https://ll.zinc.runhttps://ll.zinc.run/posts/101296https://ll.zinc.run/posts/101296Sat, 31 Oct 2026 05:25:36 GMT<b><span class="post-title-source">单机显存塞不下大模型，多台机器又各跑各的？ mesh-llm 把多机 GPU 和内存拼成一张网，对外就是一个 OpenAI 兼容接口，模型太大还能分段跑，节点随时加。</span> 适合手头有几台机器想合起来跑模型的人，先从两台试起。 你会先拼几台，两台还是更多？ </b><a href="/search/result?q=%23%E5%88%86%E5%B8%83%E5%BC%8F" title="#分布式"><b>#分布式</b></a><b> </b><a href="/search/result?q=%23LLM" title="#LLM"><b>#LLM</b></a><b> </b><a href="https://t.me/grok" target="_blank" title="@grok"><b>@grok</b></a><b> </b><a href="https://github.com/Mesh-LLM/mesh-llm" target="_blank" rel="noopener" title="https://github.com/Mesh-LLM/mesh-llm"><b>https://github.com/Mesh-LLM/mesh-llm</b></a><br /><br /><i class="emoji"><b>🌐</b></i> 链接: <a href="https://x.com/xiaoanlab/status/2116394483344400407" target="_blank" rel="noopener" title="https://x.com/xiaoanlab/status/2116394483344400407">https://x.com/xiaoanlab/status/2116394483344400407</a><br /><i class="emoji"><b>🏷️</b></i> 分组: X｜为你推荐<br /><i class="emoji"><b>🕒</b></i> 时间: 2026-10-31 12:59:00<a class="tgme_widget_message_link_preview" href="https://github.com/Mesh-LLM/mesh-llm" target="_blank" rel="noopener" title="Distributed AI/LLM for the people. Share compute privately or publicly to power your agents and chat. - Mesh-LLM/mesh-llm"> <div class="link_preview_site_name accent_color">GitHub</div> <img class="link_preview_image" alt="GitHub - Mesh-LLM/mesh-llm: Distributed AI/LLM for the people. Share compute privately or publicly to power your agents and chat." src="/static/https://cdn4.telesco.pe/file/m4I_ewM_I1BFRYpo8nnqgwg8d-2WN2WpQqyb4iDOM_aojAvUbcHLEyRu6NjNUObeDcOqqWdnipBs7VAWCAQx9fE8KZJixHfxb_zU91Kq82nd2ng49b7_LCSpQgRYMJf-i7t3FlgbBH4jL3fg-aVZhRp8l9DCFNj3zSiffNuvDZOOugyZlRvLuoDHbr7HPkQVKyjO0WCf-6w0w5h6906eX3s2wHgG-sqUeImXaILYbrrpQfhzvkG1LCalTjMflsIzjRYySZMva762j7udcYxZqQivva4Zu-fajN6tT7DInBK6Crz8mregIay4myNZMoZRLdKpcoe-uWwNzaMdoNehZQ.jpg" width="1200" height="630" loading="eager" /> <div class="link_preview_title">GitHub - Mesh-LLM/mesh-llm: Distributed AI/LLM for the people. Share compute privately or publicly to power your agents and chat.</div> <div class="link_preview_description">Distributed AI/LLM for the people. Share compute privately or publicly to power your agents and chat. - Mesh-LLM/mesh-llm</div> </a>