<div id="readability-page-1" class="page"><div id="entry-view">
        
        <main>
          <div>
              <article id="mainContent">
                <div>
                  <p><a href="/m/category"><span>카테고리 없음</span></a></p><h3>Windows WSL CUDA 설치</h3>
                </div>
                <p><cite>김 정출</cite>
                  <span> 2022. 9. 2. 11:34</span>
                </p>
                                <div>
                                <h2 data-ke-size="size26"><span>Windows&nbsp;WSL&nbsp;CUDA&nbsp;설치</span></h2>

<figure data-ke-mobilestyle="widthOrigin" data-origin-width="540" data-origin-height="304">
    <span data-url="https://k.kakaocdn.net/dn/HydRG/btrK9bpXr1Z/KIseukqg7PT2O10oIm1l01/img.png" data-lightbox="lightbox">
        <img src="https://k.kakaocdn.net/dn/HydRG/btrK9bpXr1Z/KIseukqg7PT2O10oIm1l01/img.png" onerror="this.onerror=null; this.src='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png'; this.srcset='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png';" width="540" height="304" data-origin-width="540" data-origin-height="304">
    </span>
    
</figure>
<p data-ke-size="size16"><span>Windows WSL에 CUDA를 설치하기 위해 GPU 드라이버 설치를 진행한다.</span></p>
<p data-ke-size="size16"><span>아래 링크로 이동한다.</span></p>
<p data-ke-size="size16"><span></span><a href="https://developer.nvidia.com/cuda/wsl">https://developer.nvidia.com/cuda/wsl</a><b></b></p>
<p data-ke-size="size16"><span>Get CUDA Driver 버튼을 클릭하여 Nvidia GPU driver를 다운로드 및 설치해준다.</span></p>
<figure data-ke-mobilestyle="widthOrigin" data-origin-width="1576" data-origin-height="1167">
    <span data-url="https://k.kakaocdn.net/dn/ZTxxi/btrK96ojeTi/zKAhmzTfKrv7I9oaHRyWIk/img.png" data-lightbox="lightbox">
        <img src="https://k.kakaocdn.net/dn/ZTxxi/btrK96ojeTi/zKAhmzTfKrv7I9oaHRyWIk/img.png" onerror="this.onerror=null; this.src='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png'; this.srcset='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png';" width="718" height="532" data-origin-width="1576" data-origin-height="1167">
    </span>
    
</figure>

<p data-ke-size="size16"><span>GPU 카드를 선택한다.</span></p>
<figure data-ke-mobilestyle="widthOrigin" data-origin-width="972" data-origin-height="386">
    <span data-url="https://k.kakaocdn.net/dn/MZqGZ/btrLcJkVKeu/Yt4wvKZMteIy6GWOKWTUrk/img.png" data-lightbox="lightbox">
        <img src="https://k.kakaocdn.net/dn/MZqGZ/btrLcJkVKeu/Yt4wvKZMteIy6GWOKWTUrk/img.png" onerror="this.onerror=null; this.src='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png'; this.srcset='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png';" width="718" height="285" data-origin-width="972" data-origin-height="386">
    </span>
    
</figure>

<p data-ke-size="size16"><span>설치 파일을 다운받아 실행하여 설치를 진행한다.</span></p>
<figure data-ke-mobilestyle="widthOrigin" data-origin-width="1600" data-origin-height="802">
    <span data-url="https://k.kakaocdn.net/dn/dQXJHA/btrK89MsGsS/mcYnCkKzPKDMj4rFusOK40/img.png" data-lightbox="lightbox">
        <img src="https://k.kakaocdn.net/dn/dQXJHA/btrK89MsGsS/mcYnCkKzPKDMj4rFusOK40/img.png" onerror="this.onerror=null; this.src='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png'; this.srcset='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png';" width="718" height="360" data-origin-width="1600" data-origin-height="802">
    </span>
    
</figure>

<p data-ke-size="size16"><span>설치 이후 재시작을 진행한다.</span></p>

<p data-ke-size="size16"><span>WSL에 진입하여 해당 명령어로 CUDA 설치를 진행한다.</span></p>
<pre id="code_1662086053631" data-ke-language="bash" data-ke-type="codeblock"><code>wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin

sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600

wget https://developer.download.nvidia.com/compute/cuda/11.4.0/local_installers/cuda-repo-wsl-ubuntu-11-4-local_11.4.0-1_amd64.deb

sudo dpkg -i cuda-repo-wsl-ubuntu-11-4-local_11.4.0-1_amd64.deb

sudo apt-key add /var/cuda-repo-wsl-ubuntu-11-4-local/7fa2af80.pub

sudo apt-get update

sudo apt-get -y install cuda</code></pre>
<p data-ke-size="size16"><span>nvidia-smi를 사용하기 위한 util 설치</span></p>
<pre id="code_1662086058746" data-ke-language="bash" data-ke-type="codeblock"><code>$ sudo apt install nvidia-utils-515</code></pre>

