{"id":110668,"date":"2025-06-02T14:59:44","date_gmt":"2025-06-02T06:59:44","guid":{"rendered":"https:\/\/www.ioiotimes.com\/?p=110668"},"modified":"2025-06-02T15:21:51","modified_gmt":"2025-06-02T07:21:51","slug":"red-hat%e6%8e%a8%e5%87%ballm-d%e7%a4%be%e7%be%a4%e5%b0%88%e6%a1%88%ef%bc%8c%e5%8a%a0%e9%80%9f%e6%8e%a8%e9%80%b2%e5%a4%a7%e8%a6%8f%e6%a8%a1%e5%88%86%e6%95%a3%e5%bc%8f%e7%9a%84%e7%94%9f%e6%88%90","status":"publish","type":"post","link":"https:\/\/www.ioiotimes.com\/?p=110668","title":{"rendered":"Red Hat\u63a8\u51fallm-d\u793e\u7fa4\u5c08\u6848\uff0c\u52a0\u901f\u63a8\u9032\u5927\u898f\u6a21\u5206\u6563\u5f0f\u7684\u751f\u6210\u5f0fAI\u63a8\u8ad6"},"content":{"rendered":"\n<p>\u4e16\u754c\u9818\u5148\u958b\u653e\u539f\u59cb\u78bc\u8edf\u9ad4\u89e3\u6c7a\u65b9\u6848\u4f9b\u61c9\u5546&nbsp;Red Hat&nbsp;\u65bc&nbsp;Red Hat Summit 2025&nbsp;\u6b63\u5f0f\u5ba3\u5e03\u63a8\u51fa\u5168\u65b0\u958b\u6e90\u5c08\u6848&nbsp;llm-d\uff0c\u65e8\u5728\u56de\u61c9\u672a\u4f86\u751f\u6210\u5f0f&nbsp;AI\uff08Gen AI\uff09\u6700\u95dc\u9375\u7684\u5927\u898f\u6a21\u63a8\u8ad6\u9700\u6c42\u3002llm-d&nbsp;\u904b\u7528\u7a81\u7834\u6027\u7684\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u5927\u898f\u6a21\u63a8\u8ad6\u6280\u8853\uff0c\u4e26\u63a1\u7528\u539f\u751f&nbsp;Kubernetes&nbsp;\u67b6\u69cb\u3001\u57fa\u65bc&nbsp;vLLM&nbsp;\u7684\u5206\u6563\u5f0f\u63a8\u8ad6\uff0c\u4ee5\u53ca\u667a\u6167\u578b&nbsp;AI&nbsp;\u611f\u77e5\u7db2\u8def\u8def\u7531\uff0c\u6253\u9020\u51fa\u5f37\u5927\u7684\u5927\u578b\u8a9e\u8a00\u6a21\u578b\uff08LLM\uff09\u63a8\u8ad6\u96f2\u7aef\uff0c\u4ee5\u6eff\u8db3\u6700\u56b4\u82db\u7684\u751f\u7522\u670d\u52d9\u6c34\u6e96\u76ee\u6a19\uff08service-level objectives\uff0cSLO\uff09\u3002<\/p>\n\n\n\n<p>\u5118\u7ba1\u6a21\u578b\u8a13\u7df4\u4ecd\u81f3\u95dc\u91cd\u8981\uff0c\u4f46\u751f\u6210\u5f0f\u00a0AI\u00a0\u7684\u771f\u6b63\u50f9\u503c\u66f4\u4ef0\u8cf4\u6709\u6548\u7387\u4e14\u53ef\u64f4\u5c55\u7684\u63a8\u8ad6\u80fd\u529b\uff0c\u9019\u624d\u662f\u5c07\u00a0AI\u00a0\u6a21\u578b\u8f49\u5316\u70ba\u53ef\u57f7\u884c\u7684\u6d1e\u5bdf\u548c\u4f7f\u7528\u8005\u9ad4\u9a57\u7684\u5f15\u64ce\u3002\u6839\u64da\u00a0Gartner<sup>[1]<\/sup>\u00a0\u9810\u6e2c\u6307\u51fa\uff0c\u96a8\u8457\u5e02\u5834\u8da8\u65bc\u6210\u719f\uff0c2028\u00a0\u5e74\u8d85\u904e\u00a080%\u00a0\u7684\u8cc7\u6599\u4e2d\u5fc3\u5de5\u4f5c\u8ca0\u8f09\u52a0\u901f\u5668\u5c07\u5c08\u9580\u7528\u65bc\u63a8\u8ad6\uff0c\u800c\u975e\u8a13\u7df4\u7528\u9014\uff0c\u6b64\u8da8\u52e2\u51f8\u986f\u51fa\u751f\u6210\u5f0f\u00a0AI\u00a0\u7684\u672a\u4f86\u53d6\u6c7a\u65bc\u57f7\u884c\u80fd\u529b\u3002\u7136\u800c\uff0c\u96a8\u8457\u63a8\u7406\u6a21\u578b\u65e5\u76ca\u8907\u96dc\u4e14\u9f90\u5927\uff0c\u5c0d\u8cc7\u6e90\u7684\u9700\u6c42\u4e0d\u65b7\u6500\u5347\uff0c\u7a2e\u7a2e\u56e0\u7d20\u4e0d\u50c5\u9650\u5236\u4e86\u96c6\u4e2d\u5f0f\u63a8\u8ad6\u7684\u53ef\u884c\u6027\uff0c\u5176\u884d\u751f\u7684\u9ad8\u6210\u672c\u548c\u56b4\u91cd\u7684\u5ef6\u9072\u66f4\u53ef\u80fd\u6210\u70ba\u00a0AI\u00a0\u5275\u65b0\u7684\u74f6\u9838\u3002<\/p>\n\n\n\n<p>Red