{"id":1319,"date":"2025-07-30T13:02:01","date_gmt":"2025-07-30T13:02:01","guid":{"rendered":"https:\/\/flyingtiger.io\/?p=1319"},"modified":"2025-11-14T07:49:38","modified_gmt":"2025-11-14T07:49:38","slug":"nvidia-%e8%88%87%e5%85%a8%e7%90%83%e9%9b%bb%e8%85%a6%e8%a3%bd%e9%80%a0%e5%95%86%e5%85%b1%e5%90%8c%e6%8e%a8%e5%87%ba-ai-%e9%a6%96%e6%ac%be-dgx-%e5%80%8b%e4%ba%ba%e9%81%8b%e7%ae%97%e7%b3%bb%e7%b5%b1","status":"publish","type":"post","link":"https:\/\/flyingtiger.io\/zh\/nvidia-%e8%88%87%e5%85%a8%e7%90%83%e9%9b%bb%e8%85%a6%e8%a3%bd%e9%80%a0%e5%95%86%e5%85%b1%e5%90%8c%e6%8e%a8%e5%87%ba-ai-%e9%a6%96%e6%ac%be-dgx-%e5%80%8b%e4%ba%ba%e9%81%8b%e7%ae%97%e7%b3%bb%e7%b5%b1\/","title":{"rendered":"NVIDIA \u651c\u624b\u5168\u7403\u96fb\u8166\u88fd\u9020\u5546\u63a8\u51fa AI-First DGX \u500b\u4eba\u904b\u7b97\u7cfb\u7d71"},"content":{"rendered":"

\u5b8f\u7881\u3001\u83ef\u78a9\u3001\u6234\u723e\u79d1\u6280\u3001\u6280\u5609\u3001\u60e0\u666e\u3001\u806f\u60f3\u548c\u5fae\u661f\u70ba\u4eba\u5de5\u667a\u6167\u958b\u767c\u8005\u3001\u7814\u7a76\u4eba\u54e1\u548c\u8cc7\u6599\u79d1\u5b78\u5bb6\u6253\u9020 NVIDIA DGX Spark\u3001DGX Station \u500b\u4eba\u7cfb\u7d71

<\/p>\n\n

\u53f0\u5317\u570b\u969b\u96fb\u8166\u5c55\u2014<\/strong>NVIDIA \u4eca\u5929\u5ba3\u5e03\uff0c\u53f0\u7063\u9818\u5148\u7684\u7cfb\u7d71\u88fd\u9020\u5546\u5c07\u6253\u9020\u00a0NVIDIA DGX Spark<\/u><\/a>\u00a0\u548c\u00a0DGX \u7ad9<\/u><\/a>\u2122<\/u><\/a>\u00a0\u7cfb\u7d71\u3002<\/p>\n\n

\u8207\u5b8f\u7881\u3001\u6280\u5609\u548c\u5fae\u661f\u65e5\u76ca\u589e\u9577\u7684\u5408\u4f5c\u5925\u4f34\u95dc\u4fc2\u5c07\u64f4\u5c55 DGX Spark \u548c DGX Station \u500b\u4eba AI \u8d85\u7d1a\u96fb\u8166\u7684\u53ef\u7528\u6027\u2014\u2014\u70ba\u5168\u7403\u958b\u767c\u4eba\u54e1\u3001\u8cc7\u6599\u79d1\u5b78\u5bb6\u548c\u7814\u7a76\u4eba\u54e1\u751f\u614b\u7cfb\u7d71\u63d0\u4f9b\u524d\u6240\u672a\u6709\u7684\u6027\u80fd\u548c\u6548\u7387\u3002<\/p>\n\n

\u4f01\u696d\u3001\u8edf\u9ad4\u4f9b\u61c9\u5546\u3001\u653f\u5e9c\u6a5f\u69cb\u3001\u65b0\u5275\u516c\u53f8\u548c\u7814\u7a76\u6a5f\u69cb\u9700\u8981\u5f37\u5927\u7684\u7cfb\u7d71\uff0c\u8a72\u7cfb\u7d71\u53ef\u4ee5\u4ee5\u684c\u9762\u5f62\u5f0f\u63d0\u4f9b AI \u4f3a\u670d\u5668\u7684\u6548\u80fd\u548c\u529f\u80fd\uff0c\u540c\u6642\u53c8\u4e0d\u5f71\u97ff\u8cc7\u6599\u5927\u5c0f\u3001\u5c08\u6709\u6a21\u578b\u96b1\u79c1\u6216\u53ef\u64f4\u5c55\u901f\u5ea6\u3002<\/p>\n\n

