TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Larry Hastings – Python's Infamous GIL – PyCon 2016

11 点作者 cjdrake将近 9 年前

2 条评论

tedmiston将近 9 年前
It&#x27;s helpful to know the abstract for Larry&#x27;s talk before watching:<p><pre><code> DESCRIPTION CPython&#x27;s GIL means your Python code can only run on one CPU core at a time. Can we remove it? Yes, we can... in fact we already have! But is it worth the cost? ABSTRACT CPython&#x27;s &quot;Global Interpreter Lock&quot;, or &quot;GIL&quot;, was added in 1992. It was an excellent design decision. But 24 years is a long time--today it prevents Python from capitalizing on multiple CPUs. Many people want us to remove the GIL. It turns out, removing the GIL isn&#x27;t actually that hard. In fact, I already removed it, in my experimental &quot;gilectomy&quot; branch. But the GIL is one reason CPython is so fast! The &quot;gilectomy&quot; makes CPython shockingly slow. This talk will discuss the history of the GIL, how the GIL helps make CPython fast, how the &quot;gilectomy&quot; removed the GIL, and some ways we might be able to make the &quot;gilectomy&quot; version fast enough to be useful. </code></pre> from <a href="https:&#x2F;&#x2F;us.pycon.org&#x2F;2016&#x2F;schedule&#x2F;presentation&#x2F;2101&#x2F;" rel="nofollow">https:&#x2F;&#x2F;us.pycon.org&#x2F;2016&#x2F;schedule&#x2F;presentation&#x2F;2101&#x2F;</a>
harlowja将近 9 年前
Nice handling of the crappy question at the end...