 
<code>jobs -p</code> should match the pid from <code>nvidia-smi</code>

blocks|key|4408425|text|我认为有一种更简单的方法来实现这一点。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408426|import+tensorflow+as+tf
if+tf.test.gpu_device_name():
++++print('Default+GPU+Device:+{}'.format(tf.test.gpu_device_name()))
else:
++++print("Please+install+GPU+version+of+TF")|code-block|syntax|javascript|4408427|它通常打印如下|4408428|Default+GPU+Device:+/device:GPU:0|4408429|对我来说，这似乎比那些冗长的日志更容易。|4408430|编辑：-这是针对TF+1.x版本进行的测试。我从来没有机会使用TF+2.0或更高版本做一些事情，所以请记住。|4408431|entityMap^0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|Q|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|R|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|S|8|@]|9|@]|A|$]]|$1|I|3|J|5|D|7|T|8|@]|9|@]|A|$E|F]]|$1|K|3|L|5|6|7|U|8|@]|9|@]|A|$]]|$1|M|3|N|5|6|7|V|8|@]|9|@]|A|$]]|$1|O|3|-4|5|6|7|W|8|@]|9|@]|A|$]]]|P|$]]

I think there is an easier way to achieve this.
<pre><code>import tensorflow as tf
if tf.test.gpu_device_name():
 print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))
else:
 print(&quot;Please install GPU version of TF&quot;)
</code></pre>
It usually prints like
<pre><code>Default GPU Device: /device:GPU:0
</code></pre>
This seems easier to me rather than those verbose logs.
Edit:-
This was tested for TF 1.x versions. I never had a chance to do stuff with TF 2.0 or above so keep in mind.

blocks|key|4408515|text|除了其他答案之外，以下内容还可以帮助您确保您的tensorflow版本包含GPU支持。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408516|import+tensorflow+as+tf
print(tf.test.is_built_with_cuda())|code-block|syntax|javascript|4408517|entityMap^0|0|0^^$0|@$1|2|3|4|5|6|7|I|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|J|8|@]|9|@]|A|$E|F]]|$1|G|3|-4|5|6|7|K|8|@]|9|@]|A|$]]]|H|$]]

In addition to other answers, the following should help you to make sure that your version of tensorflow includes GPU support.

<pre><code>import tensorflow as tf
print(tf.test.is_built_with_cuda())
</code></pre>

blocks|key|4408564|text|使用Tensorflow的最新更新，您可以按如下方式进行检查：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408565|tf.test.is_gpu_available(+cuda_only=False,+min_cuda_compute_capability=None)|code-block|syntax|javascript|4408566|如果Tensorflow正在使用GPU，则返回True，否则返回False。|offset|length|style|CODE|4408567|如果你需要设备device_name，你可以输入：tf.test.gpu_device_name()。从here获取更多详细信息|4408568|entityMap|0|LINK|mutability|MUTABLE|url|https://www.tensorflow.org/api_docs/python/tf/test^0|0|0|2|A|N|4|W|5|0|7|B|P|P|1G|4|0|0^^$0|@$1|2|3|4|5|6|7|W|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|X|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|Y|8|@$I|Z|J|10|K|L]|$I|11|J|12|K|L]|$I|13|J|14|K|L]]|9|@]|A|$]]|$1|M|3|N|5|6|7|15|8|@$I|16|J|17|K|L]|$I|18|J|19|K|L]]|9|@$I|1A|J|1B|1|1C]]|A|$]]|$1|O|3|-4|5|6|7|1D|8|@]|9|@]|A|$]]]|P|$Q|$5|R|S|T|A|$U|V]]]]

With the recent updates of Tensorflow, you can check it as follow : 

<pre><code>tf.test.is_gpu_available( cuda_only=False, min_cuda_compute_capability=None)
</code></pre>

This will return <code>True</code> if GPU is being used by <code>Tensorflow</code>, and return <code>False</code> otherwise. 

