<p>AI supercomputers need a new kind of network to stay in sync at massive scale.<br>

<br>

OpenAI’s <a href="https://nitter.net/markjhandley" title="Mark Handley">@markjhandley</a> and <a href="https://nitter.net/poyntingatgreg" title="Greg">@poyntingatgreg</a> join <a href="https://nitter.net/AndrewMayne" title="Andrew Mayne">@AndrewMayne</a> to discuss what it takes to move data across record numbers of chips reliably and efficiently, the new Multipath Reliable Connection (MRC) networking protocol, and why it's available for the whole industry to use.</p>

<a href="https://nitter.net/OpenAI/status/2052039800384057348#m">

<br>Video<br>

<img src="https://nitter.net/pic/media%2FHHpPbXEbIAAKv6R.jpg" style="max-width:250px;" />

</a>

<hr/>

<blockquote>

<b>OpenAI (@OpenAI)</b>

<p>

<p>We’ve partnered with <a href="https://nitter.net/AMD" title="AMD">@AMD</a>, <a href="https://nitter.net/Broadcom" title="Broadcom">@Broadcom</a>, <a href="https://nitter.net/intel" title="Intel">@Intel</a>, <a href="https://nitter.net/Microsoft" title="Microsoft">@Microsoft</a>, and <a href="https://nitter.net/nvidia" title="NVIDIA">@NVIDIA</a>, to release Multipath Reliable Connection (MRC), a new open networking protocol that helps large AI training clusters run faster and more reliably, with less wasted GPU time.<br>

<br>

<a href="https://openai.com/index/mrc-supercomputer-networking/">openai.com/index/mrc-superco…</a></p>

<img src="https://nitter.net/pic/card_img%2F2051993862323765250%2FLRnB_5s0%3Fformat%3Djpg%26name%3D800x419" style="max-width:250px;" />

</p>

<footer>

— <cite><a href="https://nitter.net/OpenAI/status/2052025532485902368#m">https://nitter.net/OpenAI/status/2052025532485902368#m</a>

</footer>

</blockquote>

为什么值得关注

能改变理解方式，而不只是重复常识；符合当前抓取需求；它提供了新的理解或解释，而不只是表面观点

来源：x，领域：tech，保留分：0.61