blocks|key|407546|text|如果您已经安装了CUDA+SDK，可以运行"deviceQuery“查看CUDA的版本|type|unstyled|depth|inlineStyleRanges|entityRanges|data|407547|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|$]]

If you have installed CUDA SDK, you can run "deviceQuery" to see the version of CUDA

blocks|key|407586|text|正如Jared在评论中提到的，从命令行：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|407587|nvcc+--version|code-block|syntax|javascript|407588|(或/usr/local/cuda/bin/nvcc+--version)提供CUDA编译器版本(与工具包版本匹配)。|offset|length|style|CODE|407589|从应用程序代码中，您可以使用以下命令查询运行时API版本|407590|cudaRuntimeGetVersion()|407591|或驱动程序API版本|407592|cudaDriverGetVersion()|407593|正如丹尼尔指出的那样，deviceQuery是一个软件开发工具包示例应用程序，可以查询上面的内容以及设备功能。|407594|正如其他人所指出的，你也可以使用(例如，在Mac或Linux上)来检查version.txt的内容。|407595|cat+/usr/local/cuda/version.txt|407596|但是，如果安装了除/usr/local/cuda符号链接版本之外的其他版本的CUDA工具包，如果PATH中的其他版本早于上述版本，则可能会报告不准确的版本，因此请谨慎使用。|407597|entityMap^0|0|0|2|Y|0|0|0|0|0|0|Z|B|0|0|9|F|1C|4|0^^$0|@$1|2|3|4|5|6|7|14|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|15|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|16|8|@$I|17|J|18|K|L]]|9|@]|A|$]]|$1|M|3|N|5|6|7|19|8|@]|9|@]|A|$]]|$1|O|3|P|5|D|7|1A|8|@]|9|@]|A|$E|F]]|$1|Q|3|R|5|6|7|1B|8|@]|9|@]|A|$]]|$1|S|3|T|5|D|7|1C|8|@]|9|@]|A|$E|F]]|$1|U|3|V|5|6|7|1D|8|@]|9|@]|A|$]]|$1|W|3|X|5|6|7|1E|8|@$I|1F|J|1G|K|L]]|9|@]|A|$]]|$1|Y|3|Z|5|D|7|1H|8|@]|9|@]|A|$E|F]]|$1|10|3|11|5|6|7|1I|8|@$I|1J|J|1K|K|L]|$I|1L|J|1M|K|L]]|9|@]|A|$]]|$1|12|3|-4|5|6|7|1N|8|@]|9|@]|A|$]]]|13|$]]

As Jared mentions in a comment, from the command line: 

<pre><code>nvcc --version
</code></pre>

(or <code>/usr/local/cuda/bin/nvcc --version</code>) gives the CUDA compiler version (which matches the toolkit version). 

From application code, you can query the runtime API version with

<pre><code>cudaRuntimeGetVersion()
</code></pre>

or the driver API version with

<pre><code>cudaDriverGetVersion()
</code></pre>

As Daniel points out, deviceQuery is an SDK sample app that queries the above, along with device capabilities.

As others note, you can also check the contents of the <code>version.txt</code> using (e.g., on Mac or Linux) 

<pre><code>cat /usr/local/cuda/version.txt
</code></pre>

However, if there is another version of the CUDA toolkit installed other than the one symlinked from <code>/usr/local/cuda</code>, this may report an inaccurate version if another version is earlier in your <code>PATH</code> than the above, so use with caution.

blocks|key|407631|text|安装CUDA后，可以通过以下方式检查版本:+nvcc+-V|type|unstyled|depth|inlineStyleRanges|entityRanges|data|407632|我已经安装了5.0和5.5，所以它提供了|407633|Cuda编译工具，版本5.5，V5.5,0|407634|此命令适用于Windows和Ubuntu。|407635|entityMap^0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|J|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|K|8|@]|9|@]|A|$]]|$1|D|3|E|5|6|7|L|8|@]|9|@]|A|$]]|$1|F|3|G|5|6|7|M|8|@]|9|@]|A|$]]|$1|H|3|-4|5|6|7|N|8|@]|9|@]|A|$]]]|I|$]]

After installing CUDA one can check the versions by: nvcc -V

I have installed both 5.0 and 5.5 so it gives 

Cuda Compilation Tools,release 5.5,V5.5,0

This command works for both Windows and Ubuntu.

blocks|key|622849|text|除上述路径外，CUDA安装路径(如果在安装过程中未更改)通常包含版本号|type|unstyled|depth|inlineStyleRanges|entityRanges|data|622850|执行which+nvcc应该会给出路径，然后就会给出版本|offset|length|style|CODE|622851|PS:这是一种快速而肮脏的方式，上面的答案更加优雅，并将在相当努力的情况下产生正确的版本|622852|entityMap^0|0|2|A|0|0^^$0|@$1|2|3|4|5|6|7|L|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|M|8|@$D|N|E|O|F|G]]|9|@]|A|$]]|$1|H|3|I|5|6|7|P|8|@]|9|@]|A|$]]|$1|J|3|-4|5|6|7|Q|8|@]|9|@]|A|$]]]|K|$]]

Apart from the ones mentioned above, your CUDA installations path (if not changed during setup) typically contains the version number

doing a <code>which nvcc</code> should give the path and that will give you the version

PS: This is a quick and dirty way, the above answers are more elegant and will result in the right version with considerable effort

blocks|key|408304|text|你可能会发现CUDA-Z很有用，下面是他们网站上的一段话：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|408305|“此程序是对其他Z-实用程序的模仿，如CPU-Z和GPU-Z。CUDA-Z显示了一些有关启用CUDA的GPU和GPGPU的基本信息。它适用于nVIDIA+Geforce、Quadro和Tesla卡、离子芯片组。”|408306|http://cuda-z.sourceforge.net/|offset|length|408307|在支持选项卡上有源代码的网址：http://sourceforge.net/p/cuda-z/code/，下载的实际上不是一个安装程序，而是可执行文件本身(没有安装，所以这是“快速”)。|408308|此实用程序提供了大量信息，如果您需要了解它是如何派生的，可以查看源代码。您可以搜索与此类似的其他实用程序。|408309|entityMap|0|LINK|mutability|MUTABLE|url|1|http://sourceforge.net/p/cuda-z/code/^0|0|0|0|U|0|0|F|11|1|0|0^^$0|@$1|2|3|4|5|6|7|U|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|V|8|@]|9|@]|A|$]]|$1|D|3|E|5|6|7|W|8|@]|9|@$F|X|G|Y|1|Z]]|A|$]]|$1|H|3|I|5|6|7|10|8|@]|9|@$F|11|G|12|1|13]]|A|$]]|$1|J|3|K|5|6|7|14|8|@]|9|@]|A|$]]|$1|L|3|-4|5|6|7|15|8|@]|9|@]|A|$]]]|M|$N|$5|O|P|Q|A|$R|E]]|S|$5|O|P|Q|A|$R|T]]]]