Hat&nbsp;\u8cc7\u6df1\u526f\u7e3d\u88c1\u66a8&nbsp;AI&nbsp;\u6280\u8853\u9577&nbsp;Brian Stevens&nbsp;\u8868\u793a\uff1a\u300c\u7531\u773e\u591a&nbsp;AI&nbsp;\u9818\u5c0e\u8005\u652f\u6301\u7684&nbsp;llm-d&nbsp;\u793e\u7fa4\u7684\u555f\u52d5\uff0c\u8c61\u5fb5\u8457\u6211\u5011\u6b63\u8655\u65bc\u6eff\u8db3\u53ef\u64f4\u5c55\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u63a8\u8ad6\u9700\u6c42\u7684\u95dc\u9375\u6642\u523b\uff0c\u4ea6\u70ba\u4f01\u696d\u5be6\u73fe\u66f4\u5ee3\u6cdb\u7684&nbsp;AI&nbsp;\u61c9\u7528\u6642\u5fc5\u9808\u514b\u670d\u7684\u91cd\u5927\u6311\u6230\u3002\u900f\u904e\u904b\u7528&nbsp;vLLM&nbsp;\u7684\u5275\u65b0\u6280\u8853\u548c&nbsp;Kubernetes&nbsp;\u7d93\u9a57\u8b49\u7684\u80fd\u529b\uff0cllm-d&nbsp;\u5354\u52a9\u4f01\u696d\u66f4\u9806\u66a2\u5730\u5728\u64f4\u5c55\u7684\u8de8\u6df7\u5408\u96f2\u74b0\u5883\u4e2d\u5be6\u73fe\u5206\u6563\u5f0f\u3001\u53ef\u64f4\u5c55\u4e14\u9ad8\u6548\u80fd\u7684&nbsp;AI&nbsp;\u63a8\u8ad6\uff0c\u80fd\u5920\u652f\u63f4\u4efb\u4f55\u6a21\u578b\u3001\u4efb\u4f55\u52a0\u901f\u5668\uff0c\u5728\u4efb\u4f55\u96f2\u7aef\u74b0\u5883\u4e2d\u904b\u884c\uff0c\u5354\u52a9\u5be6\u73fe&nbsp;AI&nbsp;\u7121\u9650\u6f5b\u529b\u7684\u9858\u666f\u3002\u300d<em><\/em><\/p>\n\n\n\n<p><\/p>\n\n\n\n<figure class=\"wp-block-image size-full\"><img loading=\"lazy\" decoding=\"async\" width=\"1200\" height=\"628\" src=\"https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2025\/06\/20250602_NEWS_1-1.jpg\" alt=\"\" class=\"wp-image-110669\" title=\"\" srcset=\"https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2025\/06\/20250602_NEWS_1-1.jpg 1200w, https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2025\/06\/20250602_NEWS_1-1-300x157.jpg 300w, https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2025\/06\/20250602_NEWS_1-1-1024x536.jpg 1024w, https:\/\/www.ioiotimes.com\/wordpress\/wp-content\/uploads\/2025\/06\/20250602_NEWS_1-1-768x402.jpg 768w\" sizes=\"(max-width: 1200px) 100vw, 1200px\" \/><\/figure>\n\n\n\n<p><\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>\u501f\u52a9llm-d\u6eff\u8db3\u5c0d\u53ef\u64f4\u5c55\u751f\u6210\u5f0fAI\u63a8\u8ad6\u7684\u9700\u6c42<\/strong><\/h3>\n\n\n\n<p>\u70ba\u61c9\u5c0d\u4e0a\u8ff0\u6311\u6230\uff0cRed Hat&nbsp;\u651c\u624b\u696d\u754c\u5925\u4f34\u5171\u540c\u63a8\u51fa&nbsp;llm-d\u3002\u9019\u9805\u5177\u524d\u77bb\u6027\u7684\u5c08\u6848\u4e0d\u50c5\u80fd\u5f37\u5316&nbsp;vLLM&nbsp;\u7684\u80fd\u529b\u4ee5\u7a81\u7834\u55ae\u4e00\u4f3a\u670d\u5668\u7684\u9650\u5236\uff0c\u540c\u6642\u4e5f\u70ba&nbsp;AI&nbsp;\u63a8\u8ad6\u91cb\u653e\u5927\u898f\u6a21\u751f\u7522\u7684\u6f5b\u529b\u3002llm-d&nbsp;\u904b\u7528&nbsp;Kubernetes&nbsp;\u7d93\u9a57\u8b49\u4e14\u5f37\u5927\u7684\u8abf\u5ea6\u80fd\u529b\uff0c\u5c07\u9032\u968e\u63a8\u8ad6\u529f\u80fd\u7121\u7e2b\u6574\u5408\u81f3\u4f01\u696d\u65e2\u6709\u7684&nbsp;IT&nbsp;\u57fa\u790e\u67b6\u69cb\u4e2d\u3002IT&nbsp;\u5718\u968a\u5f97\u4ee5\u65bc\u7d71\u4e00\u5e73