{"id":108626,"date":"2024-03-21T03:54:00","date_gmt":"2024-03-20T20:54:00","guid":{"rendered":"https:\/\/www.xinwengao.com\/cn\/pr\/20240321035400108626"},"modified":"2024-03-22T02:21:47","modified_gmt":"2024-03-21T18:21:47","slug":"supermicro-%e9%87%87%e7%94%a8%e6%96%b0%e4%b8%80%e4%bb%a3%e7%b3%bb%e7%bb%9f%e5%92%8c%e6%9c%ba%e6%9e%b6%e6%9e%b6%e6%9e%84%e4%bb%a5%e6%89%a9%e5%a4%a7%e4%ba%ba%e5%b7%a5%e6%99%ba%e8%83%bd%e4%bc%98%e5%8c%96","status":"publish","type":"post","link":"https:\/\/www.xinwengao.com\/cn\/pr\/20240321035400108626","title":{"rendered":"Supermicro \u91c7\u7528\u65b0\u4e00\u4ee3\u7cfb\u7edf\u548c\u673a\u67b6\u67b6\u6784\u4ee5\u6269\u5927\u4eba\u5de5\u667a\u80fd\u4f18\u5316\u4ea7\u54c1\u7ec4\u5408"},"content":{"rendered":"\n\n\n
\"\"<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

– Supermicro <\/b>\u91c7\u7528\u7531\u5168\u65b0<\/b> NVIDIA Blackwell <\/b>\u67b6\u6784\u89e3\u51b3\u65b9\u6848\u7ec4\u6210\u7684\u65b0\u4e00\u4ee3\u7cfb\u7edf\u548c\u673a\u67b6\u67b6\u6784\u4ee5\u6269\u5927\u4eba\u5de5\u667a\u80fd\u4f18\u5316\u4ea7\u54c1\u7ec4\u5408<\/b><\/p>\n

\u9002\u7528\u4e8e\u5927\u89c4\u6a21<\/i> CSP <\/i>\u548c<\/i> NSP <\/i>\u7684\u5f3a\u5927\u8282\u80fd\u89e3\u51b3\u65b9\u6848\u5c06\u7ed3\u5408\u5168\u65b9\u4f4d\u4e0b\u4e00\u4ee3<\/i> NVIDIA GPU <\/i>\u548c<\/i> CPU<\/i>\uff0c\u4ee5\u53ca<\/i> NVIDIA Quantum X800 <\/i>\u5e73\u53f0\u548c<\/i> NVIDIA AI Enterprise 5.0<\/i><\/p>\n

\u52a0\u5229\u798f\u5c3c\u4e9a\u5dde\u5723\u4f55\u585e<\/span>2024\u5e743\u670821\u65e5<\/span> \/\u7f8e\u901a\u793e\/ — \u5728 2024 <\/b>\u5e74<\/b> NVIDIA GPU <\/b>\u5927\u4f1a\u4e0a\uff0c\u4eba\u5de5\u667a\u80fd\u3001\u4e91\u3001\u5b58\u50a8\u548c<\/b> 5G\/<\/b>\u8fb9\u7f18\u7684\u6574\u4f53<\/b> IT <\/b>\u89e3\u51b3\u65b9\u6848\u63d0\u4f9b\u5546<\/b> Supermicro, Inc.<\/b>\uff08\u7eb3\u65af\u8fbe\u514b\u80a1\u7968\u4ee3\u7801\uff1a<\/b>SMCI<\/b>\uff09<\/b>\u5ba3\u5e03\u63a8\u51fa\u65b0\u7684\u5927\u89c4\u6a21\u751f\u6210\u5f0f\u4eba\u5de5\u667a\u80fd\u7cfb\u7edf\uff0c\u8be5\u7cfb\u7edf\u91c7\u7528 NVIDIA \u7684\u4e0b\u4e00\u4ee3\u6570\u636e\u4e2d\u5fc3\u4ea7\u54c1\uff0c\u5305\u62ec\u6700\u65b0\u7684 NVIDIA GB200 Grace™ Blackwell Superchip\u3001NVIDIA B200 Tensor Core \u548c B100 Tensor Core GPU\u3002Supermicro \u6b63\u5728\u4f18\u5316\u5176\u76ee\u524d\u7684 NVIDIA HGX™ H100\/H200 8-GPU \u7cfb\u7edf\uff0c\u4ee5\u4fbf\u652f\u6301 NVIDIA HGX™ B100 8-GPU \u548c B200\uff0c\u4ece\u800c\u7f29\u77ed\u4ea4\u4ed8\u65f6\u95f4\u3002\u6b64\u5916\uff0cSupermicro \u8fd8\u5c06\u901a\u8fc7 NVIDIA GB200\uff08\u5305\u62ec\u5177\u6709 72 \u4e2aNVIDIA Blackwell GPU \u7684\u5b8c\u6574\u673a\u67b6\u7ea7\u89e3\u51b3\u65b9\u6848\u7684 NVIDIA GB200 NVL72\uff09\u65b0\u4ea7\u54c1\uff0c\u8fdb\u4e00\u6b65\u52a0\u5f3a\u5176\u5e9e\u5927\u7684 NVIDIA MGX™ \u7cfb\u7edf\u4ea7\u54c1\u7cfb\u5217\u3002Supermicro \u8fd8\u5728\u5176\u4ea7\u54c1\u9635\u5bb9\u4e2d\u589e\u52a0\u4e86\u65b0\u7cfb\u7edf\uff0c\u5305\u62ec 4U NVIDIA HGX B200 8-GPU \u6db2\u51b7\u7cfb\u7edf\u3002<\/p>\n