If you want device <code>device_name</code> you can type : <code>tf.test.gpu_device_name()</code>.
Get more details from <a href="https://www.tensorflow.org/api_docs/python/tf/test" rel="noreferrer">here</a>

blocks|key|4408632|text|这是我用来直接从bash列出tf.session可用的设备的行：|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|4408633|python+-c+"import+os;+os.environ['TF_CPP_MIN_LOG_LEVEL']+=+'3';+import+tensorflow+as+tf;+sess+=+tf.Session();+[print(x)+for+x+in+sess.list_devices()];+print(tf.__version__);"|code-block|syntax|javascript|4408634|它将打印可用的设备和tensorflow版本，例如：|4408635|_DeviceAttributes(/job:localhost/replica:0/task:0/device:CPU:0,+CPU,+268435456,+10588614393916958794)
_DeviceAttributes(/job:localhost/replica:0/task:0/device:XLA_GPU:0,+XLA_GPU,+17179869184,+12320120782636586575)
_DeviceAttributes(/job:localhost/replica:0/task:0/device:XLA_CPU:0,+XLA_CPU,+17179869184,+13378821206986992411)
_DeviceAttributes(/job:localhost/replica:0/task:0/device:GPU:0,+GPU,+32039954023,+12481654498215526877)
1.14.0|4408636|entityMap^0|E|A|0|0|0|0^^$0|@$1|2|3|4|5|6|7|Q|8|@$9|R|A|S|B|C]]|D|@]|E|$]]|$1|F|3|G|5|H|7|T|8|@]|D|@]|E|$I|J]]|$1|K|3|L|5|6|7|U|8|@]|D|@]|E|$]]|$1|M|3|N|5|H|7|V|8|@]|D|@]|E|$I|J]]|$1|O|3|-4|5|6|7|W|8|@]|D|@]|E|$]]]|P|$]]

This is the line I am using to list devices available to <code>tf.session</code> directly from bash:

<pre><code>python -c "import os; os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'; import tensorflow as tf; sess = tf.Session(); [print(x) for x in sess.list_devices()]; print(tf.__version__);"
</code></pre>

It will print available devices and tensorflow version, for example:

<pre><code>_DeviceAttributes(/job:localhost/replica:0/task:0/device:CPU:0, CPU, 268435456, 10588614393916958794)
_DeviceAttributes(/job:localhost/replica:0/task:0/device:XLA_GPU:0, XLA_GPU, 17179869184, 12320120782636586575)
_DeviceAttributes(/job:localhost/replica:0/task:0/device:XLA_CPU:0, XLA_CPU, 17179869184, 13378821206986992411)
_DeviceAttributes(/job:localhost/replica:0/task:0/device:GPU:0, GPU, 32039954023, 12481654498215526877)
1.14.0
</code></pre>

blocks|key|4408833|text|下面还将返回您的GPU设备的名称。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408834|import+tensorflow+as+tf
tf.test.gpu_device_name()|code-block|syntax|javascript|4408835|entityMap^0|0|0^^$0|@$1|2|3|4|5|6|7|I|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|J|8|@]|9|@]|A|$E|F]]|$1|G|3|-4|5|6|7|K|8|@]|9|@]|A|$]]]|H|$]]

The following will also return the name of your GPU devices.

<pre><code>import tensorflow as tf
tf.test.gpu_device_name()
</code></pre>

blocks|key|4408857|text|我发现下面的代码片段用来测试gpu非常方便。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408858|Tensorflow+2.0测试|4408859|import+tensorflow+as+tf
import+tensorflow.compat.v1+as+tf
tf.disable_v2_behavior()
with+tf.device('/gpu:0'):
++++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()+as+sess:
++++print+(sess.run(c))|code-block|syntax|javascript|4408860|Tensorflow+1测试|4408861|import+tensorflow+as+tf
with+tf.device('/gpu:0'):
++++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()+as+sess:
++++print+(sess.run(c))|4408862|entityMap^0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|O|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|P|8|@]|9|@]|A|$]]|$1|D|3|E|5|F|7|Q|8|@]|9|@]|A|$G|H]]|$1|I|3|J|5|6|7|R|8|@]|9|@]|A|$]]|$1|K|3|L|5|F|7|S|8|@]|9|@]|A|$G|H]]|$1|M|3|-4|5|6|7|T|8|@]|9|@]|A|$]]]|N|$]]

I found below snippet is very handy to test the gpu .. 