\u53f0\u4e0a\u6eff\u8db3\u95dc\u9375\u696d\u52d9\u5de5\u4f5c\u8ca0\u8f09\u7684\u591a\u6a23\u5316\u670d\u52d9\u9700\u6c42\uff0c\u540c\u6642\u900f\u904e\u90e8\u7f72\u5275\u65b0\u6280\u8853\u5c07\u6548\u7387\u6975\u5927\u5316\uff0c\u4e26\u5927\u5e45\u964d\u4f4e\u9ad8\u6548\u80fd&nbsp;AI&nbsp;\u52a0\u901f\u5668\u5e36\u4f86\u7684\u7e3d\u9ad4\u64c1\u6709\u6210\u672c\uff08TCO\uff09\u3002<\/p>\n\n\n\n<p>llm-d&nbsp;\u63d0\u4f9b\u4e00\u7cfb\u5217\u5f37\u5927\u7684\u5275\u65b0\u529f\u80fd\uff0c\u4eae\u9ede\u5305\u62ec\uff1a<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>\u25cf\u3000<strong>vLLM\u00a0\u8fc5\u901f\u6210\u70ba\u958b\u6e90\u9818\u57df\u5be6\u8cea\u4e0a\u7684\u6a19\u6e96\u63a8\u8ad6\u4f3a\u670d\u5668<\/strong>\uff1a\u70ba\u65b0\u8208\u6a21\u578b\u63d0\u4f9b\u00a0Day 0\u00a0\u6a21\u578b\u652f\u63f4\uff0c\u53ef\u7528\u65bc\u591a\u7a2e\u52a0\u901f\u5668\uff0c\u5305\u62ec\u00a0Google Cloud\u00a0Tensor Processor Units\uff08TPU\uff09\u3002<\/li>\n\n\n\n<li>\u25cf\u3000<strong>\u9810\u586b\u8207\u89e3\u78bc\u5206\u96e2<\/strong>\uff1a\u5c07\u00a0AI\u00a0\u7684\u8f38\u5165\u5167\u5bb9\u548c\u6b0a\u6756\u751f\u6210\u968e\u6bb5\u5206\u96e2\u70ba\u7368\u7acb\u7684\u904b\u7b97\u4f5c\u696d\uff0c\u4e26\u5c07\u9019\u4e9b\u4f5c\u696d\u5206\u6563\u5230\u591a\u500b\u4f3a\u670d\u5668\u4e0a\u57f7\u884c\u3002<\/li>\n\n\n\n<li>\u25cf\u3000<strong>\u57fa\u65bc\u00a0LMCache\u00a0\u7684\u9375\u503c\uff08key-value\uff0cKV\uff09\u5feb\u53d6\u5378\u8f09<\/strong>\uff1a\u5c07\u00a0KV\u00a0\u5feb\u53d6\u7684\u8a18\u61b6\u9ad4\u8ca0\u8f09\u5f9e\u00a0GPU\u00a0\u8a18\u61b6\u9ad4\u4e2d\u79fb\u8f49\u5230\u6210\u672c\u6548\u76ca\u66f4\u9ad8\u4e14\u8cc7\u6e90\u66f4\u8c50\u5bcc\u7684\u6a19\u6e96\u5132\u5b58\u88dd\u7f6e\uff0c\u4f8b\u5982\u00a0CPU\u00a0\u8a18\u61b6\u9ad4\u6216\u7db2\u8def\u5132\u5b58\u3002<\/li>\n\n\n\n<li>\u25cf\u3000<strong>\u7531\u00a0Kubernetes\u00a0\u9a45\u52d5\u7684\u53e2\u96c6\u548c\u63a7\u5236\u5668<\/strong>\uff1a\u53ef\u5728\u5de5\u4f5c\u8ca0\u8f09\u9700\u6c42\u6ce2\u52d5\u6642\uff0c\u66f4\u6709\u6548\u5730\u8abf\u5ea6\u904b\u7b97\u8207\u5132\u5b58\u8cc7\u6e90\uff0c\u540c\u6642\u7dad\u6301\u6548\u80fd\u4e26\u964d\u4f4e\u5ef6\u9072\u3002<\/li>\n\n\n\n<li>\u25cf\u3000<strong>AI\u00a0\u611f\u77e5\u7db2\u8def\u8def\u7531<\/strong>\uff1a\u5c07\u50b3\u5165\u8acb\u6c42\u6392\u7a0b\u81f3\u6700\u6709\u53ef\u80fd\u64c1\u6709\u5148\u524d\u63a8\u8ad6\u904b\u7b97\u71b1\u5feb\u53d6\u7684\u4f3a\u670d\u5668\u548c\u52a0\u901f\u5668\u3002<\/li>\n\n\n\n<li>\u25cf\u3000<strong>\u9ad8\u6548\u80fd\u901a\u8a0a\u61c9\u7528\u7a0b\u5f0f\u4ecb\u9762<\/strong>\uff1a\u53ef\u5728\u4f3a\u670d\u5668\u4e4b\u9593\u5be6\u73fe\u66f4\u70ba\u5feb\u901f\u9ad8\u6548\u7684\u8cc7\u6599\u50b3\u8f38\uff0c\u4e26\u652f\u63f4\u00a0NVIDIA Inference Xfer