You might find CUDA-Z useful, here is a quote from their Site:

"This program was born as a parody of another Z-utilities such as CPU-Z and GPU-Z. CUDA-Z shows some basic information about CUDA-enabled GPUs and GPGPUs. It works with nVIDIA Geforce, Quadro and Tesla cards, ION chipsets."

<a href="http://cuda-z.sourceforge.net/" rel="noreferrer">http://cuda-z.sourceforge.net/</a>

On the Support Tab there is the URL for the Source Code: <a href="http://sourceforge.net/p/cuda-z/code/" rel="noreferrer">http://sourceforge.net/p/cuda-z/code/</a> and the download is not actually an Installer but the Executable itself (no installation, so this is "quick").

This Utility provides lots of information and if you need to know how it was derived there is the Source to look at. There are other Utilities similar to this that you might search for.

blocks|key|408379|text|在Ubuntu+Cuda+V8上：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|408380|$+cat+/usr/local/cuda/version.txt|code-block|syntax|javascript|408381|您还可以了解安装了哪些CUDA版本：|408382|$+ls+-l+/usr/local+%7C+grep+cuda|408383|它会给你类似这样的结果：|408384|lrwxrwxrwx++1+root+root++++9+Mar++5++2020+cuda+->+cuda-10.2
drwxr-xr-x+16+root+root+4096+Mar++5++2020+cuda-10.2
drwxr-xr-x+16+root+root+4096+Mar++5++2020+cuda-8.0.61|408385|给定一条合理的路径，cuda指向的版本应该是活动的版本(在本例中为10.2)。|offset|length|style|CODE|408386|注意:只有当您愿意假设CUDA安装在/usr/local/cuda下时，这才有效(这对于默认位置的独立安装程序是正确的，但对于带有CUDA作为软件包集成的发行版则不是这样)。参考：comment+from+@einpoklum。|408387|entityMap|0|LINK|mutability|MUTABLE|url|https://stackoverflow.com/questions/9727688/how-to-get-the-cuda-version/59868633?noredirect=1#comment117729294_59868633^0|0|0|0|0|0|0|A|4|0|2I|C|0|0^^$0|@$1|2|3|4|5|6|7|14|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|15|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|16|8|@]|9|@]|A|$]]|$1|I|3|J|5|D|7|17|8|@]|9|@]|A|$E|F]]|$1|K|3|L|5|6|7|18|8|@]|9|@]|A|$]]|$1|M|3|N|5|D|7|19|8|@]|9|@]|A|$E|F]]|$1|O|3|P|5|6|7|1A|8|@$Q|1B|R|1C|S|T]]|9|@]|A|$]]|$1|U|3|V|5|6|7|1D|8|@]|9|@$Q|1E|R|1F|1|1G]]|A|$]]|$1|W|3|-4|5|6|7|1H|8|@]|9|@]|A|$]]]|X|$Y|$5|Z|10|11|A|$12|13]]]]

On Ubuntu Cuda V8:
<pre><code>$ cat /usr/local/cuda/version.txt
 
</code></pre>
You can also get some insights into which CUDA versions are installed with:
<pre><code>$ ls -l /usr/local | grep cuda
</code></pre>
which will give you something like this:
<pre><code>lrwxrwxrwx 1 root root 9 Mar 5 2020 cuda -&gt; cuda-10.2
drwxr-xr-x 16 root root 4096 Mar 5 2020 cuda-10.2
drwxr-xr-x 16 root root 4096 Mar 5 2020 cuda-8.0.61
</code></pre>
Given a sane PATH, the version <code>cuda</code> points to should be the active one (10.2 in this case).
NOTE: This only works if you are willing to assume CUDA is installed under /usr/local/cuda (which is true for the independent installer with the default location, but not true e.g. for distributions with CUDA integrated as a package). Ref: <a href="https://stackoverflow.com/questions/9727688/how-to-get-the-cuda-version/59868633?noredirect=1#comment117729294_59868633">comment from</a> @einpoklum.

blocks|key|408438|text|在Ubuntu上：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|408439|试一试|408440|$+cat+/usr/local/cuda/version.txt或$+cat+/usr/local/cuda-8.0/version.txt|offset|length|style|CODE|408441|有时文件夹被命名为"Cuda-version“。|408442|如果上述方法都不起作用，请尝试访问$+/usr/local/并找到正确的Cuda文件夹名称。|408443|输出应类似于：CUDA+Version+8.0.61|408444|entityMap^0|0|0|0|X|Y|11|0|0|H|D|0|7|J|0^^$0|@$1|2|3|4|5|6|7|R|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|S|8|@]|9|@]|A|$]]|$1|D|3|E|5|6|7|T|8|@$F|U|G|V|H|I]|$F|W|G|X|H|I]]|9|@]|A|$]]|$1|J|3|K|5|6|7|Y|8|@]|9|@]|A|$]]|$1|L|3|M|5|6|7|Z|8|@$F|10|G|11|H|I]]|9|@]|A|$]]|$1|N|3|O|5|6|7|12|8|@$F|13|G|14|H|I]]|9|@]|A|$]]|$1|P|3|-4|5|6|7|15|8|@]|9|@]|A|$]]]|Q|$]]

On Ubuntu :

Try

<code>$ cat /usr/local/cuda/version.txt</code>
or
<code>$ cat /usr/local/cuda-8.0/version.txt</code>

Sometimes the folder is named "Cuda-version".