\u5d1b\u8d77\u00a0\u4ee3\u7406\u4eba\u5de5\u667a\u6167<\/u><\/a>\u00a0\u80fd\u5920\u81ea\u4e3b\u6c7a\u7b56\u548c\u57f7\u884c\u4efb\u52d9\u7684\u7cfb\u7d71\u9032\u4e00\u6b65\u653e\u5927\u4e86\u9019\u4e9b\u9700\u6c42\u3002\u57fa\u65bc NVIDIA Grace Blackwell \u5e73\u53f0\u7684 DGX Spark \u548c DGX Station \u5c07\u4f7f\u958b\u767c\u4eba\u54e1\u80fd\u5920\u5f9e\u684c\u9762\u5230\u8cc7\u6599\u4e2d\u5fc3\u5efa\u7acb\u539f\u578b\u3001\u9032\u884c\u5fae\u8abf\u548c\u63a8\u7406\u6a21\u578b\u3002<\/p>\n\n

NVIDIA \u5275\u59cb\u4eba\u517c\u9996\u5e2d\u57f7\u884c\u5b98\u9ec3\u4ec1\u52f3\u8868\u793a\uff1a\u201cAI \u5fb9\u5e95\u6539\u8b8a\u4e86\u8a08\u7b97\u5806\u68e7\u7684\u6bcf\u4e00\u5c64\u2014\u2014\u5f9e\u77fd\u7247\u5230\u8edf\u9ad4\u3002DGX Spark \u548c DGX Station \u662f\u5f15\u767c AI \u9769\u547d\u7684 DGX-1 \u7cfb\u7d71\u7684\u76f4\u7cfb\u5f8c\u4ee3\uff0c\u5b83\u5011\u5f9e\u96f6\u958b\u59cb\u6253\u9020\uff0c\u65e8\u5728\u70ba\u4e0b\u4e00\u4ee3 AI \u7814\u767c\u63d0\u4f9b\u52d5\u529b\u3002\u201d<\/p>\n\n

DGX Spark \u8f14\u52a9\u5275\u65b0<\/strong>
DGX Spark \u642d\u8f09 NVIDIA GB10 Grace Blackwell \u8d85\u7d1a\u6676\u7247\u548c\u7b2c\u4e94\u4ee3 Tensor Core\uff0c\u63d0\u4f9b\u9ad8\u9054 1 petaflop \u7684 AI \u904b\u7b97\u80fd\u529b\u548c 128GB \u7684\u7d71\u4e00\u5167\u5b58\uff0c\u4e26\u652f\u63f4\u5c07\u6a21\u578b\u7121\u7e2b\u5c0e\u51fa\u81f3\u00a0NVIDIA DGX\u2122 \u96f2<\/u><\/a>\u00a0\u6216\u4efb\u4f55\u52a0\u901f\u96f2\u7aef\u6216\u8cc7\u6599\u4e2d\u5fc3\u57fa\u790e\u67b6\u69cb\u3002<\/p>\n\n

DGX Spark \u5728\u7dca\u6e4a\u7684\u5c01\u88dd\u4e2d\u63d0\u4f9b\u5f37\u5927\u7684\u6548\u80fd\u548c\u529f\u80fd\uff0c\u8b93\u958b\u767c\u4eba\u54e1\u3001\u7814\u7a76\u4eba\u54e1\u3001\u8cc7\u6599\u79d1\u5b78\u5bb6\u548c\u5b78\u751f\u80fd\u5920\u7a81\u7834\u00a0\u751f\u6210\u5f0f\u4eba\u5de5\u667a\u6167<\/u><\/a>\u00a0\u4e26\u52a0\u901f\u5404\u884c\u5404\u696d\u7684\u5de5\u4f5c\u91cf\u3002<\/p>\n\n