Library\uff08NIXL\uff09\u3002<\/li>\n<\/ul>\n\n\n\n<p><\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>llm-d\u7372\u696d\u754c\u9818\u5c0e\u8005\u652f\u6301<\/strong><\/h3>\n\n\n\n<p>\u9019\u9805\u5168\u65b0\u958b\u6e90\u5c08\u6848\u5df2\u7372\u5f97\u7531\u9802\u5c16\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u6a21\u578b\u4f9b\u61c9\u5546\u3001AI&nbsp;\u52a0\u901f\u5668\u9818\u822a\u8005\u4ee5\u53ca\u4e00\u6d41&nbsp;AI&nbsp;\u96f2\u7aef\u5e73\u53f0\u7d44\u6210\u7684\u5f37\u5927\u806f\u76df\u9f0e\u529b\u652f\u6301\u3002CoreWeave\u3001Google Cloud\u3001IBM Research&nbsp;\u548c&nbsp;NVIDIA&nbsp;\u70ba\u5275\u59cb\u8ca2\u737b\u8005\uff0cAMD\u3001Cisco\u3001Hugging Face\u3001Intel\u3001Lambda&nbsp;\u548c&nbsp;Mistral AI&nbsp;\u5247\u70ba\u5408\u4f5c\u5925\u4f34\uff0c\u6b64\u9663\u5bb9\u51f8\u986f\u51fa\u696d\u754c\u70ba\u6253\u9020\u5927\u898f\u6a21&nbsp;LLM&nbsp;\u670d\u52d9\u7684\u672a\u4f86\u6240\u5c55\u73fe\u7684\u6df1\u5ea6\u5354\u4f5c\u3002llm-d&nbsp;\u793e\u7fa4\u4ea6\u7372\u5f97\u91cd\u8981\u5b78\u8853\u754c\u5275\u59cb\u652f\u6301\u8005\u7684\u97ff\u61c9\uff0c\u5305\u62ec\u52a0\u5dde\u5927\u5b78\u7684&nbsp;Sky Computing Lab\uff08vLLM&nbsp;\u7684\u5275\u59cb\u55ae\u4f4d\uff09\uff0c\u4ee5\u53ca\u829d\u52a0\u54e5\u5927\u5b78\u7684&nbsp;LMCache Lab\uff08<a href=\"https:\/\/github.com\/LMCache\/LMCache\" target=\"_blank\" rel=\"noreferrer noopener\">LMCache<\/a>&nbsp;\u7684\u5275\u59cb\u55ae\u4f4d\uff09\u3002<\/p>\n\n\n\n<p>Red Hat&nbsp;\u79c9\u6301\u5c0d\u958b\u653e\u5354\u4f5c\u7684\u627f\u8afe\uff0c\u6df1\u77e5\u5728\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u63a8\u8ad6\u5feb\u901f\u6f14\u9032\u4e0b\uff0c\u5145\u6eff\u6d3b\u529b\u4e14\u6613\u65bc\u53c3\u8207\u7684\u793e\u7fa4\u70ba\u6838\u5fc3\u95dc\u9375\u3002Red Hat&nbsp;\u5c07\u7a4d\u6975\u8015\u8018&nbsp;llm-d&nbsp;\u793e\u7fa4\u4ee5\u63a8\u52d5\u5176\u6210\u9577\uff0c\u70ba\u65b0\u6210\u54e1\u71df\u9020\u5305\u5bb9\u7684\u74b0\u5883\uff0c\u4e26\u4fc3\u9032\u5176\u6301\u7e8c\u767c\u5c55\u3002<\/p>\n\n\n\n<p><a><\/a>Red Hat&nbsp;\u7684\u9858\u666f\uff1a<strong>\u4efb\u4f55\u6a21\u578b\u3001\u4efb\u4f55\u52a0\u901f\u5668\u3001\u4efb\u4f55\u96f2\u7aef\u3002<\/strong><strong><\/strong><\/p>\n\n\n\n<p>AI&nbsp;\u7684\u672a\u4f86\u61c9\u7531\u7121\u9650\u6a5f\u6703\u6240\u5b9a\u7fa9\uff0c\u800c\u975e\u53d7\u9650\u65bc\u57fa\u790e\u67b6\u69cb\u5b64\u5cf6\u3002Red Hat&nbsp;\u5c0d\u672a\u4f86\u7684\u5c55\u671b\u662f\uff0c\u4f01\u696d\u80fd\u65bc\u4efb\u4f55\u96f2\u7aef\u3001\u4efb\u4f55\u52a0\u901f\u5668\u4e0a\u90e8\u7f72\u4efb\u4f55\u6a21\u578b\uff0c\u4e26\u5728\u5408\u7406\u6210\u672c\u7684\u524d\u63d0\u4e0b\u63d0\u4f9b\u5353\u8d8a\u4e14\u66f4\u4e00\u81f4\u7684\u4f7f\u7528\u8005\u9ad4\u9a57\u3002\u70ba\u5145\u5206\u91cb\u653e\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u6295\u8cc7\u7684\u771f\u6b63\u6f5b\u529b\uff0c\u4f01\u696d\u9700\u8981\u901a\u7528\u7684\u63a8\u8ad6\u5e73\u53f0\uff0c\u9019\u5c07\u6210\u70ba\u73fe\u4eca\u53ca\u672a\u4f86\u5be6\u73fe\u66f4\u7121\u7e2b\u3001\u9ad8\u6548\u80fd&nbsp;AI&nbsp;\u5275\u65b0\u7684\u6a19\u6e96\u3002<\/p>\n\n\n\n<p>\u6b63\u5982&nbsp;Red Hat&nbsp;\u66fe\u5c07&nbsp;Linux&nbsp;\u8f49\u578b\u70ba\u73fe\u4ee3&nbsp;IT&nbsp;\u7684\u57fa\u77f3\uff0c\u958b\u5275\u4e86\u958b\u6e90\u4f01\u696d\u7684\u5148\u6cb3\uff0c\u5982\u4eca&nbsp;Red