If none of above works, try going to
<code>$ /usr/local/</code>
And find the correct name of your Cuda folder.

Output should be similar to:
<code>CUDA Version 8.0.61</code>

blocks|key|623684|text|我得到/usr/local+-没有这样的文件或目录。尽管nvcc+-V给出了|type|unstyled|depth|inlineStyleRanges|entityRanges|data|623685|nvcc:+NVIDIA+(R)+Cuda+compiler+driver
Copyright+(c)+2005-2016+NVIDIA+Corporation
Built+on+Sun_Sep__4_22:14:01_CDT_2016
Cuda+compilation+tools,+release+8.0,+V8.0.44|code-block|syntax|javascript|623686|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|$]]

i get /usr/local - no such file or directory. Though nvcc -V gives 

<pre><code>nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2016 NVIDIA Corporation
Built on Sun_Sep__4_22:14:01_CDT_2016
Cuda compilation tools, release 8.0, V8.0.44
</code></pre>

blocks|key|408559|text|如果nvcc+--version不适用于您，则使用cat+/usr/local/cuda/version.txt|type|unstyled|depth|inlineStyleRanges|entityRanges|data|408560|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|$]]

if nvcc --version is not working for you then use cat /usr/local/cuda/version.txt

blocks|key|408671|text|如果你运行|type|unstyled|depth|inlineStyleRanges|entityRanges|data|408672|nvidia-smi|code-block|syntax|javascript|408673|您应该可以在命令输出的右上角找到CUDA版本。至少我发现CUDA版本10.0的输出，例如，|408674|📷|atomic|offset|length|408675|​|408676|entityMap|0|IMAGE|mutability|IMMUTABLE|imageUrl|https://ask.qcloudimg.com/http-save/yehe-900000/0f302b47bedff7c6429ef811f79cf231.png|imageAlt^0|0|0|0|0|1|0|0|0^^$0|@$1|2|3|4|5|6|7|Y|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|Z|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|10|8|@]|9|@]|A|$]]|$1|I|3|J|5|K|7|11|8|@]|9|@$L|12|M|13|1|14]]|A|$]]|$1|N|3|O|5|6|7|15|8|@]|9|@]|A|$]]|$1|P|3|-4|5|6|7|16|8|@]|9|@]|A|$]]]|Q|$R|$5|S|T|U|A|$V|W|X|-4]]]]

If you run

<pre><code>nvidia-smi
</code></pre>

You should find the CUDA Version on the top right corner of the comand's output. At least I found that output for CUDA version 10.0 e.g.,
<a href="https://i.stack.imgur.com/0iIOs.png" rel="noreferrer"><img src="https://i.stack.imgur.com/0iIOs.png" alt="enter image description here"></a>

blocks|key|623904|text|如果您在linux上运行：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|623905|dpkg+-l+%7C+grep+cuda|code-block|syntax|javascript|623906|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 running on linux:

<pre><code>dpkg -l | grep cuda
</code></pre>

blocks|key|624024|text|使用以下命令检查Conda的CUDA安装：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|624025|conda+list+cudatoolkit|code-block|syntax|javascript|624026|并使用以下命令检查conda安装的CUDNN版本：|624027|conda+list+cudnn|624028|如果您希望通过CONDA安装/更新CUDA和CUDNN，请使用以下命令：|624029|conda+install+-c+anaconda+cudatoolkit
conda+install+-c+anaconda+cudnn|624030|或者，您也可以使用以下命令检查CUDA安装：|624031|nvidia-smi|624032|或|624033|nvcc+--version|624034|如果您通过Anaconda包使用tensorflow-gpu+(您可以通过在控制台中打开Python进行验证，并检查启动时默认python是否显示Anaconda，Inc.，或者您可以运行哪个python并检查位置)，那么手动安装CUDA和CUDNN很可能不起作用。您将不得不通过conda进行更新。|624035|如果您想要手动安装CUDA、CUDNN或tensorflow-gpu，您可以查看此处的说明https://www.tensorflow.org/install/gpu|offset|length|624036|entityMap|0|LINK|mutability|MUTABLE|url|https://www.tensorflow.org/install/gpu^0|0|0|0|0|0|0|0|0|0|0|0|19|12|0|0^^$0|@$1|2|3|4|5|6|7|1A|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|1B|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|1C|8|@]|9|@]|A|$]]|$1|I|3|J|5|D|7|1D|8|@]|9|@]|A|$E|F]]|$1|K|3|L|5|6|7|1E|8|@]|9|@]|A|$]]|$1|M|3|N|5|D|7|1F|8|@]|9|@]|A|$E|F]]|$1|O|3|P|5|6|7|1G|8|@]|9|@]|A|$]]|$1|Q|3|R|5|D|7|1H|8|@]|9|@]|A|$E|F]]|$1|S|3|T|5|6|7|1I|8|@]|9|@]|A|$]]|$1|U|3|V|5|D|7|1J|8|@]|9|@]|A|$E|F]]|$1|W|3|X|5|6|7|1K|8|@]|9|@]|A|$]]|$1|Y|3|Z|5|6|7|1L|8|@]|9|@$10|1M|11|1N|1|1O]]|A|$]]|$1|12|3|-4|5|6|7|1P|8|@]|9|@]|A|$]]]|13|$14|$5|15|16|17|A|$18|19]]]]

Use the following command to check CUDA installation by Conda:

<pre><code>conda list cudatoolkit
</code></pre>

And the following command to check CUDNN version installed by conda:

<pre><code>conda list cudnn
</code></pre>

If you want to install/update CUDA and CUDNN through CONDA, please use the following commands:

<pre><code>conda install -c anaconda cudatoolkit
conda install -c anaconda cudnn
</code></pre>

Alternatively you can use following commands to check CUDA installation:

<pre><code>nvidia-smi
</code></pre>

OR

<pre><code>nvcc --version
</code></pre>

If you are using tensorflow-gpu through Anaconda package (You can verify this by simply opening Python in console and check if the default python shows Anaconda, Inc. when it starts, or you can run which python and check the location), then manually installing CUDA and CUDNN will most probably not work. You will have to update through conda instead. 