<h3>Tensorflow 2.0 Test</h3>

<pre><code>import tensorflow as tf
import tensorflow.compat.v1 as tf
tf.disable_v2_behavior()
with tf.device('/gpu:0'):
 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() as sess:
 print (sess.run(c))
</code></pre>

<h3>Tensorflow 1 Test</h3>

<pre class="lang-py prettyprint-override"><code>import tensorflow as tf
with tf.device('/gpu:0'):
 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() as sess:
 print (sess.run(c))
</code></pre>

blocks|key|4408901|text|如果您使用的是TensorFlow+2.0，则可以使用此for循环来显示设备：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408902|with+tf.compat.v1.Session()+as+sess:
++devices+=+sess.list_devices()
devices|code-block|syntax|javascript|4408903|entityMap^0|0|0^^$0|@$1|2|3|4|5|6|7|I|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|J|8|@]|9|@]|A|$E|F]]|$1|G|3|-4|5|6|7|K|8|@]|9|@]|A|$]]]|H|$]]

If you are using TensorFlow 2.0, you can use this for loop to show the devices:

<pre><code>with tf.compat.v1.Session() as sess:
 devices = sess.list_devices()
devices
</code></pre>

blocks|key|4408932|text|如果您使用的是tensorflow+2.x，请使用：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408933|sess+=+tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(log_device_placement=True))|code-block|syntax|javascript|4408934|entityMap^0|0|0^^$0|@$1|2|3|4|5|6|7|I|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|J|8|@]|9|@]|A|$E|F]]|$1|G|3|-4|5|6|7|K|8|@]|9|@]|A|$]]]|H|$]]

if you are using tensorflow 2.x use:

<pre><code>sess = tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(log_device_placement=True))
</code></pre>

blocks|key|4408977|text|对于Tensorflow+2.0|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408978|import+tensorflow+as+tf

tf.test.is_gpu_available(
++++cuda_only=False,
++++min_cuda_compute_capability=None
)|code-block|syntax|javascript|4408979|源here|offset|length|4408980|其他选项是：|4408981|tf.config.experimental.list_physical_devices('GPU')|4408982|entityMap|0|LINK|mutability|MUTABLE|url|https://www.tensorflow.org/api_docs/python/tf/test/is_gpu_available?version=stable^0|0|0|1|4|0|0|0|0^^$0|@$1|2|3|4|5|6|7|W|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|X|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|Y|8|@]|9|@$I|Z|J|10|1|11]]|A|$]]|$1|K|3|L|5|6|7|12|8|@]|9|@]|A|$]]|$1|M|3|N|5|D|7|13|8|@]|9|@]|A|$E|F]]|$1|O|3|-4|5|6|7|14|8|@]|9|@]|A|$]]]|P|$Q|$5|R|S|T|A|$U|V]]]]

<h3>For Tensorflow 2.0</h3>
<pre><code>import tensorflow as tf

tf.test.is_gpu_available(
 cuda_only=False,
 min_cuda_compute_capability=None
)
</code></pre>
source <a href="https://www.tensorflow.org/api_docs/python/tf/test/is_gpu_available?version=stable" rel="nofollow noreferrer">here</a>
other option is:
<pre><code>tf.config.experimental.list_physical_devices('GPU')
</code></pre>

blocks|key|4408815|text|在Jupyter或您的集成开发环境中运行以下命令，以检查Tensorflow是否使用了图形处理器：tf.config.list_physical_devices('GPU')|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|4408816|entityMap^0|1D|12|0^^$0|@$1|2|3|4|5|6|7|H|8|@$9|I|A|J|B|C]]|D|@]|E|$]]|$1|F|3|-4|5|6|7|K|8|@]|D|@]|E|$]]]|G|$]]