Hat&nbsp;\u4e5f\u5c07\u64d8\u5283&nbsp;AI&nbsp;\u63a8\u8ad6\u7684\u672a\u4f86\u3002vLLM&nbsp;\u64c1\u6709\u6210\u70ba\u6a19\u6e96\u5316\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u63a8\u8ad6\u95dc\u9375\u6838\u5fc3\u7684\u6f5b\u529b\uff0cRed Hat&nbsp;\u4e0d\u50c5\u6301\u7e8c\u81f4\u529b\u58ef\u5927&nbsp;vLLM&nbsp;\u793e\u7fa4\uff0c\u4e5f\u540c\u6642\u63a8\u52d5\u5927\u898f\u6a21\u5206\u6563\u5f0f\u63a8\u8ad6\u7684&nbsp;<a href=\"https:\/\/www.redhat.com\/en\/about\/press-releases\/red-hat-launches-llm-d-community-powering-distributed-gen-ai-inference-scale\" target=\"_blank\" rel=\"noreferrer noopener\">llm-d<\/a>\uff0c\u5171\u540c\u6253\u9020\u84ec\u52c3\u767c\u5c55\u7684\u751f\u614b\u7cfb\u3002Red Hat&nbsp;\u7684\u9858\u666f\u975e\u5e38\u660e\u78ba\uff0c\u4ea6\u5373\u7121\u8ad6&nbsp;AI&nbsp;\u6a21\u578b\u3001\u57fa\u5c64\u52a0\u901f\u5668\u6216\u90e8\u7f72\u74b0\u5883\u70ba\u4f55\uff0cvLLM&nbsp;\u7686\u80fd\u6210\u70ba\u65b0\u578b\u614b\u8de8\u6df7\u5408\u96f2\u63a8\u8ad6\u7684\u958b\u6e90\u6a19\u6e96\u6b0a\u5a01\u3002<\/p>\n\n\n\n<p>AMD AI&nbsp;\u7522\u54c1\u7ba1\u7406\u4e8b\u696d\u90e8\u526f\u7e3d\u88c1&nbsp;Ramine Roane&nbsp;\u8868\u793a\uff1a\u300cAMD&nbsp;\u5f88\u81ea\u8c6a\u80fd\u6210\u70ba&nbsp;llm-d&nbsp;\u793e\u7fa4\u7684\u5275\u59cb\u6210\u54e1\uff0c\u8ca2\u737b\u6211\u5011\u5728\u9ad8\u6548\u80fd&nbsp;GPU&nbsp;\u9818\u57df\u7684\u5c08\u696d\uff0c\u63a8\u52d5&nbsp;AI&nbsp;\u63a8\u8ad6\u6280\u8853\u7684\u9032\u6b65\u4e26\u6eff\u8db3\u4e0d\u65b7\u6f14\u9032\u7684\u4f01\u696d&nbsp;AI&nbsp;\u9700\u6c42\u3002\u96a8\u8457\u4f01\u696d\u8ffd\u6c42\u66f4\u5927\u898f\u6a21\u548c\u66f4\u9ad8\u6548\u7387\u7684\u751f\u6210\u5f0f&nbsp;AI\uff0cAMD&nbsp;\u671f\u5f85\u900f\u904e&nbsp;llm-d&nbsp;\u5c08\u6848\u4f86\u6eff\u8db3\u696d\u754c\u65e5\u76ca\u8907\u96dc\u7684\u9700\u6c42\u3002\u300d<em><\/em><\/p>\n\n\n\n<p>\u601d\u79d1\u958b\u6e90\u8a08\u756b\u8fa6\u516c\u5ba4\u526f\u7e3d\u88c1\u66a8&nbsp;Cisco DevNet&nbsp;\u8ca0\u8cac\u4eba&nbsp;Shannon McFarland&nbsp;\u8868\u793a\uff1a\u300cllm-d&nbsp;\u5c08\u6848\u662f\u5be6\u7528\u578b\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u5411\u524d\u9081\u51fa\u4ee4\u4eba\u632f\u596e\u7684\u4e00\u6b65\u3002\u958b\u767c\u4eba\u54e1\u80fd\u904b\u7528&nbsp;llm-d&nbsp;\u4ee5\u7a0b\u5f0f\u5316\u7684\u65b9\u5f0f\u6574\u5408\u548c\u64f4\u5c55\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u63a8\u8ad6\uff0c\u5728\u73fe\u4ee3&nbsp;AI&nbsp;\u9818\u57df\u4e2d\u91cb\u653e\u66f4\u9ad8\u5c64\u6b21\u7684\u5275\u65b0\u548c\u6548\u7387\u3002\u601d\u79d1\u5f88\u81ea\u8c6a\u80fd\u6210\u70ba&nbsp;llm-d&nbsp;\u793e\u7fa4\u7684\u4e00\u54e1\uff0c\u6211\u5011\u6b63\u651c\u624b\u63a2\u7d22\u5be6\u969b\u61c9\u7528\u60c5\u5883\uff0c\u5354\u52a9\u4f01\u696d\u4ee5\u66f4\u6709\u6548\u4e14\u66f4\u5177\u6548\u7387\u7684\u65b9\u5f0f\u61c9\u7528&nbsp;AI\u3002\u300d<em><\/em><\/p>\n\n\n\n<p>CoreWeave&nbsp;\u5de5\u7a0b\u8cc7\u6df1\u526f\u7e3d\u88c1&nbsp;Chen