If you want to install CUDA, CUDNN, or tensorflow-gpu manually, you can check out the instructions here <a href="https://www.tensorflow.org/install/gpu" rel="noreferrer">https://www.tensorflow.org/install/gpu</a>

blocks|key|624071|text|以编程方式使用CUDA+Runtime+API+C%2B%2B+wrappers|type|unstyled|depth|inlineStyleRanges|entityRanges|offset|length|data|624072|auto+v1+=+cuda::version::maximum_supported_by_driver();
auto+v2+=+cuda::version::runtime();|code-block|syntax|javascript|624073|这为您提供了一个cuda::version_t结构，您可以对其进行比较，还可以打印/流传输，例如：|style|CODE|624074|if+(v2+<+cuda::version_t{+8,+0+}+)+{
++++std::cerr+<<+"CUDA+version+"+<<+v2+<<+"+is+insufficient."+std::endl;
}|624075|entityMap|0|LINK|mutability|MUTABLE|url|https://docs.nvidia.com/cuda/cuda-runtime-api/index.html^0|7|T|0|0|0|8|F|0|0^^$0|@$1|2|3|4|5|6|7|W|8|@]|9|@$A|X|B|Y|1|Z]]|C|$]]|$1|D|3|E|5|F|7|10|8|@]|9|@]|C|$G|H]]|$1|I|3|J|5|6|7|11|8|@$A|12|B|13|K|L]]|9|@]|C|$]]|$1|M|3|N|5|F|7|14|8|@]|9|@]|C|$G|H]]|$1|O|3|-4|5|6|7|15|8|@]|9|@]|C|$]]]|P|$Q|$5|R|S|T|C|$U|V]]]]

Programmatically with the <a href="https://docs.nvidia.com/cuda/cuda-runtime-api/index.html" rel="nofollow noreferrer">CUDA Runtime API C++ wrappers</a>:
<pre><code>auto v1 = cuda::version::maximum_supported_by_driver();
auto v2 = cuda::version::runtime();
</code></pre>
This gives you a <code>cuda::version_t</code> structure, which you can compare and also print/stream e.g.:
<pre><code>if (v2 &lt; cuda::version_t{ 8, 0 } ) {
 std::cerr &lt;&lt; &quot;CUDA version &quot; &lt;&lt; v2 &lt;&lt; &quot; is insufficient.&quot; std::endl;
}
</code></pre>

blocks|key|408972|text|您可以使用以下命令检查CUDA的版本|type|unstyled|depth|inlineStyleRanges|entityRanges|data|408973|nvcc+-V|code-block|syntax|javascript|408974|或者您可以使用|408975|nvcc+--version|408976|或者，您可以检查CUDA使用的位置|408977|whereis+cuda+|408978|然后做|408979|cat+location/of/cuda/you/got/from/above/command|408980|entityMap^0|0|0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|U|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|V|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|W|8|@]|9|@]|A|$]]|$1|I|3|J|5|D|7|X|8|@]|9|@]|A|$E|F]]|$1|K|3|L|5|6|7|Y|8|@]|9|@]|A|$]]|$1|M|3|N|5|D|7|Z|8|@]|9|@]|A|$E|F]]|$1|O|3|P|5|6|7|10|8|@]|9|@]|A|$]]|$1|Q|3|R|5|D|7|11|8|@]|9|@]|A|$E|F]]|$1|S|3|-4|5|6|7|12|8|@]|9|@]|A|$]]]|T|$]]

You can check the version of CUDA using 

<pre><code>nvcc -V
</code></pre>

or you can use

<pre><code>nvcc --version
</code></pre>

or You can check the location of where the CUDA is using 

<pre><code>whereis cuda 
</code></pre>

and then do

<pre><code>cat location/of/cuda/you/got/from/above/command
</code></pre>

blocks|key|409032|text|我们有三种方法来检查版本:在我的例子中，下面是输出：-方式1:-|type|unstyled|depth|inlineStyleRanges|entityRanges|data|409033|cat+/usr/local/cuda/version.txt|code-block|syntax|javascript|409034|输出：-|409035|CUDA+Version+10.1.243|409036|方式2：-|409037|nvcc+--version|409038|409039|nvcc:+NVIDIA+(R)+Cuda+compiler+driver
Copyright+(c)+2005-2017+NVIDIA+Corporation
Built+on+Fri_Nov__3_21:07:56_CDT_2017
Cuda+compilation+tools,+release+9.1,+V9.1.85|409040|方式3：-|409041|/usr/local/cuda/bin/nvcc+--version|409042|409043|nvcc:+NVIDIA+(R)+Cuda+compiler+driver
Copyright+(c)+2005-2019+NVIDIA+Corporation
Built+on+Sun_Jul_28_19:07:16_PDT_2019
Cuda+compilation+tools,+release+10.1,+V10.1.243|409044|方式4：-|409045|nvidia-smi
NVIDIA-SMI+450.36.06++++Driver+Version:+450.36.06++++CUDA+Version:+11.0|409046|输出不同。不知道为什么会这样。|409047|entityMap^0|0|0|0|0|0|0|0|0|0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|16|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|17|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|18|8|@]|9|@]|A|$]]|$1|I|3|J|5|D|7|19|8|@]|9|@]|A|$E|F]]|$1|K|3|L|5|6|7|1A|8|@]|9|@]|A|$]]|$1|M|3|N|5|D|7|1B|8|@]|9|@]|A|$E|F]]|$1|O|3|H|5|6|7|1C|8|@]|9|@]|A|$]]|$1|P|3|Q|5|D|7|1D|8|@]|9|@]|A|$E|F]]|$1|R|3|S|5|6|7|1E|8|@]|9|@]|A|$]]|$1|T|3|U|5|D|7|1F|8|@]|9|@]|A|$E|F]]|$1|V|3|H|5|6|7|1G|8|@]|9|@]|A|$]]|$1|W|3|X|5|D|7|1H|8|@]|9|@]|A|$E|F]]|$1|Y|3|Z|5|6|7|1I|8|@]|9|@]|A|$]]|$1|10|3|11|5|D|7|1J|8|@]|9|@]|A|$E|F]]|$1|12|3|13|5|6|7|1K|8|@]|9|@]|A|$]]|$1|14|3|-4|5|6|7|1L|8|@]|9|@]|A|$]]]|15|$]]

