跳到正文
实时
NVIDIA Technical Blog· Farshad Ghodsian·· 10 小时前58

NVIDIA Vera Rubin NVL72 在 MLPerf Training v6.1 中达到 2 倍 Blackwell Ultra 性能

NVIDIA Vera Rubin NVL72 Achieves up to 2x Blackwell Ultra Performance in MLPerf Training v6.1

要点速览

NVIDIA Vera Rubin NVL72 在 MLPerf Training v6.1 评测中达到 2 倍 Blackwell Ultra 的性能。其 CUDA 软件栈已高度成熟。

正文

As AI agents tackle increasingly complex workflows, reducing time-to-train is critical for rapidly iterating on both models and the agents they power. MLPerf...

As AI agents tackle increasingly complex workflows, reducing time-to-train is critical for rapidly iterating on both models and the agents they power. MLPerf Training v6.1 marks the debut of the NVIDIA Vera Rubin architecture, and these early results demonstrate that its CUDA software stack is already highly mature. In its first appearance, the Vera Rubin NVL72 platform delivered the fastest time…

Source

来源: NVIDIA Technical Blog · developer.nvidia.com