Goldberg&nbsp;\u8868\u793a\uff1a\u300cCoreWeave&nbsp;\u5f88\u9ad8\u8208\u80fd\u6210\u70ba&nbsp;llm-d&nbsp;\u5c08\u6848\u7684\u5275\u59cb\u8ca2\u737b\u8005\uff0c\u4e26\u6df1\u5316\u6211\u5011\u5c0d\u958b\u6e90&nbsp;AI&nbsp;\u7684\u9577\u671f\u627f\u8afe\u3002\u5f9e\u65e9\u671f\u8207&nbsp;EleutherAI&nbsp;\u5408\u4f5c\uff0c\u6211\u5011\u81f3\u4eca\u4ecd\u6301\u7e8c\u63a8\u52d5\u5927\u898f\u6a21\u63a8\u8ad6\u7684\u5de5\u4f5c\uff0c\u6211\u5011\u4e00\u76f4\u81f4\u529b\u65bc\u8b93\u53d6\u5f97\u5f37\u5927\u7684&nbsp;AI&nbsp;\u57fa\u790e\u67b6\u69cb\u8b8a\u5f97\u66f4\u52a0\u8f15\u9b06\u3002\u6211\u5011\u5f88\u9ad8\u8208\u80fd\u8207\u773e\u591a\u512a\u79c0\u7684\u5925\u4f34\u53ca\u66f4\u5ee3\u6cdb\u7684\u958b\u767c\u793e\u7fa4\u5408\u4f5c\uff0c\u651c\u624b\u6253\u9020\u5f48\u6027\u3001\u9ad8\u6548\u80fd\u7684\u63a8\u8ad6\u5f15\u64ce\uff0c\u70ba\u958b\u653e\u4e14\u53ef\u4e92\u901a\u7684&nbsp;AI&nbsp;\u5960\u5b9a\u57fa\u790e\uff0c\u52a0\u901f\u5275\u65b0\u3002\u300d<\/p>\n\n\n\n<p>Google Cloud AI&nbsp;\u8207\u904b\u7b97\u57fa\u790e\u67b6\u69cb\u526f\u7e3d\u88c1\u66a8\u7e3d\u7d93\u7406&nbsp;Mark Lohmeyer&nbsp;\u8868\u793a\uff1a\u300c\u96a8\u8457\u4f01\u696d\u958b\u59cb\u5927\u898f\u6a21\u90e8\u7f72&nbsp;AI&nbsp;\u4e26\u70ba\u5176\u4f7f\u7528\u8005\u5275\u9020\u50f9\u503c\uff0c\u9ad8\u6548\u7387\u7684&nbsp;AI&nbsp;\u63a8\u8ad6\u8b8a\u5f97\u6975\u70ba\u91cd\u8981\u3002\u96a8\u8457\u6211\u5011\u6b65\u5165\u63a8\u8ad6\u7684\u65b0\u6642\u4ee3\uff0cGoogle Cloud&nbsp;\u5f88\u69ae\u5e78\u80fd\u4f5c\u70ba&nbsp;llm-d&nbsp;\u5c08\u6848\u7684\u5275\u59cb\u8ca2\u737b\u8005\uff0c\u5728\u6211\u5011\u65e2\u6709\u7684\u958b\u6e90\u8ca2\u737b\u57fa\u790e\u4e0a\u5ef6\u4f38\u6295\u5165\u3002\u9019\u500b\u65b0\u793e\u7fa4\u5c07\u6210\u70ba\u5927\u898f\u6a21\u5206\u6563\u5f0f&nbsp;AI&nbsp;\u63a8\u8ad6\u7684\u95dc\u9375\u50ac\u5316\u5291\uff0c\u5354\u52a9\u4f7f\u7528\u8005\u63d0\u9ad8\u5de5\u4f5c\u8ca0\u8f09\u6548\u7387\uff0c\u4e26\u70ba\u5176\u57fa\u790e\u67b6\u69cb\u8cc7\u6e90\u63d0\u4f9b\u66f4\u591a\u9078\u64c7\u3002\u300d<\/p>\n\n\n\n<p>Hugging Face&nbsp;\u7522\u54c1\u8ca0\u8cac\u4eba&nbsp;Jeff Boudier&nbsp;\u8868\u793a\uff1a\u300c\u6211\u5011\u76f8\u4fe1\u6bcf\u5bb6\u516c\u53f8\u90fd\u61c9\u8a72\u80fd\u5920\u5efa\u7acb\u548c\u904b\u884c\u81ea\u5bb6\u7684\u6a21\u578b\u3002\u85c9\u7531&nbsp;vLLM&nbsp;\u5229\u7528&nbsp;Hugging Face transformers&nbsp;\u51fd\u5f0f\u5eab\u4f5c\u70ba\u6a21\u578b\u5b9a\u7fa9\u7684\u55ae\u4e00\u4e8b\u5be6\u4f86\u6e90\uff0c\u4e0d\u540c\u5927\u5c0f\u898f\u6a21\u7684\u6a21\u578b\u90fd\u80fd\u652f\u63f4\u6587\u5b57\u3001\u97f3\u8a0a\u3001\u5716\u7247\u548c\u5f71\u50cf&nbsp;AI&nbsp;\u61c9\u7528\u7a0b\u5f0f\u3002800&nbsp;\u842c\u540d&nbsp;AI&nbsp;\u5efa\u7f6e\u8005\u900f\u904e&nbsp;Hugging Face&nbsp;\u5e73\u53f0\uff0c\u5728\u5168\u7403\u793e\u7fa4\u516c\u958b\u5206\u4eab\u7684\u903e&nbsp;200&nbsp;\u842c\u500b&nbsp;AI&nbsp;\u6a21\u578b\u548c\u8cc7\u6599\u96c6\u4e0a\u5354\u4f5c\u3002\u6211\u5011\u5f88\u9ad8\u8208\u80fd\u652f\u63f4&nbsp;llm-d&nbsp;\u5c08\u6848\uff0c\u8b93\u958b\u767c\u4eba\u54e1\u80fd\u5c07\u9019\u4e9b\u61c9\u7528\u7a0b\u5f0f\u64f4\u5c55\u5230\u66f4\u5927\u898f\u6a21\u3002\u300d<\/p>\n\n\n\n<p>IBM