We have three ways to check Version:
In my case below is the output:-
Way 1:-

<pre><code>cat /usr/local/cuda/version.txt
</code></pre>

Output:-

<pre><code>CUDA Version 10.1.243
</code></pre>

Way2:-

<pre><code>nvcc --version
</code></pre>

Output:-

<pre><code>nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2017 NVIDIA Corporation
Built on Fri_Nov__3_21:07:56_CDT_2017
Cuda compilation tools, release 9.1, V9.1.85
</code></pre>

Way3:-

<pre><code>/usr/local/cuda/bin/nvcc --version
</code></pre>

Output:-

<pre><code>nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2019 NVIDIA Corporation
Built on Sun_Jul_28_19:07:16_PDT_2019
Cuda compilation tools, release 10.1, V10.1.243
</code></pre>

Way4:-

<pre><code>nvidia-smi
NVIDIA-SMI 450.36.06 Driver Version: 450.36.06 CUDA Version: 11.0
</code></pre>

Outputs are not same. Don't know why it's happening.

blocks|key|624213|text|如果nvcc和nvidia-smi之间存在版本不匹配，则使用不同版本的cuda作为驱动程序和运行时环境then。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|624214|为了确保使用相同版本的CUDA驱动程序，您需要做的是在系统路径上获取CUDA。|624215|首先运行whereis+cuda并找到cuda驱动程序的位置。|624216|然后转到.bashrc并修改path变量，并使用变量'LD_LIBRARY_PATH‘设置搜索的目录优先顺序。|624217|例如|624218|$+whereis+cuda
cuda:+/usr/lib/cuda+/usr/include/cuda.h+/usr/local/cuda|code-block|syntax|javascript|624219|CUDA安装在/usr/local/cuda，现在我们需要转到.bashrc并添加路径变量，如下所示：|624220|vim++~/.bashrc
export+PATH="/usr/local/cuda/bin:${PATH}"|624221|在此行之后，将目录搜索路径设置为：|624222|export+LD_LIBRARY_PATH="/usr/local/cuda/lib64:${LD_LIBRARY_PATH}"|624223|然后保存.bashrc文件。并将其刷新为：|624224|$+source+~/.bashrc|624225|这将确保您的nvcc+-V和nvidia-smi使用相同版本的驱动程序。|624226|entityMap^0|0|0|0|0|0|0|0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|14|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|15|8|@]|9|@]|A|$]]|$1|D|3|E|5|6|7|16|8|@]|9|@]|A|$]]|$1|F|3|G|5|6|7|17|8|@]|9|@]|A|$]]|$1|H|3|I|5|6|7|18|8|@]|9|@]|A|$]]|$1|J|3|K|5|L|7|19|8|@]|9|@]|A|$M|N]]|$1|O|3|P|5|6|7|1A|8|@]|9|@]|A|$]]|$1|Q|3|R|5|L|7|1B|8|@]|9|@]|A|$M|N]]|$1|S|3|T|5|6|7|1C|8|@]|9|@]|A|$]]|$1|U|3|V|5|L|7|1D|8|@]|9|@]|A|$M|N]]|$1|W|3|X|5|6|7|1E|8|@]|9|@]|A|$]]|$1|Y|3|Z|5|L|7|1F|8|@]|9|@]|A|$M|N]]|$1|10|3|11|5|6|7|1G|8|@]|9|@]|A|$]]|$1|12|3|-4|5|6|7|1H|8|@]|9|@]|A|$]]]|13|$]]

If there is a version mismatch between nvcc and nvidia-smi then different versions of cuda are used as driver and run time environemtn.
To ensure same version of CUDA drivers are used what you need to do is to get CUDA on system path.
First run whereis cuda and find the location of cuda driver.
Then go to .bashrc and modify the path variable and set the directory precedence order of search using variable 'LD_LIBRARY_PATH'.
for instance
<pre><code>$ whereis cuda
cuda: /usr/lib/cuda /usr/include/cuda.h /usr/local/cuda
</code></pre>
CUDA is installed at /usr/local/cuda, now we need to to .bashrc and add the path variable as:
<pre><code>vim ~/.bashrc
export PATH=&quot;/usr/local/cuda/bin:${PATH}&quot;
</code></pre>
and after this line set the directory search path as:
<pre><code>export LD_LIBRARY_PATH=&quot;/usr/local/cuda/lib64:${LD_LIBRARY_PATH}&quot;
</code></pre>
Then save the .bashrc file. And refresh it as:
<pre><code>$ source ~/.bashrc
</code></pre>
This will ensure you have nvcc -V and nvidia-smi to use the same version of drivers.

blocks|key|409113|text|您还可以使用：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|409114|nvidia-smi+%7C+grep+"CUDA+Version:"+|code-block|syntax|javascript|409115|来检索显式行。|409116|entityMap^0|0|0|0^^$0|@$1|2|3|4|5|6|7|K|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|L|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|M|8|@]|9|@]|A|$]]|$1|I|3|-4|5|6|7|N|8|@]|9|@]|A|$]]]|J|$]]

You could also use:
<pre><code>nvidia-smi | grep &quot;CUDA Version:&quot; 
</code></pre>
To retrieve the explicit line.

