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。

DRAM chip failures reveal surprising hardware vulnerabilities

91 点作者 mud_dauber超过 9 年前

10 条评论

fivesigma超过 9 年前
Thank you Intel for disabling ECC capabilities on consumer grade CPUs.<p>With smaller and more dense DRAM chips every year, it takes an even smaller amount of disturbance to flip a bit. ECC needs to go mainsteam yesterday.
评论 #10621252 未加载
评论 #10622155 未加载
PhantomGremlin超过 9 年前
They make an interesting claim:<p><pre><code> Between 12 percent and 45 percent of machines at Google experience at least one DRAM error per year. This is orders of magnitude more frequent than earlier estimates had suggested. </code></pre> We just had a big discussion on this topic a few days ago. <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=10598629" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=10598629</a>
评论 #10620567 未加载
crististm超过 9 年前
Judging from Jeff Atwood&#x27;s post only three days ago, a lot of people converted or seem to believe now that bit flips in RAMs are a thing of the past and ECC is just an &quot;enterprisey&quot; thing.<p><a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=10598629" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=10598629</a><p>He didn&#x27;t dare himself to not use ECC for the database... :)
评论 #10620933 未加载
评论 #10620586 未加载
nano_o超过 9 年前
The article seems to be based on the following paper, published in 2012: &quot;Cosmic Rays Don&#x27;t Strike Twice: Understanding the Nature of DRAM Errors and the Implications for System Design&quot;. Newer research on the topic has since been published, for example this year: &quot;Revisiting Memory Errors in Large-Scale Production Data Centers: Analysis and Modeling of New Trends from the Field&quot; (from Facebook) and &quot;Memory Errors in Modern Systems: The Good, The Bad, and The Ugly&quot;.
sevensor超过 9 年前
The idea that cosmic rays are responsible for DRAM failures seems laughable. I&#x27;m surprised anybody ever thought that.<p>I worked in a DRAM fab; I saw how it was made and tested. It&#x27;s a month-long process involving dozens of machines. Particulate contamination, watermarks, process variation -- there are tons of things that can go wrong. And then testing. There&#x27;s just no way to test thoroughly enough to catch all of the possible issues with timing and crosstalk. Not to mention electromigration and dielectric breakdown. You can do burn-in, but it&#x27;s not going to catch everything. We even know which die are most failure-prone (usually edge die, due to process uniformity issues.) If they pass test, they get shipped!<p><i>Of course</i> these issues are manufacturing related.
评论 #10622492 未加载
评论 #10622632 未加载
moconnor超过 9 年前
&quot;DRAM chips are a little like people: Their faults are not so much in their stars as in themselves. And like so many people, they can function perfectly well once they compensate for a few small flaws.&quot;<p>A beautifully whimsical way to end a fascinating technical article!
mschuster91超过 9 年前
Does anyone know of a consumer-grade laptop with ECC RAM?
评论 #10620327 未加载
评论 #10620339 未加载
评论 #10621739 未加载
评论 #10624253 未加载
gima超过 9 年前
Source paper appears to be: <a href="http:&#x2F;&#x2F;www.cs.toronto.edu&#x2F;~bianca&#x2F;papers&#x2F;ASPLOS2012.pdf" rel="nofollow">http:&#x2F;&#x2F;www.cs.toronto.edu&#x2F;~bianca&#x2F;papers&#x2F;ASPLOS2012.pdf</a><p>Based on the list of authors at the beginning of the article and the credited source given in the figures (&quot;Ioan Stefanovici, Andy Hwang &amp; Bianca Schroeder&quot; and &quot;Source: Hwang, Stefanovici, and Schroeder, Proceedings of ASPLOS XVII, 2012&quot;)
mozumder超过 9 年前
So how do we enable page retirement in popular operating systems on machines with ECC?
tempodox超过 9 年前
Can I pay the publisher to omit those blinking distractions? If I want to watch a movie while reading an article, I bring one myself.
评论 #10620466 未加载