Run this command in Jupyter or your IDE to check if Tensorflow is using a GPU or not: <code>tf.config.list_physical_devices('GPU')</code>

blocks|key|4409042|text|使用tensorflow+2.0+>=|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4409043|import+tensorflow+as+tf
sess+=+tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(log_device_placement=True))|code-block|syntax|javascript|4409044|​|4409045|📷|atomic|offset|length|4409046|4409047|entityMap|0|IMAGE|mutability|IMMUTABLE|imageUrl|https://ask.qcloudimg.com/http-save/yehe-900000/c245598c87122cf796d7fe2b9ba04e01.png|imageAlt^0|0|0|0|0|1|0|0|0^^$0|@$1|2|3|4|5|6|7|X|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|Y|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|Z|8|@]|9|@]|A|$]]|$1|I|3|J|5|K|7|10|8|@]|9|@$L|11|M|12|1|13]]|A|$]]|$1|N|3|H|5|6|7|14|8|@]|9|@]|A|$]]|$1|O|3|-4|5|6|7|15|8|@]|9|@]|A|$]]]|P|$Q|$5|R|S|T|A|$U|V|W|-4]]]]

With tensorflow 2.0 &gt;=
<pre><code>import tensorflow as tf
sess = tf.compat.v1.Session(config=tf.compat.v1.ConfigProto(log_device_placement=True))
</code></pre>
<a href="https://i.stack.imgur.com/VEWI0.png" rel="noreferrer"><img src="https://i.stack.imgur.com/VEWI0.png" alt="enter image description here" /></a>

blocks|key|4408890|text|Tensorflow+2.1|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408891|可以使用nvidia-smi验证GPU上的内存使用情况的简单计算。|4408892|import+tensorflow+as+tf+

c1+=+[]
n+=+10

def+matpow(M,+n):
++++if+n+<+1:+#Abstract+cases+where+n+<+1
++++++++return+M
++++else:
++++++++return+tf.matmul(M,+matpow(M,+n-1))

with+tf.device('/gpu:0'):
++++a+=+tf.Variable(tf.random.uniform(shape=(10000,+10000)),+name="a")
++++b+=+tf.Variable(tf.random.uniform(shape=(10000,+10000)),+name="b")
++++c1.append(matpow(a,+n))
++++c1.append(matpow(b,+n))|code-block|syntax|javascript|4408893|entityMap^0|0|0|0^^$0|@$1|2|3|4|5|6|7|K|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|L|8|@]|9|@]|A|$]]|$1|D|3|E|5|F|7|M|8|@]|9|@]|A|$G|H]]|$1|I|3|-4|5|6|7|N|8|@]|9|@]|A|$]]]|J|$]]

<h1>Tensorflow 2.1</h1>

A simple calculation that can be verified with nvidia-smi for memory usage on the GPU.

<pre><code>import tensorflow as tf 

c1 = []
n = 10

def matpow(M, n):
 if n &lt; 1: #Abstract cases where n &lt; 1
 return M
 else:
 return tf.matmul(M, matpow(M, n-1))

with tf.device('/gpu:0'):
 a = tf.Variable(tf.random.uniform(shape=(10000, 10000)), name="a")
 b = tf.Variable(tf.random.uniform(shape=(10000, 10000)), name="b")
 c1.append(matpow(a, n))
 c1.append(matpow(b, n))
</code></pre>

blocks|key|4408905|text|>>>+import+tensorflow+as+tf+
>>>+tf.config.list_physical_devices('GPU')