blocks|key|624283|text|如果你已经安装了PyTorch，你可以简单地在你的集成开发环境中运行以下代码：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|624284|import+torch

print(torch.version.cuda)|code-block|syntax|javascript|624285|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 have PyTorch installed, you can simply run the following code in your IDE:
<pre><code>import torch

print(torch.version.cuda)
</code></pre>

blocks|key|409275|text|在Windows+10上，我在'C:\Program+Files\NVIDIA+Corporation\NVSMI‘中找到了nvidia-smi.exe；在cd进入该文件夹(在我的例子中不在路径中)之后，显示了'.\nvidia-smi.exe’|type|unstyled|depth|inlineStyleRanges|entityRanges|data|409276|📷|atomic|offset|length|409277|​|409278|entityMap|0|IMAGE|mutability|IMMUTABLE|imageUrl|https://ask.qcloudimg.com/http-save/yehe-900000/f05b5504321eed9f5c658a97bc0237fa.png|imageAlt^0|0|0|1|0|0|0^^$0|@$1|2|3|4|5|6|7|R|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|S|8|@]|9|@$E|T|F|U|1|V]]|A|$]]|$1|G|3|H|5|6|7|W|8|@]|9|@]|A|$]]|$1|I|3|-4|5|6|7|X|8|@]|9|@]|A|$]]]|J|$K|$5|L|M|N|A|$O|P|Q|-4]]]]

On Windows 10, I found nvidia-smi.exe in 'C:\Program Files\NVIDIA Corporation\NVSMI'; after cd into that folder (was not in the PATH in my case) and '.\nvidia-smi.exe' it showed
<a href="https://i.stack.imgur.com/xIXjU.png" rel="noreferrer"><img src="https://i.stack.imgur.com/xIXjU.png" alt="enter image description here" /></a>

blocks|key|409410|text|其他受访者已经描述了哪些命令可用于检查CUDA版本。在这里，我将描述如何将这些命令的输出转换为"10.2“、"11.0”等形式的环境变量。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|409411|简单地说，您可以使用|409412|nvcc+--version|code-block|syntax|javascript|409413|了解CUDA版本。我想这应该是你的第一站。如果您安装了CUDA的多个版本，此命令应打印出路径中最高版本的副本。|409414|输出如下所示：|409415|nvcc:+NVIDIA+(R)+Cuda+compiler+driver
Copyright+(c)+2005-2020+NVIDIA+Corporation
Built+on+Thu_Jun_11_22:26:38_PDT_2020
Cuda+compilation+tools,+release+11.0,+V11.0.194
Build+cuda_11.0_bu.TC445_37.28540450_0|409416|我们可以将此输出传递给sed，以便只挑选出MAJOR.MINOR发行版本号。|409417|CUDA_VERSION=$(nvcc+--version+%7C+sed+-n+'s/%5E.*release+$[0-9\.]\%2B$.*$/\1/p')|409418|如果nvcc不在您的路径上，您应该能够通过指定nvcc的默认位置的完整路径来运行它。|409419|/usr/local/cuda/bin/nvcc+--version|409420|它的输出与上面相同，并且可以用相同的方式进行解析。|409421|或者，您可以从version.txt文件中找到CUDA版本。|409422|cat+/usr/local/cuda/version.txt|409423|它的输出|409424|CUDA+Version+10.1.243|409425|可以使用sed进行解析，以便只找出MAJOR.MINOR发行版本号。|409426|CUDA_VERSION=$(cat+/usr/local/cuda/version.txt+%7C+sed+'s/.*+$[0-9]\%2B\.[0-9]\%2B$.*/\1/')|409427|请注意，有时version.txt文件引用的是与nvcc+--version不同的CUDA安装。在这种情况下，nvcc版本应该是您实际使用的版本。|offset|length|style|CODE|409428|我们可以将这三种方法结合在一起，以便可靠地获得CUDA版本，如下所示：|409429|if+nvcc+--version+2&>+/dev/null;+then
++++#+Determine+CUDA+version+using+default+nvcc+binary
++++CUDA_VERSION=$(nvcc+--version+%7C+sed+-n+'s/%5E.*release+$[0-9\.]\%2B$.*$/\1/p');

elif+/usr/local/cuda/bin/nvcc+--version+2&>+/dev/null;+then
++++#+Determine+CUDA+version+using+/usr/local/cuda/bin/nvcc+binary
++++CUDA_VERSION=$(/usr/local/cuda/bin/nvcc+--version+%7C+sed+-n+'s/%5E.*release+$[0-9\.]\%2B$.*$/\1/p');

elif+[+-f+"/usr/local/cuda/version.txt"+];+then
++++#+Determine+CUDA+version+using+/usr/local/cuda/version.txt+file
++++CUDA_VERSION=$(cat+/usr/local/cuda/version.txt+%7C+sed+'s/.*+$[0-9]\%2B\.[0-9]\%2B$.*/\1/')

else
++++CUDA_VERSION=""

fi|409430|此环境变量对于下游安装非常有用，例如当pip安装为正确的CUDA版本编译的pytorch副本时。|409431|python+-m+pip+install+\
++++"torch==1.9.0%2Bcu${CUDA_VERSION/./}"+\
++++"torchvision==0.10.0%2Bcu${CUDA_VERSION/./}"+\
++++-f+https://download.pytorch.org/whl/torch_stable.html|409432|同样，您可以在未安装CUDA的情况下安装pytorch的CPU版本。|409433|if+[+"$CUDA_VERSION"+=+""+];+then
++++MOD="%2Bcpu";
++++echo+"Warning:+Installing+CPU-only+version+of+pytorch"
else
++++MOD="%2Bcu${CUDA_VERSION/./}";
++++echo+"Installing+pytorch+with+$MOD"
fi