\"\"<\/a>
<\/span> <\/div>\n

Supermicro \u603b\u88c1\u517c\u9996\u5e2d\u6267\u884c\u5b98\u6881\u89c1\u540e\u8868\u793a\uff1a"\u6211\u4eec\u4e13\u6ce8\u4e8e\u4eba\u5de5\u667a\u80fd\u7684\u6784\u5efa\u5757\u67b6\u6784\u548c\u673a\u67b6\u7ea7\u6574\u4f53 IT\uff0c\u6709\u52a9\u4e8e\u6211\u4eec\u8bbe\u8ba1\u4e0b\u4e00\u4ee3\u7cfb\u7edf\uff0c\u4ee5\u6ee1\u8db3\u57fa\u4e8e NVIDIA Blackwell \u67b6\u6784 GPU \u7684\u66f4\u9ad8\u8981\u6c42\uff0c\u4f8b\u5982\u91c7\u7528 NVIDIA HGX B200 8-GPU \u7684\u65b0 4U \u6db2\u51b7\u7cfb\u7edf\u4ee5\u53ca NVIDIA GB200 NVL72 \u7684\u5b8c\u5168\u96c6\u6210\u76f4\u8fbe\u82af\u7247\u6db2\u51b7\u673a\u67b6\u3002\u8fd9\u4e9b\u65b0\u4ea7\u54c1\u7684\u5f00\u53d1\u4e43\u57fa\u4e8e Supermicro \u548c NVIDIA \u4e45\u7ecf\u8003\u9a8c\u7684 HGX \u548c MGX \u7cfb\u7edf\u67b6\u6784\uff0c\u5e76\u9488\u5bf9 NVIDIA Blackwell GPU \u7684\u65b0\u529f\u80fd\u8fdb\u884c\u4e86\u4f18\u5316\u3002Supermicro \u62e5\u6709\u5c06 1kW GPU \u6574\u5408\u5230\u5404\u79cd\u98ce\u51b7\u548c\u6db2\u51b7\u7cfb\u7edf\u4e2d\u7684\u4e13\u4e1a\u77e5\u8bc6\uff0c\u4ee5\u53ca\u6bcf\u6708 5,000 \u4e2a\u673a\u67b6\u7684\u673a\u67b6\u89c4\u6a21\u751f\u4ea7\u80fd\u529b\uff0c\u6211\u4eec\u9884\u8ba1\u5c06\u7387\u5148\u90e8\u7f72\u91c7\u7528 NVIDIA Blackwell GPU \u7684\u5168\u673a\u67b6\u96c6\u7fa4\uff0c\u5e76\u5728\u5e02\u573a\u4e0a\u63a8\u51fa\u3002"<\/p>\n

Supermicro \u76f4\u8fbe\u82af\u7247\u7684\u6db2\u4f53\u51b7\u5374\u6280\u672f\u5c06\u5141\u8bb8\u589e\u52a0\u6700\u65b0 GPU \u7684\u6563\u70ed\u8bbe\u8ba1\u529f\u7387 (TDP)\uff0c\u5e76\u5145\u5206\u53d1\u6325 NVIDIA Blackwell GPU \u7684\u6f5c\u529b\u3002Supermicro \u642d\u8f7d NVIDIA Blackwell \u7684 HGX \u548c MGX \u7cfb\u7edf\u662f\u672a\u6765\u4eba\u5de5\u667a\u80fd\u57fa\u7840\u8bbe\u65bd\u7684\u6784\u5efa\u5757\uff0c\u5c06\u4e3a\u6570\u4e07\u4ebf\u53c2\u6570\u7684\u4eba\u5de5\u667a\u80fd\u8bad\u7ec3\u548c\u5b9e\u65f6\u4eba\u5de5\u667a\u80fd\u63a8\u7406\u5e26\u6765\u7a81\u7834\u6027\u7684\u6027\u80fd\u3002<\/p>\n

\u4e00\u7cfb\u5217\u7ecf\u8fc7 GPU \u4f18\u5316\u7684 Supermicro \u7cfb\u7edf\u5c06\u4e3a NVIDIA Blackwell B200 \u548c B100 Tensor Core GPU \u505a\u597d\u51c6\u5907\uff0c\u5e76\u9488\u5bf9\u6700\u65b0\u7684 NVIDIA AI Enterprise \u8f6f\u4ef6\u8fdb\u884c\u9a8c\u8bc1\uff0c\u8be5\u8f6f\u4ef6\u589e\u52a0\u4e86\u5bf9 NVIDIA NIM<\/a> \u63a8\u7406\u5fae\u670d\u52a1\u7684\u652f\u6301\u3002Supermicro \u7cfb\u7edf\u5305\u62ec\uff1a<\/p>\n