Research&nbsp;\u6df7\u5408\u96f2\u8207&nbsp;AI&nbsp;\u5e73\u53f0\u526f\u7e3d\u88c1&nbsp;Priya Nagpurkar&nbsp;\u8868\u793a\uff1a\u300cIBM&nbsp;\u8a8d\u70ba&nbsp;AI&nbsp;\u7684\u4e0b\u4e00\u968e\u6bb5\u91cd\u9ede\u5728\u65bc\u6548\u7387\u548c\u898f\u6a21\u3002\u6211\u5011\u5c08\u6ce8\u65bc\u900f\u904e\u4f01\u696d\u80fd\u6709\u6548\u90e8\u7f72\u7684&nbsp;AI&nbsp;\u89e3\u6c7a\u65b9\u6848\uff0c\u70ba\u4f01\u696d\u91cb\u653e\u50f9\u503c\u3002\u4f5c\u70ba&nbsp;llm-d&nbsp;\u7684\u5275\u59cb\u8ca2\u737b\u8005\uff0cIBM&nbsp;\u5f88\u81ea\u8c6a\u80fd\u6210\u70ba\u5efa\u7f6e\u5dee\u7570\u5316\u3001\u786c\u9ad4\u4e2d\u7acb\uff08hardware agnostic\uff09\u7684\u5206\u6563\u5f0f&nbsp;AI&nbsp;\u63a8\u8ad6\u5e73\u53f0\u7684\u95dc\u9375\u4e00\u54e1\u3002\u6211\u5011\u671f\u5f85\u6301\u7e8c\u70ba\u9019\u500b\u793e\u7fa4\u7684\u6210\u9577\u548c\u6210\u529f\u505a\u51fa\u8ca2\u737b\uff0c\u4ee5\u6539\u8b8a&nbsp;AI&nbsp;\u63a8\u8ad6\u7684\u672a\u4f86\u3002\u300d<\/p>\n\n\n\n<p>Intel&nbsp;\u8cc7\u6599\u4e2d\u5fc3\u8207&nbsp;AI&nbsp;\u8edf\u9ad4\u89e3\u6c7a\u65b9\u6848\u66a8\u751f\u614b\u7cfb\u526f\u7e3d\u88c1&nbsp;Bill Pearson&nbsp;\u8868\u793a\uff1a\u300cllm-d&nbsp;\u7684\u63a8\u51fa\u5c07\u6210\u70ba\u696d\u754c\u63a8\u52d5\u5927\u898f\u6a21&nbsp;AI&nbsp;\u8f49\u578b\u7684\u95dc\u9375\u8f49\u6369\u9ede\uff0cIntel&nbsp;\u5f88\u69ae\u5e78\u80fd\u4ee5\u5275\u59cb\u652f\u6301\u8005\u8eab\u5206\u53c3\u8207\u5176\u4e2d\u3002Intel&nbsp;\u53c3\u8207&nbsp;llm-d&nbsp;\u662f\u6211\u5011\u8207&nbsp;Red Hat&nbsp;\u9577\u9054\u6578\u5341\u5e74\u5408\u4f5c\u7684\u6700\u65b0\u91cc\u7a0b\u7891\uff0c\u65e8\u5728\u900f\u904e\u958b\u6e90\u89e3\u6c7a\u65b9\u6848\u8ce6\u80fd\u4f01\u696d\u5f97\u4ee5\u5728\u4efb\u4f55\u5730\u9ede\u3001\u4efb\u4f55\u6240\u9078\u5e73\u53f0\u4e0a\u9032\u884c\u90e8\u7f72\u3002\u6211\u5011\u671f\u5f85\u900f\u904e&nbsp;llm-d&nbsp;\u793e\u7fa4\u9032\u4e00\u6b65\u64f4\u5c55\u548c\u5efa\u69cb&nbsp;AI&nbsp;\u5275\u65b0\u3002\u300d<\/p>\n\n\n\n<p>NVIDIA&nbsp;\u5de5\u7a0b&nbsp;AI&nbsp;\u6846\u67b6\u526f\u7e3d\u88c1&nbsp;Ujval Kapasi&nbsp;\u8868\u793a\uff1a\u300cllm-d&nbsp;\u5c08\u6848\u662f\u958b\u6e90&nbsp;AI&nbsp;\u751f\u614b\u7cfb\u4e2d\u7684\u91cd\u8981\u4e00\u74b0\uff0c\u540c\u6642\u5c55\u73fe&nbsp;NVIDIA&nbsp;\u652f\u6301\u4ee5\u5354\u4f5c\u63a8\u52d5\u751f\u6210\u5f0f&nbsp;AI&nbsp;\u5275\u65b0\u7684\u7acb\u5834\u3002\u53ef\u64f4\u5c55\u4e14\u9ad8\u6548\u80fd\u7684\u63a8\u8ad6\u662f\u4e0b\u4e00\u6ce2\u751f\u6210\u5f0f\u548c\u4ee3\u7406&nbsp;AI&nbsp;\u7684\u95dc\u9375\u3002\u6211\u5011\u6b63\u8207&nbsp;Red Hat&nbsp;\u53ca\u5176\u4ed6\u652f\u6301\u6b64\u5c08\u6848\u7684\u5925\u4f34\u5408\u4f5c\uff0c\u4fc3\u9032&nbsp;llm-d&nbsp;\u793e\u7fa4\u7684\u53c3\u8207\u548c\u63a1\u7528\uff0c\u4e26\u900f\u904e\u50cf\u662f&nbsp;NIXL&nbsp;\u9019\u6a23\u7684&nbsp;NVIDIA Dynamo&nbsp;\u5275\u65b0\u6280\u8853\uff0c\u5354\u52a9\u52a0\u5feb&nbsp;llm-d&nbsp;\u7684\u767c\u5c55\u3002\u300d<\/p>\n\n\n\n<p><\/p>\n\n\n\n<p><\/p>\n\n\n\n<p>\u3000\u3000\u3000<\/p>\n\n\n\n<h4 