python+-m+pip+install+\
++++"torch==1.9.0${MOD}"+\
++++"torchvision==0.10.0${MOD}"+\
++++-f+https://download.pytorch.org/whl/torch_stable.html|409434|但是要小心这一点，因为当你想要GPU支持的时候，你可能会意外地安装一个只有CPU的版本。例如，如果您在没有GPU的服务器登录节点上运行安装脚本，而您的作业将部署到具有GPU的节点上。在这种情况下，登录节点通常不会安装CUDA。|409435|entityMap^0|0|0|0|0|0|0|0|0|0|0|0|0|0|0|0|0|0|O|E|0|0|0|0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|1W|8|@]|9|@]|A|$]]|$1|B|3|C|5|6|7|1X|8|@]|9|@]|A|$]]|$1|D|3|E|5|F|7|1Y|8|@]|9|@]|A|$G|H]]|$1|I|3|J|5|6|7|1Z|8|@]|9|@]|A|$]]|$1|K|3|L|5|6|7|20|8|@]|9|@]|A|$]]|$1|M|3|N|5|F|7|21|8|@]|9|@]|A|$G|H]]|$1|O|3|P|5|6|7|22|8|@]|9|@]|A|$]]|$1|Q|3|R|5|F|7|23|8|@]|9|@]|A|$G|H]]|$1|S|3|T|5|6|7|24|8|@]|9|@]|A|$]]|$1|U|3|V|5|F|7|25|8|@]|9|@]|A|$G|H]]|$1|W|3|X|5|6|7|26|8|@]|9|@]|A|$]]|$1|Y|3|Z|5|6|7|27|8|@]|9|@]|A|$]]|$1|10|3|11|5|F|7|28|8|@]|9|@]|A|$G|H]]|$1|12|3|13|5|6|7|29|8|@]|9|@]|A|$]]|$1|14|3|15|5|F|7|2A|8|@]|9|@]|A|$G|H]]|$1|16|3|17|5|6|7|2B|8|@]|9|@]|A|$]]|$1|18|3|19|5|F|7|2C|8|@]|9|@]|A|$G|H]]|$1|1A|3|1B|5|6|7|2D|8|@$1C|2E|1D|2F|1E|1F]]|9|@]|A|$]]|$1|1G|3|1H|5|6|7|2G|8|@]|9|@]|A|$]]|$1|1I|3|1J|5|F|7|2H|8|@]|9|@]|A|$G|H]]|$1|1K|3|1L|5|6|7|2I|8|@]|9|@]|A|$]]|$1|1M|3|1N|5|F|7|2J|8|@]|9|@]|A|$G|H]]|$1|1O|3|1P|5|6|7|2K|8|@]|9|@]|A|$]]|$1|1Q|3|1R|5|F|7|2L|8|@]|9|@]|A|$G|H]]|$1|1S|3|1T|5|6|7|2M|8|@]|9|@]|A|$]]|$1|1U|3|-4|5|6|7|2N|8|@]|9|@]|A|$]]]|1V|$]]

Other respondents have already described which commands can be used to check the CUDA version. Here, I'll describe how to turn the output of those commands into an environment variable of the form &quot;10.2&quot;, &quot;11.0&quot;, etc.
To recap, you can use
<pre class="lang-sh prettyprint-override"><code>nvcc --version
</code></pre>
to find out the CUDA version.
I think this should be your first port of call.
If you have multiple versions of CUDA installed, this command should print out the version for the copy which is highest on your PATH.
The output looks like this:
<pre><code>nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2020 NVIDIA Corporation
Built on Thu_Jun_11_22:26:38_PDT_2020
Cuda compilation tools, release 11.0, V11.0.194
Build cuda_11.0_bu.TC445_37.28540450_0
</code></pre>
We can pass this output through sed to pick out just the MAJOR.MINOR release version number.
<pre class="lang-sh prettyprint-override"><code>CUDA_VERSION=$(nvcc --version | sed -n 's/^.*release $[0-9\.]\+$.*$/\1/p')
</code></pre>
If nvcc isn't on your path, you should be able to run it by specifying the full path to the default location of nvcc instead.
<pre class="lang-sh prettyprint-override"><code>/usr/local/cuda/bin/nvcc --version
</code></pre>
The output of which is the same as above, and it can be parsed in the same way.
Alternatively, you can find the CUDA version from the version.txt file.
<pre class="lang-sh prettyprint-override"><code>cat /usr/local/cuda/version.txt
</code></pre>
The output of which
<pre><code>CUDA Version 10.1.243
</code></pre>
can be parsed using sed to pick out just the MAJOR.MINOR release version number.
<pre class="lang-sh prettyprint-override"><code>CUDA_VERSION=$(cat /usr/local/cuda/version.txt | sed 's/.* $[0-9]\+\.[0-9]\+$.*/\1/')
</code></pre>
Note that sometimes the version.txt file refers to a different CUDA installation than the <code>nvcc --version</code>. In this scenario, the nvcc version should be the version you're actually using.
We can combine these three methods together in order to robustly get the CUDA version as follows:
<pre class="lang-sh prettyprint-override"><code>if nvcc --version 2&amp;&gt; /dev/null; then
 # Determine CUDA version using default nvcc binary
 CUDA_VERSION=$(nvcc --version | sed -n 's/^.*release $[0-9\.]\+$.*$/\1/p');

elif /usr/local/cuda/bin/nvcc --version 2&amp;&gt; /dev/null; then
 # Determine CUDA version using /usr/local/cuda/bin/nvcc binary
 CUDA_VERSION=$(/usr/local/cuda/bin/nvcc --version | sed -n 's/^.*release $[0-9\.]\+$.*$/\1/p');

elif [ -f &quot;/usr/local/cuda/version.txt&quot; ]; then
 # Determine CUDA version using /usr/local/cuda/version.txt file
 CUDA_VERSION=$(cat /usr/local/cuda/version.txt | sed 's/.* $[0-9]\+\.[0-9]\+$.*/\1/')

else
 CUDA_VERSION=&quot;&quot;

fi
</code></pre>
This environment variable is useful for downstream installations, such as when pip installing a copy of pytorch that was compiled for the correct CUDA version.
<pre class="lang-sh prettyprint-override"><code>python -m pip install \
 &quot;torch==1.9.0+cu${CUDA_VERSION/./}&quot; \
 &quot;torchvision==0.10.0+cu${CUDA_VERSION/./}&quot; \
 -f https://download.pytorch.org/whl/torch_stable.html
</code></pre>
Similarly, you could install the CPU version of pytorch when CUDA is not installed.
<pre class="lang-sh prettyprint-override"><code>if [ &quot;$CUDA_VERSION&quot; = &quot;&quot; ]; then
 MOD=&quot;+cpu&quot;;
 echo &quot;Warning: Installing CPU-only version of pytorch&quot;
else
 MOD=&quot;+cu${CUDA_VERSION/./}&quot;;
 echo &quot;Installing pytorch with $MOD&quot;
fi

python -m pip install \
 &quot;torch==1.9.0${MOD}&quot; \
 &quot;torchvision==0.10.0${MOD}&quot; \
 -f https://download.pytorch.org/whl/torch_stable.html
</code></pre>
But be careful with this because you can accidentally install a CPU-only version when you meant to have GPU support.
For example, if you run the install script on a server's login node which doesn't have GPUs and your jobs will be deployed onto nodes which do have GPUs. In this case, the login node will typically not have CUDA installed.