2020-05-10+14:58:16.243814:+I+tensorflow/stream_executor/platform/default/dso_loader.cc:44]+Successfully+opened+dynamic+library+libcuda.so.1
2020-05-10+14:58:16.262675:+I+tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981]+successful+NUMA+node+read+from+SysFS+had+negative+value+(-1),+but+there+must+be+at+least+one+NUMA+node,+so+returning+NUMA+node+zero
2020-05-10+14:58:16.263119:+I+tensorflow/core/common_runtime/gpu/gpu_device.cc:1555]+Found+device+0+with+properties:
pciBusID:+0000:01:00.0+name:+GeForce+GTX+1060+6GB+computeCapability:+6.1
coreClock:+1.7715GHz+coreCount:+10+deviceMemorySize:+5.93GiB+deviceMemoryBandwidth:+178.99GiB/s
2020-05-10+14:58:16.263143:+I+tensorflow/stream_executor/platform/default/dso_loader.cc:44]+Successfully+opened+dynamic+library+libcudart.so.10.1
2020-05-10+14:58:16.263188:+I+tensorflow/stream_executor/platform/default/dso_loader.cc:44]+Successfully+opened+dynamic+library+libcublas.so.10
2020-05-10+14:58:16.264289:+I+tensorflow/stream_executor/platform/default/dso_loader.cc:44]+Successfully+opened+dynamic+library+libcufft.so.10
2020-05-10+14:58:16.264495:+I+tensorflow/stream_executor/platform/default/dso_loader.cc:44]+Successfully+opened+dynamic+library+libcurand.so.10
2020-05-10+14:58:16.265644:+I+tensorflow/stream_executor/platform/default/dso_loader.cc:44]+Successfully+opened+dynamic+library+libcusolver.so.10
2020-05-10+14:58:16.266329:+I+tensorflow/stream_executor/platform/default/dso_loader.cc:44]+Successfully+opened+dynamic+library+libcusparse.so.10
2020-05-10+14:58:16.266357:+I+tensorflow/stream_executor/platform/default/dso_loader.cc:44]+Successfully+opened+dynamic+library+libcudnn.so.7
2020-05-10+14:58:16.266478:+I+tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981]+successful+NUMA+node+read+from+SysFS+had+negative+value+(-1),+but+there+must+be+at+least+one+NUMA+node,+so+returning+NUMA+node+zero
2020-05-10+14:58:16.266823:+I+tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981]+successful+NUMA+node+read+from+SysFS+had+negative+value+(-1),+but+there+must+be+at+least+one+NUMA+node,+so+returning+NUMA+node+zero
2020-05-10+14:58:16.267107:+I+tensorflow/core/common_runtime/gpu/gpu_device.cc:1697]+Adding+visible+gpu+devices:+0
[PhysicalDevice(name='/physical_device:GPU:0',+device_type='GPU')]|type|code-block|depth|inlineStyleRanges|entityRanges|data|syntax|javascript|4408906|根据@AmitaiIrron的建议：|unstyled|4408907|此部分指示找到了gpu|4408908|2020-05-10+14:58:16.263119:+I+tensorflow/core/common_runtime/gpu/gpu_device.cc:1555]+Found+device+0+with+properties:

pciBusID:+0000:01:00.0+name:+GeForce+GTX+1060+6GB+computeCapability:+6.1
coreClock:+1.7715GHz+coreCount:+10+deviceMemorySize:+5.93GiB+deviceMemoryBandwidth:+178.99GiB/s|4408909|在这里，它被添加为可用的物理设备|4408910|2020-05-10+14:58:16.267107:+I+tensorflow/core/common_runtime/gpu/gpu_device.cc:1697]+Adding+visible+gpu+devices:+0

[PhysicalDevice(name='/physical_device:GPU:0',+device_type='GPU')]|4408911|entityMap^0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|Q|8|@]|9|@]|A|$B|C]]|$1|D|3|E|5|F|7|R|8|@]|9|@]|A|$]]|$1|G|3|H|5|F|7|S|8|@]|9|@]|A|$]]|$1|I|3|J|5|6|7|T|8|@]|9|@]|A|$B|C]]|$1|K|3|L|5|F|7|U|8|@]|9|@]|A|$]]|$1|M|3|N|5|6|7|V|8|@]|9|@]|A|$B|C]]|$1|O|3|-4|5|F|7|W|8|@]|9|@]|A|$]]]|P|$]]

<pre><code>&gt;&gt;&gt; import tensorflow as tf 
&gt;&gt;&gt; tf.config.list_physical_devices('GPU')

2020-05-10 14:58:16.243814: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1
2020-05-10 14:58:16.262675: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-05-10 14:58:16.263119: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1555] Found device 0 with properties:
pciBusID: 0000:01:00.0 name: GeForce GTX 1060 6GB computeCapability: 6.1
coreClock: 1.7715GHz coreCount: 10 deviceMemorySize: 5.93GiB deviceMemoryBandwidth: 178.99GiB/s
2020-05-10 14:58:16.263143: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1
2020-05-10 14:58:16.263188: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10
2020-05-10 14:58:16.264289: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10
2020-05-10 14:58:16.264495: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10
2020-05-10 14:58:16.265644: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10
2020-05-10 14:58:16.266329: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10
2020-05-10 14:58:16.266357: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-05-10 14:58:16.266478: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-05-10 14:58:16.266823: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:981] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2020-05-10 14:58:16.267107: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1697] Adding visible gpu devices: 0
[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]
</code></pre>

