https://milknocandy.github.io/tech/ Recent content in Teches on PaperMoon's blog https://milknocandy.github.io/%3Clink%20or%20path%20of%20image%20for%20opengraph,%20twitter-cards%3E https://milknocandy.github.io/%3Clink%20or%20path%20of%20image%20for%20opengraph,%20twitter-cards%3E Hugo -- 0.154.3 en Mon, 23 Mar 2026 12:30:23 +0800 https://milknocandy.github.io/tech/2026-1-20-nvcc/ Tue, 20 Jan 2026 17:32:00 +0800 https://milknocandy.github.io/tech/2026-1-20-nvcc/ <h2 id="1-installation">1 Installation</h2> <h3 id="11-installation-for-python-environment">1.1 Installation for python environment</h3> <p>To install <code>nvcc</code> inside a Python virtual environment, conda is all you need. It lets you pick the exact CUDA version for each env, so you never clash with system-wide installs. Please follow the steps below to install <code>nvcc</code>:</p> <ol> <li>Search for all available <code>nvcc</code> version</li> </ol> <div class="highlight"><div class="chroma"> <table class="lntable"><tr><td class="lntd"> <pre tabindex="0" class="chroma"><code><span class="lnt" id="hl-0-1"><a class="lnlinks" href="#hl-0-1"> 1</a> </span><span class="lnt" id="hl-0-2"><a class="lnlinks" href="#hl-0-2"> 2</a> </span><span class="lnt" id="hl-0-3"><a class="lnlinks" href="#hl-0-3"> 3</a> </span><span class="lnt" id="hl-0-4"><a class="lnlinks" href="#hl-0-4"> 4</a> </span><span class="lnt" id="hl-0-5"><a class="lnlinks" href="#hl-0-5"> 5</a> </span><span class="lnt" id="hl-0-6"><a class="lnlinks" href="#hl-0-6"> 6</a> </span><span class="lnt" id="hl-0-7"><a class="lnlinks" href="#hl-0-7"> 7</a> </span><span class="lnt" id="hl-0-8"><a class="lnlinks" href="#hl-0-8"> 8</a> </span><span class="lnt" id="hl-0-9"><a class="lnlinks" href="#hl-0-9"> 9</a> </span><span class="lnt" id="hl-0-10"><a class="lnlinks" href="#hl-0-10">10</a> </span><span class="lnt" id="hl-0-11"><a class="lnlinks" href="#hl-0-11">11</a> </span><span class="lnt" id="hl-0-12"><a class="lnlinks" href="#hl-0-12">12</a> </span><span class="lnt" id="hl-0-13"><a class="lnlinks" href="#hl-0-13">13</a> </span><span class="lnt" id="hl-0-14"><a class="lnlinks" href="#hl-0-14">14</a> </span><span class="lnt" id="hl-0-15"><a class="lnlinks" href="#hl-0-15">15</a> </span><span class="lnt" id="hl-0-16"><a class="lnlinks" href="#hl-0-16">16</a> </span><span class="lnt" id="hl-0-17"><a class="lnlinks" href="#hl-0-17">17</a> </span><span class="lnt" id="hl-0-18"><a class="lnlinks" href="#hl-0-18">18</a> </span><span class="lnt" id="hl-0-19"><a class="lnlinks" href="#hl-0-19">19</a> </span><span class="lnt" id="hl-0-20"><a class="lnlinks" href="#hl-0-20">20</a> </span></code></pre></td> <td class="lntd"> <pre tabindex="0" class="chroma"><code class="language-bash" data-lang="bash"><span class="line"><span class="cl"><span class="c1"># activate your conda environment first</span> </span></span><span class="line"><span class="cl">$ conda activate xxenv </span></span><span class="line"><span class="cl"><span class="c1"># search for available cudatoolkit versions</span> </span></span><span class="line"><span class="cl">$ conda search cudatoolkit --channel conda-forge </span></span><span class="line"><span class="cl"><span class="c1"># output like this:</span> </span></span><span class="line"><span class="cl">Loading channels: <span class="k">done</span> </span></span><span class="line"><span class="cl"><span class="c1"># Name Version Build Channel </span> </span></span><span class="line"><span class="cl">cudatoolkit 5.5rc1 p0 anaconda/pkgs/pro </span></span><span class="line"><span class="cl">cudatoolkit 5.5.1 p0 anaconda/pkgs/pro </span></span><span class="line"><span class="cl">cudatoolkit 6.0 p0 anaconda/pkgs/pro </span></span><span class="line"><span class="cl">cudatoolkit 7.0 <span class="m">1</span> anaconda/pkgs/pro </span></span><span class="line"><span class="cl">cudatoolkit 7.5 <span class="m">0</span> anaconda/pkgs/free </span></span><span class="line"><span class="cl">cudatoolkit 7.5 <span class="m">2</span> anaconda/pkgs/free </span></span><span class="line"><span class="cl">cudatoolkit 8.0 <span class="m">1</span> anaconda/pkgs/free </span></span><span class="line"><span class="cl">cudatoolkit 8.0 <span class="m">3</span> anaconda/pkgs/free </span></span><span class="line"><span class="cl">cudatoolkit 9.0 h13b8566_0 anaconda/pkgs/main </span></span><span class="line"><span class="cl">cudatoolkit 9.0 h13b8566_0 pkgs/main </span></span><span class="line"><span class="cl">cudatoolkit 9.2 <span class="m">0</span> anaconda/pkgs/main </span></span><span class="line"><span class="cl">cudatoolkit 9.2 <span class="m">0</span> pkgs/main </span></span><span class="line"><span class="cl">... </span></span></code></pre></td></tr></table> </div> </div><ol start="2"> <li>Install specific version of <code>nvcc</code> If you install nvcc within the base environment, you may run into errors such as <code>OSError: [Errno 39] Directory not empty: 'xxx/anaconda3/lib/ossl-modules'</code>. Therefore, I recommend installing it in a manually created virtual environment instead.</li> </ol> <div class="highlight"><div class="chroma"> <table class="lntable"><tr><td class="lntd"> <pre tabindex="0" class="chroma"><code><span class="lnt" id="hl-1-1"><a class="lnlinks" href="#hl-1-1">1</a> </span><span class="lnt" id="hl-1-2"><a class="lnlinks" href="#hl-1-2">2</a> </span></code></pre></td> <td class="lntd"> <pre tabindex="0" class="chroma"><code class="language-bash" data-lang="bash"><span class="line"><span class="cl"><span class="c1"># Install specific version of `nvcc`</span> </span></span><span class="line"><span class="cl">$ conda install -c nvidia <span class="nv">cudatoolkit</span><span class="o">=</span>11.8 -y <span class="c1"># or conda install -c conda-forge cudatoolkit=x.x</span> </span></span></code></pre></td></tr></table> </div> </div><p>Now let us check what we have obtained by running this.</p>