返回全部动态

NVIDIA 支持八节点 DGX Spark 集群,统一内存达 1TB

原标题:Running Frontier-Class AI Models Across 8 NVIDIA DGX Spark Nodes

NVIDIA Technical Blog一手来源产品发布质量 74

AI 摘要

NVIDIA 在 GTC Berlin 宣布正式支持八节点 DGX Spark 集群,将统一内存扩展至 1TB,可本地运行前沿级模型,包括 Z.ai 的完整 GLM 5.3(753B)。NVIDIA Sync 的 Cluster Assistant 负责集群设置与设备管理,八节点支持将于年底推出。基准测试显示,八节点集群上 GLM 5.3 Flash 生成速度约 80 tokens/s,Qwen3.8-27B 解码吞吐从单节点约 12 TPS 提升至八节点 52 TPS。

以上摘要由 AI 生成,可能存在误差。事实请以原文为准。

正文节选

AI developers aren’t just running a single model anymore—they’re orchestrating multiple models in parallel, chaining inference calls, and experimenting with frontier-class architectures that demand hundreds of gigabytes of memory. NVIDIA DGX Spark was built to bring AI development to the desk, and now you’ll be able to link even more Spark nodes together. At GTC Berlin, NVIDIA announced official support for eight-node DGX Spark clusters, scaling unified memory to 1 TB and opening the door to fro


发布时间:2026-10-23 00:00
抓取时间:2026-10-23 00:20
来源机构:NVIDIA
阅读原文developer.nvidia.com