As suggested by @AmitaiIrron:

This section indicates that a gpu was found

<pre><code>2020-05-10 14:58:16.263119: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1555] Found device 0 with properties:

pciBusID: 0000:01:00.0 name: GeForce GTX 1060 6GB computeCapability: 6.1
coreClock: 1.7715GHz coreCount: 10 deviceMemorySize: 5.93GiB deviceMemoryBandwidth: 178.99GiB/s
</code></pre>

And here that it got added as an available physical device

<pre><code>2020-05-10 14:58:16.267107: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1697] Adding visible gpu devices: 0

[PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]
</code></pre>

blocks|key|4408916|text|在tensorflow+website上被列为检查TF是否使用图形处理器的“官方”方式的TF2.4%2B|type|unstyled|depth|inlineStyleRanges|entityRanges|offset|length|data|4408917|>>>+import+tensorflow+as+tf
>>>+print("Num+GPUs+Available:+",+len(tf.config.experimental.list_physical_devices('GPU')))
Num+GPUs+Available:++2|code-block|syntax|javascript|4408918|entityMap|0|LINK|mutability|MUTABLE|url|https://www.tensorflow.org/guide/gpu^0|1|I|0|0|0^^$0|@$1|2|3|4|5|6|7|Q|8|@]|9|@$A|R|B|S|1|T]]|C|$]]|$1|D|3|E|5|F|7|U|8|@]|9|@]|C|$G|H]]|$1|I|3|-4|5|6|7|V|8|@]|9|@]|C|$]]]|J|$K|$5|L|M|N|C|$O|P]]]]

<h3>For TF2.4+ listed as the &quot;official&quot; way on <a href="https://www.tensorflow.org/guide/gpu" rel="nofollow noreferrer">tensorflow website</a> to check if TF is using GPU or Not</h3>
<pre class="lang-py prettyprint-override"><code>&gt;&gt;&gt; import tensorflow as tf
&gt;&gt;&gt; print(&quot;Num GPUs Available: &quot;, len(tf.config.experimental.list_physical_devices('GPU')))
Num GPUs Available: 2
</code></pre>

blocks|key|4408973|text|在TF的新版本(>2.1)中，检查TF是否使用GPU的推荐方法是：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4408974|tf.config.list_physical_devices('GPU')|code-block|syntax|javascript|4408975|entityMap^0|0|0^^$0|@$1|2|3|4|5|6|7|I|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|J|8|@]|9|@]|A|$E|F]]|$1|G|3|-4|5|6|7|K|8|@]|9|@]|A|$]]]|H|$]]

In the new versions of TF(&gt;2.1) the recommended way for checking whether TF is using GPU is:
<pre><code>tf.config.list_physical_devices('GPU')
</code></pre>

blocks|key|4409198|text|我找到了最简单、最全面的方法。只需设置tf.debugging.set_log_device_placement(True)，您应该会看到操作是否真的在GPU上运行，例如Executing+op+_EagerConst+in+device+/job:localhost/replica:0/task:0/device:GPU:0|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|4409199|文档中有更多内容：https://www.tensorflow.org/guide/gpu#logging_device_placement|4409200|entityMap|0|LINK|mutability|MUTABLE|url|https://www.tensorflow.org/guide/gpu#logging_device_placement^0|J|17|2D|27|0|9|1P|0|0^^$0|@$1|2|3|4|5|6|7|P|8|@$9|Q|A|R|B|C]|$9|S|A|T|B|C]]|D|@]|E|$]]|$1|F|3|G|5|6|7|U|8|@]|D|@$9|V|A|W|1|X]]|E|$]]|$1|H|3|-4|5|6|7|Y|8|@]|D|@]|E|$]]]|I|$J|$5|K|L|M|E|$N|O]]]]