DGX Station \u63a8\u52d5 AI \u5275\u65b0<\/strong>
DGX Station \u5c08\u70ba\u6700\u56b4\u82db\u7684 AI \u5de5\u4f5c\u8ca0\u8f09\u800c\u6253\u9020\uff0c\u642d\u8f09 NVIDIA GB300 Grace Blackwell Ultra Desktop Superchip\uff0c\u53ef\u63d0\u4f9b\u9ad8\u9054 20 petaflops \u7684 AI \u6548\u80fd\u548c 784GB \u7684\u7d71\u4e00\u7cfb\u7d71\u8a18\u61b6\u9ad4\u3002\u8a72\u7cfb\u7d71\u9084\u5305\u542b\uff1a\u00a0NVIDIA ConnectX\u00ae-8 \u8d85\u7d1a\u7db2\u8def\u5361<\/u><\/a>\uff0c\u652f\u63f4\u9ad8\u9054800Gb\/s\u7684\u7db2\u8def\u901f\u5ea6\uff0c\u5be6\u73fe\u9ad8\u901f\u9023\u63a5\u548c\u591a\u7ad9\u64f4\u5c55\u3002<\/p>\n\n

DGX Station \u53ef\u4ee5\u4f5c\u70ba\u55ae\u4e00\u4f7f\u7528\u8005\u7684\u7368\u7acb\u684c\u9762\uff0c\u4f7f\u7528\u672c\u5730\u8cc7\u6599\u904b\u884c\u9ad8\u7d1a AI \u6a21\u578b\uff0c\u4e5f\u53ef\u4ee5\u4f5c\u70ba\u591a\u500b\u4f7f\u7528\u8005\u7684\u6309\u9700\u96c6\u4e2d\u5f0f\u8a08\u7b97\u7bc0\u9ede\u3002\u8a72\u7cfb\u7d71\u652f\u63f4\u00a0NVIDIA \u591a\u57f7\u884c\u500b\u9ad4 GPU<\/u><\/a>\u00a0\u6280\u8853\u5c07\u5176\u5283\u5206\u70ba\u591a\u9054\u4e03\u500b\u5be6\u4f8b\u2014\u2014\u6bcf\u500b\u5be6\u4f8b\u90fd\u6709\u81ea\u5df1\u7684\u9ad8\u983b\u5bec\u8a18\u61b6\u9ad4\u3001\u5feb\u53d6\u548c\u904b\u7b97\u6838\u5fc3\u2014\u2014\u4f5c\u70ba\u8cc7\u6599\u79d1\u5b78\u548c\u4eba\u5de5\u667a\u6167\u958b\u767c\u5718\u968a\u7684\u500b\u4eba\u96f2\u7aef\u3002<\/p>\n\n

\u70ba\u4e86\u63d0\u4f9b\u958b\u767c\u8005\u719f\u6089\u7684\u4f7f\u7528\u8005\u9ad4\u9a57\uff0cDGX Spark \u548c DGX Station \u6cbf\u8972\u4e86\u9a45\u52d5\u5de5\u696d\u7d1a AI \u5de5\u5ee0\u7684\u8edf\u9ad4\u67b6\u69cb\u3002\u5169\u6b3e\u7cfb\u7d71\u5747\u63a1\u7528 NVIDIA DGX \u4f5c\u696d\u7cfb\u7d71\uff0c\u9810\u5148\u914d\u7f6e\u4e86\u6700\u65b0\u7684 NVIDIA AI \u8edf\u9ad4\u5806\u758a\uff0c\u4e26\u5305\u542b\u4ee5\u4e0b\u5b58\u53d6\u6b0a\u9650\uff1a\u00a0NVIDIA NIM\u2122 \u5fae\u670d\u52d9<\/u><\/a>\u00a0\u548c\u00a0NVIDIA \u85cd\u5716<\/u><\/a>.<\/p>\n\n

\u958b\u767c\u4eba\u54e1\u53ef\u4ee5\u4f7f\u7528 PyTorch\u3001Jupyter \u548c Ollama \u7b49\u5e38\u7528\u5de5\u5177\u5728 DGX Spark \u4e0a\u9032\u884c\u539f\u578b\u8a2d\u8a08\u3001\u5fae\u8abf\u548c\u63a8\u7406\uff0c\u4e26\u7121\u7e2b\u90e8\u7f72\u5230 DGX Cloud \u6216\u4efb\u4f55\u52a0\u901f\u8cc7\u6599\u4e2d\u5fc3\u6216\u96f2\u7aef\u57fa\u790e\u67b6\u69cb\u3002<\/p>\n\n