<p data-ke-size="size16"><span>GPU Driver 확인</span></p>
<pre id="code_1662086064279" data-ke-language="bash" data-ke-type="codeblock"><code>$ nvidia-smi</code></pre>
<figure data-ke-mobilestyle="widthOrigin" data-origin-width="1313" data-origin-height="521">
    <span data-url="https://k.kakaocdn.net/dn/cb6RKR/btrLc8dQ5MF/WGu06ChZkedhTrWGbj6W0K/img.png" data-lightbox="lightbox">
        <img src="https://k.kakaocdn.net/dn/cb6RKR/btrLc8dQ5MF/WGu06ChZkedhTrWGbj6W0K/img.png" onerror="this.onerror=null; this.src='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png'; this.srcset='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png';" width="718" height="285" data-origin-width="1313" data-origin-height="521">
    </span>
    
</figure>
<p data-ke-size="size16">Docker 설치 이후에 GPU 사용도 가능<b></b></p>
<pre id="code_1662092921038" data-ke-language="bash" data-ke-type="codeblock"><code>$ docker run --gpus all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark</code></pre>
<figure data-ke-mobilestyle="widthOrigin" data-origin-width="1522" data-origin-height="604">
    <span data-url="https://k.kakaocdn.net/dn/bq1eU1/btrLcJllWdh/dHtwngYFCDhZ1SAcexAzhk/img.png" data-lightbox="lightbox">
        <img src="https://k.kakaocdn.net/dn/bq1eU1/btrLcJllWdh/dHtwngYFCDhZ1SAcexAzhk/img.png" onerror="this.onerror=null; this.src='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png'; this.srcset='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png';" data-origin-width="1522" data-origin-height="604">
    </span>
    
</figure>

<pre id="code_1662093166872" data-ke-language="bash" data-ke-type="codeblock"><code>$ docker run --rm --gpus all nvidia/cuda:11.7.1-base-ubuntu18.04 nvidia-smi</code></pre>
<figure data-ke-mobilestyle="widthOrigin" data-origin-width="1310" data-origin-height="713">
    <span data-url="https://k.kakaocdn.net/dn/bMOOJ3/btrLaCOeKNk/x87Eg4GIPmAN8bPJvIsreK/img.png" data-lightbox="lightbox">
        <img src="https://k.kakaocdn.net/dn/bMOOJ3/btrLaCOeKNk/x87Eg4GIPmAN8bPJvIsreK/img.png" onerror="this.onerror=null; this.src='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png'; this.srcset='//t1.daumcdn.net/tistory_admin/static/images/no-image-v1.png';" data-origin-width="1310" data-origin-height="713">
    </span>
    
</figure>
<p data-ke-size="size16"><span>감사합니다.</span></p>

            
            
                </div>
                            
            
              </article>
              
              
              
              
              
                
              
            </div>
        </main>
        
      </div></div>

Jeongchul

Windows WSL CUDA 설치
 
Windows WSL에 CUDA를 설치하기 위해 GPU 드라이버 설치를 진행한다.
아래 링크로 이동한다.
https://developer.nvidia.com/cuda/wsl
Get CUDA Driver 버튼을 클릭하여 Nvidia GPU driver를 다운로드 및 설치해준다.
 
GPU 카드를 선택한다.
 
설치 파일을 다운받아 실행하여 설치를 진행한다.
 
설치 이후 재시작을 진행한다.
 
WSL에 진입하여 해당 명령어로 CUDA 설치를 진행한다.
wget https://developer.download.nvidia.com/compute/cuda/repos/wsl-ubuntu/x86_64/cuda-wsl-ubuntu.pin
sudo mv cuda-wsl-ubuntu.pin /etc/apt/preferences.d/cuda-repository-pin-600
wget https://developer.download.nvidia.com/compute/cuda/11.4.0/local_installers/cuda-repo-wsl-ubuntu-11-4-local_11.4.0-1_amd64.deb
sudo dpkg -i cuda-repo-wsl-ubuntu-11-4-local_11.4.0-1_amd64.deb
sudo apt-key add /var/cuda-repo-wsl-ubuntu-11-4-local/7fa2af80.pub
sudo apt-get update
sudo apt-get -y install cuda
nvidia-smi를 사용하기 위한 util 설치
$ sudo apt install nvidia-utils-515
 
GPU Driver 확인
$ nvidia-smi
Docker 설치 이후에 GPU 사용도 가능
$ docker run --gpus all nvcr.io/nvidia/k8s/cuda-sample:nbody nbody -gpu -benchmark
 
$ docker run --rm --gpus all nvidia/cuda:11.7.1-base-ubuntu18.04 nvidia-smi
감사합니다.

Windows WSL CUDA 설치

댓글