class=\"wp-block-heading has-text-align-right has-very-light-gray-to-cyan-bluish-gray-gradient-background has-background\">\ud83d\udfe6COMPUTEX \u6d3b\u52d5 \u2013 <a href=\"https:\/\/reurl.cc\/vQbdRN\" data-type=\"link\" data-id=\"https:\/\/reurl.cc\/lzOKNA\" target=\"_blank\" rel=\"noreferrer noopener\">\u770b\u5341\u9293\u5c55\u5834\u4ecb\u7d39\u9001\u76ae\u9769\u5361\u5957\u5305<\/a><br>\ud83d\udfe6COMPUTEX \u6d3b\u52d5 \u2013 <a href=\"http:\/\/reurl.cc\/lzOKNA\" target=\"_blank\" rel=\"noreferrer noopener\">\u770b ZOTAC \u5c55\u5834\u65b0\u54c1\u62ff\u9650\u91cf\u5c0f\u591c\u71c8<\/a><br>\ud83d\udfe6COMPUTEX \u6d3b\u52d5&#8211; <a href=\"https:\/\/reurl.cc\/OYnMNr\" target=\"_blank\" rel=\"noreferrer noopener\">\u770b MSI \u5c55\u524d\u6703\u62ff\u8d85\u840c\u597d\u79ae<\/a><br>\ud83d\udfe6\u641e\u754f\u6709\u79ae\uff01\u53ea\u8981\u65bc\u7ad9\u4e0a <a href=\"https:\/\/www.facebook.com\/ioioTIMES\/\" data-type=\"link\" data-id=\"https:\/\/www.facebook.com\/ioioTIMES\/\" target=\"_blank\" rel=\"noreferrer noopener\">FB \u4efb\u4e00\u5e16\u767c\u8a00<\/a>\u5c31\u6709\u6a5f\u6703\u7372\u5f97\u300c\u641e\u754f\u734e\u795e\u79d8\u79ae\u54c1\u300d<br>\ud83d\udfe6\u73fe\u5728\u5c31\u52a0\u5165&nbsp;ioioTIMES \u81c9\u66f8\u7c89\u7d72\u5718&nbsp;\u66f4\u591a\u4e92\u52d5\u3001\u66f4\u591a\u597d\u5eb7\u650f\u62b5\u52a0!!<br>\ud83d\udfe6<strong>\u6211\u5011\u6709<a href=\"https:\/\/today.line.me\/tw\/v2\/publisher\/103117\" target=\"_blank\" rel=\"noreferrer noopener nofollow\">LINE TODAY<\/a>\u983b\u9053\u4e86\uff0c\u5feb\u4f86\u8ffd\u8e2a\u6211\u5011\u5427!!&#8211;\u6700\u65b0\u79d1\u6280\u65b0\u805e \u76e1\u5728\u4f60<\/strong>\u624b<\/h4>\n\n\n\n<p><\/p>\n","protected":false},"excerpt":{"rendered":"<p>\u4e16\u754c\u9818\u5148\u958b\u653e\u539f\u59cb\u78bc\u8edf\u9ad4\u89e3\u6c7a\u65b9\u6848\u4f9b\u61c9\u5546<\/p>\n","protected":false},"author":3,"featured_media":110669,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"rank_math_lock_modified_date":false,"footnotes":""},"categories":[13],"tags":[13736,1586,13737],"class_list":["post-110668","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news","tag-llm-d","tag-red-hat","tag-ai"],"_links":{"self":[{"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/posts\/110668"}],"collection":[{"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=110668"}],"version-history":[{"count":2,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/posts\/110668\/revisions"}],"predecessor-version":[{"id":110681,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/posts\/110668\/revisions\/110681"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=\/wp\/v2\/media\/110669"}],"wp:attachment":[{"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=110668"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=110668"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.ioiotimes.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=110668"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}