\u6234\u723e\u79d1\u6280\u662f\u9996\u6279\u958b\u767c DGX Spark \u548c DGX Station \u7684\u5168\u7403\u7cfb\u7d71\u88fd\u9020\u5546\u4e4b\u4e00\uff0c\u81f4\u529b\u65bc\u6eff\u8db3\u4f01\u696d\u5c0d\u5f37\u5927\u7684\u672c\u5730\u5316 AI \u904b\u7b97\u89e3\u6c7a\u65b9\u6848\u65e5\u76ca\u589e\u9577\u7684\u9700\u6c42\u3002<\/p>\n\n

\u6234\u723e\u79d1\u6280\u96c6\u5718\u8463\u4e8b\u9577\u517c\u9996\u5e2d\u57f7\u884c\u5b98\u9081\u514b\u723e\u00b7\u6234\u723e\u8868\u793a\uff1a\u300c\u6d88\u8cbb\u8005\u548c\u4f01\u696d\u660e\u986f\u8f49\u5411\u512a\u5148\u8003\u616e\u80fd\u5920\u8655\u7406\u4e0b\u4e00\u4ee3\u667a\u6167\u5de5\u4f5c\u8ca0\u8f09\u7684\u7cfb\u7d71\u3002\u5c0d NVIDIA DGX Spark \u548c NVIDIA DGX Station \u7684\u95dc\u6ce8\u6a19\u8a8c\u8457\u684c\u9762\u8a08\u7b97\u65b0\u6642\u4ee3\u7684\u5230\u4f86\uff0c\u5c07\u91cb\u653e\u672c\u5730 AI \u6027\u80fd\u7684\u5168\u90e8\u6f5b\u529b\u3002\u6211\u5011\u7684\u7522\u54c1\u7d44\u5408\u65e8\u5728\u6eff\u8db3\u9019\u4e9b\u9700\u6c42\u3002\u70ba\u4f01\u696d\u63d0\u4f9b\u4e86\u6574\u5408\u548c\u8655\u7406\u5927\u578b AI \u5de5\u4f5c\u8ca0\u8f09\u7684\u57fa\u790e\u8a2d\u65bd\u3002<\/p>\n\n

\u60e0\u666e\u516c\u53f8\u900f\u904e\u63d0\u4f9b\u9019\u4e9b\u65b0\u89e3\u6c7a\u65b9\u6848\uff0c\u4f7f\u4f01\u696d\u80fd\u5920\u5145\u5206\u767c\u63ee\u4eba\u5de5\u667a\u6167\u6548\u80fd\u7684\u6f5b\u529b\uff0c\u5f9e\u800c\u589e\u5f37\u4eba\u5de5\u667a\u6167\u904b\u7b97\u7684\u672a\u4f86\u3002<\/p>\n\n

\u60e0\u666e\u516c\u53f8\u7e3d\u88c1\u517c\u57f7\u884c\u9577 Enrique Lores \u8868\u793a\uff1a\u300c\u900f\u904e\u8207 NVIDIA \u7684\u5408\u4f5c\uff0c\u6211\u5011\u5c07\u63d0\u4f9b\u4e00\u5957\u5168\u65b0\u7684 AI \u9a45\u52d5\u8a2d\u5099\u548c\u9ad4\u9a57\uff0c\u4ee5\u9032\u4e00\u6b65\u63a8\u9032\u60e0\u666e\u7684\u672a\u4f86\u5de5\u4f5c\u76ee\u6a19\uff0c\u5be6\u73fe\u696d\u52d9\u589e\u9577\u548c\u8077\u696d\u6210\u5c31\u3002\u501f\u52a9 HP ZGX\uff0c\u6211\u5011\u6b63\u5728\u91cd\u65b0\u5b9a\u7fa9\u684c\u9762\u8a08\u7b97\u2014\u2014\u70ba\u958b\u767c\u4eba\u54e1\u548c\u7814\u7a76\u4eba\u54e1\u5e36\u4f86\u6578\u64da\u4e2d\u5fc3\u7684\u8fed\u4ee3\u6027\u80fd\uff0c\u4ee5\u66f4\u5feb\u5730\u91cb\u653e\u7684\u8fed\u4ee3\u6027\u80fd\uff0c\u5f9e\u800c\u66f4\u5feb\u5730\u91cb\u653e\u6578\u64da\u4e2d\u5fc3\u3002