I found the most simple and comprehensive approach. Just set <code>tf.debugging.set_log_device_placement(True)</code> and you should see if ops are actually run on GPU e.g. <code>Executing op _EagerConst in device /job:localhost/replica:0/task:0/device:GPU:0</code>
More in the docs: <a href="https://www.tensorflow.org/guide/gpu#logging_device_placement" rel="nofollow noreferrer">https://www.tensorflow.org/guide/gpu#logging_device_placement</a>

blocks|key|4409257|text|也许你可以试试这个：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|4409258|python+-c“将tensorflow导入为tf；打印(tf.reduce_sum(tf.random.normal(1000,1000)”|4409259|查看系统是否返回张量|4409260|根据the+site的说法|offset|length|4409261|entityMap|0|LINK|mutability|MUTABLE|url|https://www.tensorflow.org/install/pip#virtual-environment-install^0|0|0|0|2|8|0|0^^$0|@$1|2|3|4|5|6|7|R|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|S|8|@]|9|@]|A|$]]|$1|D|3|E|5|6|7|T|8|@]|9|@]|A|$]]|$1|F|3|G|5|6|7|U|8|@]|9|@$H|V|I|W|1|X]]|A|$]]|$1|J|3|-4|5|6|7|Y|8|@]|9|@]|A|$]]]|K|$L|$5|M|N|O|A|$P|Q]]]]

maybe try this:
python -c &quot;import tensorflow as tf;print(tf.reduce_sum(tf.random.normal([1000, 1000])))&quot;
to see if the system returns the tensor
according to <a href="https://www.tensorflow.org/install/pip#virtual-environment-install" rel="nofollow noreferrer">the site</a>

I have installed tensorflow in my ubuntu 16.04 using the second answer <a href="https://devtalk.nvidia.com/default/topic/936429/-solved-tensorflow-with-gpu-in-anaconda-env-ubuntu-16-04-cuda-7-5-cudnn-/" rel="noreferrer">here</a> with ubuntu's builtin apt cuda installation.

Now my question is how can I test if tensorflow is really using gpu? I have a gtx 960m gpu. When I <code>import tensorflow</code> this is the output

<pre><code>I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcublas.so locally
I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcudnn.so locally
I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcufft.so locally
I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcuda.so.1 locally
I tensorflow/stream_executor/dso_loader.cc:105] successfully opened CUDA library libcurand.so locally
</code></pre>

Is this output enough to check if tensorflow is using gpu ?

How to tell if tensorflow is using gpu acceleration from inside python shell?

翻译质量差，导致语言生硬或混乱。

没有提供实际的解决方法或示例。

解答不清晰，无法理解或解决问题。

页面排版不美观，阅读体验差。

文章

问答

视频

学习中心

腾讯云实验室

直播

竞赛

腾讯云代码分析专区

腾讯iOA零信任安全管理系统专区

腾讯云架构师技术同盟交流圈

腾讯云数据库专区

腾讯云顾问专区

腾讯云原生专区

腾讯混元专区

腾讯云TCE专区

腾讯云Lighthouse专区

腾讯云HAI专区

腾讯云Edgeone专区

腾讯云存储专区

腾讯云智能专区

腾讯轻联专区 

腾讯云开发专区

TAPD专区

腾讯轻量云游戏服专区

腾讯云最具价值专家

腾讯云架构师技术同盟

腾讯云创作之星

腾讯云开发者先锋

腾讯云代码助手

云原生构建

TAPD 敏捷项目管理

Cloud Studio

SDK中心

API中心

命令行工具

涵盖代码开发、场景应用、自动测试全流程，助你从零构建专属AI助手

一站式MCP教程库，解锁AI应用新玩法

我已经在我的ubuntu16.04中使用第二个答案在ubuntu的内置apt cuda安装中安装了tensorflow。现在我的问题是，我如何测试tensorflow是否真的在使用gpu？我有一个gtx 960m的gpu。当我使用import tensorflow命令时，这是输出I tensorflow/stream_...

问如何判断tensorflow是否在python shell内部使用gpu加速？
EN

Stack Overflow用户

社区

活动

圈层

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问如何判断tensorflow是否在python shell内部使用gpu加速？EN