blocks|key|409468|text|使用tensorflow：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|409469|import+tensorflow+as+tf
from+tensorflow.python.platform+import+build_info+as+build
print(f"tensorflow+version:+{tf.__version__}")
print(f"Cuda+Version:+{build.build_info['cuda_version']}")
print(f"Cudnn+version:+{build.build_info['cudnn_version']}")|code-block|syntax|javascript|409470|tensorflow版本:+2.4.0|409471|Cuda版本:+11.0|409472|Cudnn版本:8|409473|entityMap^0|0|0|0|0|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]]|$1|G|3|H|5|6|7|Q|8|@]|9|@]|A|$]]|$1|I|3|J|5|6|7|R|8|@]|9|@]|A|$]]|$1|K|3|L|5|6|7|S|8|@]|9|@]|A|$]]|$1|M|3|-4|5|6|7|T|8|@]|9|@]|A|$]]]|N|$]]

Using tensorflow:
<pre><code>import tensorflow as tf
from tensorflow.python.platform import build_info as build
print(f&quot;tensorflow version: {tf.__version__}&quot;)
print(f&quot;Cuda Version: {build.build_info['cuda_version']}&quot;)
print(f&quot;Cudnn version: {build.build_info['cudnn_version']}&quot;)
</code></pre>
tensorflow version: 2.4.0
Cuda Version: 11.0
Cudnn version: 8

blocks|key|409519|text|在我的cuda-11.6.0安装中，可以在/usr/local/cuda/version.json中找到相关信息。它包含完整的版本号(11.6.0，而不是nvidia-smi显示的11.6。|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|409520|可以按如下方式检索信息：|409521|python+-c+'import+json;+print(json.load(open("/usr/local/cuda/version.json"))["cuda"]["version"])'|code-block|syntax|javascript|409522|entityMap^0|L|S|25|A|0|0|0^^$0|@$1|2|3|4|5|6|7|O|8|@$9|P|A|Q|B|C]|$9|R|A|S|B|C]]|D|@]|E|$]]|$1|F|3|G|5|6|7|T|8|@]|D|@]|E|$]]|$1|H|3|I|5|J|7|U|8|@]|D|@]|E|$K|L]]|$1|M|3|-4|5|6|7|V|8|@]|D|@]|E|$]]]|N|$]]

On my cuda-11.6.0 installation, the information can be found in <code>/usr/local/cuda/version.json</code>. It contains the full version number (11.6.0 instead of 11.6 as shown by <code>nvidia-smi</code>.
The information can be retrieved as follows:
<pre><code>python -c 'import json; print(json.load(open(&quot;/usr/local/cuda/version.json&quot;))[&quot;cuda&quot;][&quot;version&quot;])'
</code></pre>

blocks|key|409573|text|如果您安装了多个CUDA，则系统中加载的CUDA是与"nvcc“关联的CUDA。因此，"nvcc+--version“会显示您想要的内容。|type|unstyled|depth|inlineStyleRanges|entityRanges|data|409574|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|$]]

If you have multiple CUDA installed, the one loaded in your system is CUDA associated with &quot;nvcc&quot;. Therefore, &quot;nvcc --version&quot; shows what you want.

blocks|key|624666|text|您还可以尝试：|type|unstyled|depth|inlineStyleRanges|entityRanges|data|624667|nvidia-smi+-q+%7C+grep+CUDA|code-block|syntax|javascript|624668|它输出类似以下内容：|624669|CUDA+Version++++++++++++++++++++++++++++++:+11.1|624670|entityMap^0|0|0|0|0^^$0|@$1|2|3|4|5|6|7|M|8|@]|9|@]|A|$]]|$1|B|3|C|5|D|7|N|8|@]|9|@]|A|$E|F]]|$1|G|3|H|5|6|7|O|8|@]|9|@]|A|$]]|$1|I|3|J|5|D|7|P|8|@]|9|@]|A|$E|F]]|$1|K|3|-4|5|6|7|Q|8|@]|9|@]|A|$]]]|L|$]]

You could also try:
<pre><code>nvidia-smi -q | grep CUDA
</code></pre>
which outputs something like:
<pre><code>CUDA Version : 11.1
</code></pre>

blocks|key|624712|text|在使用CUDA+11.6.1的Windows+11上，这对我很有效：|type|unstyled|depth|inlineStyleRanges|offset|length|style|CODE|entityRanges|data|624713|cat+"C:\Program+Files\NVIDIA+GPU+Computing+Toolkit\CUDA\v11.6\version.json"|code-block|syntax|javascript|624714|entityMap^0|3|B|F|A|0|0^^$0|@$1|2|3|4|5|6|7|M|8|@$9|N|A|O|B|C]|$9|P|A|Q|B|C]]|D|@]|E|$]]|$1|F|3|G|5|H|7|R|8|@]|D|@]|E|$I|J]]|$1|K|3|-4|5|6|7|S|8|@]|D|@]|E|$]]]|L|$]]

On <code>Windows 11</code> with <code>CUDA 11.6.1</code>, this worked for me:
<pre><code>cat &quot;C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.6\version.json&quot;
</code></pre>

Is there any quick command or script to check for the version of CUDA installed?

I found the manual of 4.0 under the installation directory but I'm not sure whether it is of the actual installed version or not.

How to get the CUDA version?

集成开发环境

软件开发

服务器

编译器

Windows

Python

Ubuntu

Linux

是否有快速命令或脚本来检查所安装的CUDA版本？我在安装目录下找到了4.0的手册，但我不确定它是否是实际安装的版本。

问如何获取CUDA版本？
EN

社区

活动

资源

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问如何获取CUDA版本？EN

社区

活动

资源

关于

腾讯云开发者

热门产品

热门推荐

更多推荐

问如何获取CUDA版本？
EN