\u64f4\u5c55\u53ef\u7528\u6027\u548c\u5408\u4f5c\u5925\u4f34\u751f\u614b\u7cfb\u7d71<\/strong>
DGX Spark \u5c07\u65bc 7 \u6708\u8d77\u5728\u5b8f\u7881\u3001\u83ef\u78a9\u3001\u6234\u723e\u79d1\u6280\u3001\u6280\u5609\u3001\u60e0\u666e\u3001\u806f\u60f3\u3001\u5fae\u661f\u4ee5\u53ca\u5168\u7403\u901a\u8def\u5408\u4f5c\u5925\u4f34\u8655\u767c\u552e\u3002 DGX Spark \u73fe\u5df2\u958b\u653e\u9810\u8a02\u00a0nvidia.com<\/u><\/a>\u00a0\u4e26\u900f\u904e NVIDIA \u5408\u4f5c\u5925\u4f34\u3002<\/p>\n\n

\u9810\u8a08\u83ef\u78a9\u3001\u6234\u723e\u79d1\u6280\u3001\u6280\u5609\u3001\u60e0\u666e\u548c\u5fae\u661f\u5c07\u65bc\u4eca\u5e74\u7a0d\u5f8c\u63a8\u51fa DGX Station\u3002<\/p>\n\n

\u89c0\u770b\u00a0\u53f0\u5317\u570b\u969b\u96fb\u8166\u5c55<\/u>\u00a0<\/a>\u8046\u807d\u9ec3\u4ec1\u52f3\u7684\u4e3b\u984c\u6f14\u8b1b\u4e26\u5728 NVIDIA GTC \u53f0\u5317\u4e86\u89e3\u66f4\u591a\u8cc7\u8a0a\u3002<\/p>","protected":false},"excerpt":{"rendered":"

\u5b8f\u7881\u3001\u83ef\u78a9\u3001\u6234\u723e\u79d1\u6280\u3001\u6280\u5609\u3001\u60e0\u666e\u3001\u806f\u60f3\u548c\u5fae\u661f\u5c07\u70ba\u4eba\u5de5\u667a\u6167\u958b\u767c\u8005\u3001\u7814\u7a76\u4eba\u54e1\u548c\u8cc7\u6599\u79d1\u5b78\u5bb6\u6253\u9020 NVIDIA DGX Spark \u548c DGX Station \u500b\u4eba\u7cfb\u7d71\u2014\u2014\u5728\u53f0\u5317\u570b\u969b\u96fb\u8166\u5c55 (COMPUTEX) \u4e0a\uff0cNVIDIA \u4eca\u65e5\u5ba3\u5e03\uff0c\u53f0\u7063\u9818\u5148\u7684\u7cfb\u7d71\u88fd\u9020\u5546\u5c07\u6253\u9020 NVIDIA DGX \u548c XDGIA \u7cfb\u7d71\u3002\u8207\u5b8f\u7881\u3001\u6280\u5609\u548c\u5fae\u661f\u4e0d\u65b7\u6df1\u5316\u7684\u5408\u4f5c\u95dc\u4fc2\u5c07\u64f4\u5927 DGX Spark \u548c DGX Station \u7684\u4f9b\u61c9\u7bc4\u570d [\u2026]<\/p>","protected":false},"author":2,"featured_media":1320,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"two_page_speed":[],"footnotes":""},"categories":[1],"tags":[],"class_list":["post-1319","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/posts\/1319","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/comments?post=1319"}],"version-history":[{"count":0,"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/posts\/1319\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/media\/1320"}],"wp:attachment":[{"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/media?parent=1319"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/categories?post=1319"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/flyingtiger.io\/zh\/wp-json\/wp\/v2\/tags?post=1319"}],"curies":[{"name":"\u53ef\u6fd5\u6027\u7c89\u5291","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}